CN101365950A - Methods and materials for identifying the origin of a carcinoma of unknown primary origin - Google Patents

Abstract

The present invention provides a method of identifying origin of a metastasis of unknown origin by obtaining a sample containing metastatic cells; measuring Biomarkers associated with at least two different carcinomas; combining the data from the Biomarkers into an algorithm where the algorithm normalizes the Biomarkers against a reference; and imposes a cut-off which optimizes sensitivity and specificity of each Biomarker, weights the prevalence of the carcinomas and selects a tissue of origin determining origin based on highest probability determined by the algorithm or determining that the carcinoma is not derived from a particular set of carcinomas; and optionally measuring Biomarkers specific for one or more additional different carcinoma, and repeating the steps for additional Biomarkers.

Description

Be used to differentiate the method and the material of the origin of the not clear cancer in former initiation source

Invention field

The invention provides the material, method, algorithm, kit of the origin that is used to differentiate the not clear cancer in former initiation source etc.

Background of invention

Former to send out cancer (CUP) not clear be one group of heterogeneous, that biopsy confirms malignant tumour, wherein has metastatic disease, and it does not have identifiable primary tumor site or tissue of origin (ToO).This problem is represented about 3-5% of all cancers, makes it become the 7th kind of modal malignant tumour.Ghosh etc. (2005); With (2004) such as Mintzer.Patient's prognosis and therapeutic scheme depend on the origin of primary tumo(u)r, and this has emphasized differentiating the needs of primary tumor site.Greco etc. (2004); Lembersky etc. (1996); With (1994) such as Schlag.

At present, many kinds of methods are used to address this problem.Following several method is shown among Fig. 1-2.The blood serum tumor mark can be used for antidiastole.Although they lack enough specificitys, they can be used in combination with pathology and clinical information.Ghosh etc. (2005).Immunohistochemistry (IHC) method can be used to differentiate the tumour pedigree, but considerably less IHC mark is 100% specific.Therefore, the virologist often uses one group of IHC mark.Several researchs have used 4-14 kind IHC mark to confirm the accuracy of 66-88%.Brown etc. (1997); DeYoung etc. (2000); With (2005a) such as Dennis.More expensive deagnostic test comprises formation method, chest x-ray for example, computed tomography (CT) scanning, and PET (positron emission tomography) (PET) scanning.Every kind of such method can be differentiated former in the 30-50% case.Ghosh etc. (2005); With (2003) such as Pavlidis.Although these complicated technology are arranged, the ability of (ante mortem) resolution CUP case only is 20-30% before dead.Pavlidis etc. (2003); With (2004) such as Varadhachary.

A kind of promising new method is to analyze the ability that complete genomic gene expression overview is differentiated the tumour origin.Ma etc. (2006); Dennis etc. (2005b); Su etc. (2001); Ramaswamy etc. (2001); Bloom etc. (2004); Giordano etc. (2001); With 20060094035.These studies confirm that the feasibility of differentiating tissue of origin based on the gene expression overview.Be used for clinical application for these are expressed the profile analysis technology, must overcome 2 big obstacles.At first, because the gene expression profile analysis carries out at former tissue fully,, in metastasis, keep their tissue specific expression with confirmation so must verify the genetic marker material standed at the transfer tissue.Secondly, gene expression profile analysis technology must be able to be utilized (FFPE) formalin fixed, paraffin-embedded tissue, because immobilized tissue sample is the standard material in the present practice.Formalin fixed causes the degraded (Lewis etc. (2001) of RNA; With (1999) such as Masuda), so existing microarray rules can not be carried out reliably.Bibikova etc. (2004).In addition, described profile analysis technology must be strong, reproducible and can be approaching easily.

Show that quantitatively RTPCR (qRTPCR) produces reliable result from the FFPE tissue.Abrahamsen etc. (2003); Specht etc. (2001); Godfrey etc. (2000); With (2004) such as Cronin.Therefore, a more practical method is to use complete genomic method as discovering tool, and exploitation is based on the diagnostic assay of stronger technology.Ramaswamy(2004)。But this example need be developed littler gene sets.Oien and colleague thereof use the serial analysis (SAGE) of gene expression to differentiate 61 kinds of tumor markers, and they have therefrom developed the RTPCR method based on 11 genes of 5 tumor types.Dennis etc. (2002).With another research that SAGE and qRTPCR combine, developed the group of 5 genes of 4 tumor types, and reached 81% accuracy.Buckhaults etc. (2003).Nearest research combines the microarray profile analysis with qRTPCR, but has been to use 79 kinds of marks.Tothill etc. (2005).

Summary of the invention

The invention provides the method for the origin of the not clear metastasis of following discriminating origin: the sample that obtains to contain transitional cell; Measure different Cancer-Related biomarkers with at least 2 kinds; To advance in the algorithm wherein said algorithm from the data combination of biomarker: the relative described biomarker of reference standardization; With employing cutoff, the sensitivity and the specificity of every kind of biomarker of its optimization, make the morbidity rate weighted sum of cancer select tissue of origin; Based on the maximum probability of measuring by algorithm, determine origin, or determine that this cancer is not to be derived from the particular cancer set; Randomly measure the specific biomarker of one or more other different carcinoma, and be other biomarker repeating said steps.

The accompanying drawing summary

Fig. 1-2 has described the art methods of the origin of differentiating the not clear metastasis of origin.

Fig. 3 has described existing C UP diagnosis algorithm.

Fig. 4 has described the microarray data of the intensity that shows one group of 2 gene in the tissue.(A) prostate stem cell antigen (PSCA).(B) coagulation factors V (F5).Bar chart has shown intensity on the y-axle, and has shown tissue on the x-axle.Panc Ca, the pancreas cancer; Panc N, normal pancreas.

Fig. 5 has described the electrophoresis pattern that obtains from the Agilent biological analyser.Use 3 hours (A) or 16 hours (B) protease K digestings, from FFPE separate tissue RNA.Sample C22 (redness) is 1 age piece, and sample C23 (blueness) is 5 age pieces.The size ladder is shown in green.

Fig. 6 has described the contrast of the Ct value that obtains from 3 kinds of different qRTPCR methods: the random hexamer reverse transcription causes, carry out qPCR (RH 2 steps) with the cDNA that obtains subsequently, gene in reverse transcription-specific (reverse primer) causes, carry out qPCR (GSP 2 steps) with the cDNA that obtains subsequently, or in one-step reaction gene-specific initiation and qRTPCR (GSP1 step).To divide from the RNA of 11 samples in 3 kinds of methods, and measure the rna level of 3 genes: beta-actin (A), HUMSPB (B), and TTF (C).Indicate the intermediate value Ct value that obtains with every kind of method with solid line.

Fig. 7 has described CUP assay plate figure.

A series of figure among Fig. 8 have described the mensuration performance of certain RNA concentration range.

Fig. 9 is a width of cloth experimental work process flow diagram: the mark candidate thing is specified and checking (9A); Optimization and prediction algorithm are set up and test (9B) with measuring.

Figure 10 has described the expression of tissue-specific genetic marker material standed in FFPE metastatic carcinoma and former gland cancer of prostate of 10 kinds of selections.For every width of cloth figure, X-axis is represented standardized marker expression value.

Figure 11 has described the mensuration optimization.The electrophoresis pattern that (A and B) obtains from the Agilent biological analyser.Use 3 hours (A) or 16 hours (B) protease K digestings, from FFPE separate tissue RNA.Sample C22 (redness) is 1 age piece, and sample C23 (blueness) is 5 age pieces.The size ladder is shown in green.The contrast of the Ct value that (C and D) obtains from 3 kinds of different qRTPCR methods: the random hexamer reverse transcription causes, carry out qPCR (RH 2 steps) with the cDNA that obtains subsequently, gene in reverse transcription-specific (reverse primer) causes, carry out qPCR (GSP 2 steps) with the cDNA that obtains subsequently, or in one-step reaction gene-specific initiation and qRTPCR (GSP1 step).To divide from the RNA of 11 samples in 3 kinds of methods, and measure the rna level of 2 genes: beta-actin (C), HUMSPB (D).Indicate the intermediate value Ct value that obtains with every kind of method with solid line.

Figure 12 is the thermal map (heatmap) that shows the relative expression level of 10 mark groups between 239 samples.The higher expression of red indication.

Describe in detail

Differentiate the original site among not clear (CUP) the metastatic carcinoma patient in former initiation source, can realize the application of particular treatment, and can prolong survival. Then by reverse transcriptase polymerase chain reaction (RT-PCR), being derived from the 205 FFPE metastatic carcinomas of these 6 kinds of tissues and being derived from the MET of other type of cancer, verified the mark candidate thing, to measure specificity. Select the 10-genetic marker, it can predict the metastatic carcinoma tissue of origin of these 6 kinds of type of cancer. Then, for these 10 mark optimization RNA separate and the qRTPCR method, and qRTPCR mensuration is applied to one group of 260 metastatic tumour, thereby produces 78% overall accuracy. At last, 48 independent set of shifting sample have been tested. Importantly, 37 samples in this set have known primary carcinoma, or show as at first CUP, but differentiated that subsequently and this mensuration shows 78% the degree of accuracy.

Biomarker is any mark of expression of the marker gene of indication.This mark can be direct or indirect, and can measure under given physiological parameter, and compares with internal contrast, normal structure or another kind of cancer, and the overexpression of this gene or expression are not enough.Biomarker includes but not limited to, nucleic acid (overexpression and expression are not enough, and direct and indirect).Use nucleic acid as biomarker, can comprise any method known in the art, include but not limited to, measure DNA cloning, RNA, microRNA, loss of heterozygosity (LOH), single nucleotide polymorphism (SNPs, Brookes (1999)), microsatellite DNA, dna methylation is not enough or excessively methylate.Use albumen to comprise any method known in the art, include but not limited to as biomarker, measuring amount, activity is modified as glycosylation, phosphorylation, ADP-ribosylation, ubiquitination etc., or immunohistochemistry (IHC).Other biomarker comprises imaging, cell count and Apoptosis mark.

The gene of indication provided herein is those genes relevant with specific tumors or types of organization.Marker gene can be relevant with many cancer types; but prerequisite is; this expression of gene is that specific a kind of tumour or types of organization are fully relevant with using algorithm described herein to be accredited as specific origin; this gene can be used for claimed invention, to determine the tissue of origin of the not clear cancer (CUP) in former initiation source.Many genes relevant known in the art with one or more cancers.The invention provides preferred marker gene and even preferred marker gene combination.They are described in detail in this article.

Be meant types of organization's (lung, colon etc.) or histological type (gland cancer, squamous cell carcinoma etc.) as " origin " mentioned in " tissue of origin ", this depends on the certain medical environment, and will be understood by those skilled in the art.

When marker gene contained the sequence of SEQ ID NO appointment, it was corresponding with this sequence.When constant gene segment C or fragment contain when being enough to distinguish that it is the part of the canonical sequence of sequence of this gene or its complement, it is corresponding with the sequence of such gene.When its RNA, mRNA or cDNA and composition when hybridization with such sequence (for example probe), perhaps, under the situation of peptide or albumen, it is during by such mRNA coding, and gene expression product is corresponding with such sequence.When it contains when being enough to distinguish that it is the part of the reference gene expression product of sequence of this gene or gene expression product or its complement, the section of gene expression product or fragment are corresponding with the sequence of such gene or gene expression product.

Described in this instructions and claimed method of the present invention, composition, article and kit comprise one or more marker gene." mark " or " marker gene " is used in reference in running through this instructions and corresponding gene of any gene and gene expression product, and its overexpression or express not enough relevant with tumour or types of organization.Preferred marker gene has been described in table 1 in more detail.

Table 1

The CUP group

SEQ ID NO：	Title	The chip title	Sequence
					1	SP-B	209810_at	gaaaaaccagccactgctttacaggacagggggttgaagctgagccccgc ctcacacccacccccatgcactcaaagattggattttacagctacttgcaatt caaaattcagaagaataaaaaatgggaacatacagaactctaaaagataga catcagaaattgttaagttaagctttttcaaaaaatcagcaattccccagcgta gtcaagggtggacactgcacgctctggcatgatgggatggcgaccgggc aagctttcttcctcgagatgctctgctgcttgagagctattgctttgttaagatat aaaaaggggtttctttttgtctttctgtaaggtggacttccagattttgattgaaa gtcctagggtgattctatttctgctgtgatttatctgctgaaagctcagctggg gttgtgcaagctagggacccattcctgtgtaatacaatgtctgcaccaatgct
2	TTFI	211024_s_at	gtgattcaaatgggttttccacgctagggcggggcacagattggagagggc tctgtgctgacatggctctggactctaaagaccaaacttcactctgggcaca ctctgccagcaaagaggactcgcttgtaaataccaggatttttttttttttttgaa gggaggacgggagctggggagaggaaagagtcttcaacataacccactt gtcactgacacaaaggaagtgccccctccccggcaccctctggccgccta ggctcagcggcgaccgccctccgcgaaaatagtttgtttaatgtgaacttgt agctgtaaaacgctgtcaaaagttggactaaatgcctagtttttagtaatctgt acattttgttgtaaaaagaaaaaccactcccagtccccagcccttcacatttttt atgggcattgacaaatctgtgtatattatttggcagtttggtatttgcggcgtca gtctttttctgttgtaact
				3	DSG3	205595_at	ccatcccatagaagtccagcagacaggatttgttaagtgccagactttgtca ggaagtcaaggagcttctgctttgtccgcctctgggtctgtccagccagctgt ttccatccctgaccctctgcagcatggtaactatttagtaacggagacttactc ggcttctggttccctcgtgcaaccttccactgcaggctttgatccacttctcac acaaaatgtgatagtgacagaaagggtgatctgtcccatttccagtgttcctg gcaacctagctggcccaacgcagctacgagggtcacatactatgctctgta cagaggatccttgctcccgtctaatatgaccagaatgagctggaataccaca ctgaccaaatctggatctttggactaaagtattcaaaatagcatagcaaagct cactgtattgggctaataatttggcacttattagcttctctcataaactgatcac gattataaattaaatgtttgggttcataccccaaaagcaatatgttgtcactcct aattctcaagtac
4	HPT1	209847_at	ctgcacccacctacttagatatttcatgtgctatagacattagagagatttttca tttttccatgacatttttcctctctgcaaatggcttagctacttgtgtttttccctttt ggggcaagacagactcattaaatattctgtacattttttctttatcaaggagata tatcagtgttgtctcatagaactgcctggattccatttatgttttttctgattccatc ctgtgtccccttcatccttgactcctttggtatttcactgaatttcaaacatttgtc
				5	PSCA	205319_at	ttcctgaggcacatcctaacgcaagtttgaccatgtatgtttgcaccccttttcc ccnaaccctgaccttcccatgggccttttccaggattccnaccnggcagatc agttttagtganacanatccgcntgcagatggcccctccaaccntttntgttg ntgtttccatggcccagcattttccacccttaaccctgtgttcaggcacttnttc ccccaggaagccttccctgcccaccccatttatgaattgagccaggtttggt ccgtggtgtcccccgcacccagcaggggacaggcaatcaggagggccc agtaaaggctgagatgaagtggactgagtagaactggaggacaagagttg acgtgagttcctgggagtttccagagatg

6	F5	204713_s_at	atcctctacagccagatgtcacagggatacgtctactttcacttggtgctgga gaattcanaagtcaagaacatgctaagcntaagggacccaaggtagaaag agatcaagcagcaaagcacaggttctcctggatgaaattactagcacataa agttgggagacacctaagccaagacactggttctccttccggaatgaggcc ctgggaggaccttcctagccaagacactggttctccttccagaatgaggcc ctggaaggaccctcctagtgatctgttactcttaaaacaaagtaactcatctaa gattttggttgggagatggcatttggcttctgagaaaggtagctatgaaataat ccaagatactgatgaagacacagctgttaacaattggctgatcagccccca gaatgcctcacgtgcttggggagaaagcacccctcttgccaacaagcctgg aaag
				7	MGB1	206378_at	gcagcagcctcaccatgaagttgctgatggtcctcatgctggcggccctctc ccagcactgctacgcaggctctggctgccccttattggagaatgtgatttcca agacaatcaatccacaagtgtctaagactgaatacaaagaacttcttcaaga gttcatagacgacaatgccactacaaatgccatagatgaattgaaggaatgt tttcttaaccaaacggatgaaactctgagcaatgttgaggtgtttatgcaatta atatatgacagcagtctttgtgatttattttaactttctgcaagacctttggctca cagaactgcagggtatggtgagaaaccaactacggattgctgcaaaccac accttctctttcttatgtctttttact
8	PDEF	220192_x_at	gagtggggcccttaaactggattcaaaaaatgctctaaacataggaatggtt gaagaggtcttgcagtcttcagatgaaactaaatctctagaagaggcacaa gaatggctaaagcaattcatccaagggccaccggaagtaattagagctttg aaaaaatctgtttgttcaggcagagagctatatttggaggaagcattacagaa cgaaagagatcttttaggaacagtttggggtgggcctgcaaatttagaggct attgctaagaaaggaaaatttaataaataattggtttttcgtgtggatgtactcc aagtaaagctccagtgactaatatgtataaatgttaaatgatattaaatatgaa catcagttaaaaaaaaaattctttaaggctactattaatatgcagacttactttta atcatttgaaatctgaactcatttacctcatttcttgccaattactcccttgggtat ttactgcgta
				9	PSA	204582_s_at	tggtgtaattttgtcctctctgtgtcctggggaatactggccatgcctggagac atatcactcaatttctctgaggacacagataggatggggtgtctgtgttatttgt ggggtacagagatgaaagaggggtgggatccacactgagagagtggag agtgacatgtgctggacactgtccatgaagcactgagcagaagctggagg cacaacgcaccagacactcacagcaaggatggagctgaaaacataaccc actctgtcc
10	WT1	206067_s_at	atagatgtacatacctccttgcacaaatggaggggaattcattttcatcactgg gagtgtccttagtgtataaaaaccatgctggtatatggcttcaagttgtaaaaa tgaaagtgactttaaaagaaaataggggatggtccaggatctccactgataa gactgtttttaagtaacttaaggacctttgggtctacaagtatatgtgaaaaaa atgagacttactgggtgaggaaatccattgtttaaagatggtcgtgtgtgtgt gtgtgtgtgtgtgtgtgttgtgttgtgttttgttttttaagggagggaatttattatt taccgttgcttgaaattactgtgtaaatatatgtctgataatgatttgctctttgac aactaaaattaggactgtataagtactagatgcatcactgggtgttgatcttac aagat

The invention provides the method for the origin of the not clear metastasis of following discriminating origin: measure different Cancer-Related biomarkers with at least 2 kinds in the sample that contains transitional cell; To advance in the algorithm wherein said algorithm from the data combination of biomarker: the relative described biomarker of reference standardization; With employing cutoff, the sensitivity and the specificity of every kind of biomarker of its optimization, make the morbidity rate weighted sum of cancer select tissue of origin; Based on the maximum probability of measuring by algorithm, determine origin, or determine that this cancer is not to be derived from the particular cancer set; Randomly measure the specific biomarker of one or more other different carcinoma, and be other biomarker repeating said steps.

The invention provides the method for the origin of the not clear metastasis of following discriminating origin: the sample that obtains to contain transitional cell; Measure different Cancer-Related biomarkers with at least 2 kinds; To advance in the algorithm wherein said algorithm i from the data combination of biomarker) the relative described biomarker of reference standardization; Ii) adopt cutoff, the sensitivity and the specificity of every kind of biomarker of its optimization make the morbidity rate weighted sum of cancer select tissue of origin; Based on the maximum probability of measuring by algorithm, determine origin, or determine that this cancer is not to be derived from the particular cancer set; Randomly measure to the specific biomarker of one or more other different carcinoma with for other biomarker repeating step c) and d).

In one embodiment, described marker gene is selected from: i) SP-B, TTF, DSG3, KRT6F, p73H or SFTPC; Ii) F5, PSCA, ITGB6, KLK10, CLDN18, TR10 or FKBP10; And/or iii) CDH17, CDX1 or FABP1.Preferably, described marker gene is SP-B, TTF, DSG3, KRT6F, p73H and/or SFTPC.More preferably, described marker gene is SP-B, TTF and/or DSG3.Described marker gene can also comprise or replace with KRT6F, p73H and/or SFTPC.

In one embodiment, described marker gene is F5, PSCA, ITGB6, KLK10, CLDN18, TR10 and/or FKBP10.More preferably, described marker gene is F5 and/or PSCA.Preferably, described marker gene can comprise or replace with ITGB6, KLK10, CLDN18, TR10 and/or FKBP10.

In another embodiment, described marker gene is CDH17, CDX1 and/or FABP1, preferred CDH17.Described marker gene can also comprise or replace with CDX1 and/or FABP1.

In one embodiment, use at least one among the SEQ ID NO:11-58, measure gene expression.

The present invention also comprises the method for following measurement gene expression: obtain and measure at least one formation among the amplicon SEQID NO:14,18,22,26,30,34,38,42,46,50,54 and/or 58.

In one embodiment, described marker gene can be selected from the sex-specific marker of selecting at least one from following: i) under male patient's situation, and KLK3, KLK2, NGEP or NPY; Or ii) under the female patient situation, PDEF, MGB, PIP, B305D, B726 or GABA-Pi; And/or WT1, PAX8, STAR or EMX2.Preferably, described marker gene is KLK2 or KLK3.In this embodiment, described marker gene can comprise or replace with NGEP and/or NPY.In one embodiment, described marker gene is PDEF, MGB, PIP, B305D, B726 or GABA-Pi, preferred PDEF and MGB.In this embodiment, described marker gene can comprise or replace with PIP, B305D, B726 or GABA-Pi.In one embodiment, described marker gene is WT1, PAX8, STAR or EMX2, preferred WT1.In this embodiment, described marker gene can comprise or replace with PAX8, STAR or EMX2.

The invention provides following method, obtain additional clinical information, comprise metastasis site, to determine the origin of cancer; Obtain the biomarker set for cancer the best, it comprises following step: use the metastasis of known origin, measure its biomarker, and contrast the biomarker of described biomarker and the not clear metastasis of origin; Fail to understand the origin of metastasis and differentiate its suitable treatment by determining to originate from, orientation treatment is provided; And the origin by measuring the not clear metastasis of origin and differentiate its corresponding prognosis, prognosis is provided.

The present invention provides the method for following discovery biomarker in addition: measure mark expression of gene level in the specific metastasis, the biomarker of measurement markers gene, to determine its expression, provide or any method known in the art according to this paper, the evaluation of markers expression of gene, and determine that whether described marker gene is effectively specific to the origin tumour.

The present invention provides composition in addition, and it contains the sequence of at least one separation that is selected from SEQ ID NO:11-58.The present invention provides kit in addition, and it is used to carry out the mensuration according to method provided herein, and contains biological markers detection reagent in addition.

The present invention provides microarray or the genetic chip that is used to carry out methods described herein in addition.

The present invention provides diagnosis/prognosis combination (portfolio) in addition, it comprises the isolated nucleic acid sequences of the assortment of genes described herein, their complement or its part, wherein said combination be enough to measure or the characterising biological sample in gene expression, with compare from the cell of different carcinoma or normal structure, described biological sample has transitional cell.

Any means of the present invention can also comprise, measures at least one expression of gene of constitutive expression in sample.

Preferably, the mark of pancreas cancer be coagulation factors V (F5), prostate stem cell antigen (PSCA), integrin, β 6 (ITGB6), KLK10 (KLK10), close protein 18 (CLDN18), three-piece albumen isoform (trio isoform) (TR10) and be similar to the protein-bonded albumen FLJ22041 (FKBP10) that infers of FK506.Preferably, measure the biomarker of F5 and PSCA together.Except F5 and/or PSCA, or as an alternative, can measure the biomarker of ITGB6, KLK10, CLDN18, TR10 and FKBP10.F5 for example is described in 20040076955; 20040005563; In WO2004031412.PSCA is described in for example WO1998040403; 20030232350; In WO2004063355.ITGB6 is described in for example WO2004018999; In 6339148.KLK10 is described in for example WO2004077060; In 20030235820.CLDN18 is described in for example WO2004063355; In WO2005005601.TR10 for example is described in 20020055627.FKBP10 for example is described among the WO2000055320.

Preferably, the marker gene of colon cancer is intestines peptide-relevant transport protein HPT-1 (CDH17), tail type homeobox transcription factor 1 (CDX1) and fatty acid binding protein 1 (FABP1).Preferably, the biomarker of independent measurement CDH17.Except the biomarker of CDH17, or as an alternative, can measure the biomarker of CDX1 and FABP1.CDH17 is described in for example Takamura etc. (2004); In WO2004063355.CDX1 is described in for example Pilozzi etc. (2004); 20050059008; In 20010029020.FABP1 is described in for example Borchers etc. (1997); Chan etc. (1985); Chen etc. (1986); In (1985) such as Lowe.

Preferably, the marker gene of lung cancer is surfactant albumen-B (SP-B), thyroid gland transcription factor (TTF), desmoglein 3 (DSG3), keratin 6 isoform 6F (KRT6F), p53-related gene (p73H) and surfactant PROTEIN C (SFTPC).Preferably, measure the biomarker of SP-B, TTF and DSG3 together.In the biomarker of SP-B, TTF and/or DSG3 any, or as an alternative, can measure the biomarker of KRT6F, p73H and SFTPC.SP-B is described in for example Pilot-Mathias etc. (1989); 20030219760; In 20030232350.TTF is described in for example Jones etc. (2005); US20040219575; WO1998056953; WO2002073204; 20030138793; In WO2004063355.DSG3 is described in for example Wan etc. (2003); 20030232350; WO2004030615; In WO2002101357.KRT6F is described in for example Takahashi etc. (1995); 20040146862; In 20040219572.P73H is described in for example Senoo etc. (1998); In 20030138793.SFTPC for example is described among Glasser etc. (1988).

Described marker gene can further be selected from sex-specific marker, for example, under male patient's situation, KLK3, KLK2, NGEP or NPY; Or under the situation of female patient, PDEF, MGB, PIP, B305D, B726 or GABA-Pi; And/or WT1, PAX8, STAR or EMX2.

Preferably, the marker gene of breast cancer be the prostate epithelium factor (PDEF) of deriving, mammary gland globin (mammaglobin) (MG), prolactin-inducible protein (PIP), B305D, B726 and GABA-π.Preferably, measure the biomarker of PDEF and MG together.Except the biomarker of PDEF and/or MG, or as an alternative, can measure the biomarker of PIP, B305D, B726 and GABA-Pi.PDEF is described in for example WO2004030615; WO2000006589; WO2001073032; Wallace etc. (2005); Feldman etc. (2003); In (2000) such as Oettgen.MG is described in for example WO2004030615; 20030124128; Fleming etc. (2000); Watson etc. (1996 and 1998); In 5668267.PIP is described in for example Autiero etc. (2002); Clark etc. (1999); In (1987) such as Myal etc. (1991) and Murphy.B305D, B726 and GABA-Pi are described in (2005) such as Reinholz.NGEP for example is described among Bera etc. (2004).

Preferably, the mark of oophoroma is WilmShi knurl 1 (WT1), PAX8, steroids generation acute regulation protein (STAR) and EMX2.Preferably, measure the biomarker of WT1.Except the biomarker of WT1, or as an alternative, can measure the biomarker of STAR and EMX2.WTl for example is described in 5350840; 6232073; 6225051; 20040005563; In (2003) such as Bentov.PAX8 for example is described in 20050037010; Poleev etc. (1992); (2003) such as Di Palma; Marques etc. (2002); Cheung etc. (2003); Goldstein etc. (2002); Oji etc. (2003); Rauscher etc. (1993); Zapata-Benavides etc. (2002); In (2003) such as Dwight.STAR is described in for example Gradi etc. (1995); In (2003) such as Kim.EMX2 for example is described among Noonan etc. (2001).

Preferably, the mark of prostate cancer is KLK3, KLK2, NGEP and NPY.Preferably, measure the biomarker of KLK3.Except KLK3, or as an alternative, can measure the biomarker of KLK2, NGEP and NPY.KLK2 and KLK3 for example are described among Magklara etc. (2002).KLK2 for example is described in 20030215835; In 5786148.KLK3 for example is described in 6261766.

Described method also can comprise, obtains additional clinical information, comprises metastasis site, to determine the origin of cancer.Process flow diagram is provided in Fig. 3.

The present invention provides the following method that obtains for the biomarker set of cancer the best in addition: use the metastasis of known origin, measure its biomarker, and contrast the biomarker of described biomarker and the not clear metastasis of origin.

The present invention provides the following method that orientation treatment is provided in addition: determine the origin of the not clear metastasis of origin and differentiate its suitable treatment according to methods described herein.

The present invention provides the following method that prognosis is provided in addition: measure the origin of the not clear metastasis of origin and differentiate its corresponding prognosis according to methods described herein.

The present invention provides the method for finding biomarker in addition, it comprises, measure mark expression of gene level in the specific metastasis, the biomarker of measurement markers gene, to determine its expression, according to methods described herein evaluation of markers expression of gene, and determine that whether described marker gene is effectively specific to the origin tumour.

The present invention provides composition in addition, and it comprises the sequence of at least one separation that is selected from SEQ ID NO:11-58.

The present invention provides kit, article, microarray or genetic chip, diagnosis/prognosis combination that is used to carry out mensuration described herein and patient's report of reporting the result who obtains by the inventive method in addition.

Only rarely find, only specific nucleic acid sequence in tissue sample existence or do not exist, just have diagnosis or prognostic value.On the other hand, about the information of the expression of various albumen, peptide or mRNA, it is important to seem day by day.Only have the existence of nucleotide sequence (such sequence is called " gene ") in genome of the potentiality of expressing protein, peptide or mRNA itself, can not determine whether albumen, peptide or mRNA express in given cell.Whether the given gene of energy expressing protein, peptide or mRNA expresses, and in what degree such expression takes place, if take place, by many complicated factors decisions.Do not consider to understand and estimate the difficulty of these factors, mensuration gene expression can provide the useful information about the generation of critical event, for example tumour generation of described critical event, transfer, Apoptosis and other relevant clinically phenomenon.In the gene expression overview, can find to have the relevant indication of activity or non-activity degree about gene.Gene expression overview of the present invention is used to CUP that diagnosis and treatment patient are provided.

Specimen preparation need be collected patient's sample.Patient's sample of Shi Yonging is the sample that contains diseased cells under a cloud in the method for the invention, the cell that described diseased cells for example obtains from brief summary in organizing fine needle aspiration (FNA).The preparation of organizing greatly that obtains from biopsy or operation sample and laser capture microdissection also is suitable for.Laser capture microdissection (LCM) technology is the mode of the cell that will study of a kind of selection, and it minimizes the variability that is caused by the cell type unevenness.As a result, can easily detect the medium or little variation that the marker gene between normal or optimum and the cancerous cells is expressed.Sample also can comprise the circulation epithelial cell that extracts from peripheral blood.These can obtain according to many methods, but most preferred method is 6136182 described magnetic separation techniques.In case obtained containing the sample of target cell, just used the biomarker of the gene in the appropriate combination to obtain the gene expression overview.

The method for optimizing of establishing the gene expression overview comprises, measures by can encoding proteins or the amount of the RNA that produces of the gene of peptide.This can realize by following method: reverse transcriptase PCR (RT-PCR), competition RT-PCR, real-time RT-PCR, differential RT-PCR, rna blot analysis and other correlation test.Although may use single PCR reaction to carry out these technology, the complementary DNA (cDNA) or the complementary RNA (cRNA) that preferably increase and produce from mRNA, and by microarray analysis it.Many different array configurations and their production method are well known by persons skilled in the art, and for example are documented in 5445934; 5532128; 5556752; 5242974; 5384261; 5405783; 5412087; 5424186; 5429807; 5436327; 5472672; 5527681; 5529756; 5545531; 5554501; 5561071; 5571639; 5593839; 5599695; 5624711; 5658734; In 5700637.

Microarray technology allows to measure simultaneously the steady-state mRNA level of thousands of genes, thereby for identifying as the effect of beginning, stagnation or the adjusting of uncontrolled cell proliferation provides strong tool.Be extensive use of two kinds of microarray technologies at present.CDNA array and oligonucleotide arrays.Though there is structural differences in these chips, all basically downstream data analyses are identical with output.The result of these analyses generally is the measured value of the signal intensity that receives of the label probe of the cDNA sequence from be used for detecting the sample with the nucleic acid array hybridizing of microarray known location.Usually, signal intensity and the cDNA that in sample cell, expresses and therefore the amount of mRNA is proportional.A large amount of this technology can obtain with useful.Determine gene expression preferable methods can referring to, 6271002; 6218122; 6218114; With 6004755.

Can analyze expression by more described signal intensity.This is preferably undertaken by the rate matrix that generates gene expression intensity in test specimen and the control sample.For example, can the gene expression intensity of illing tissue be compared with the expression intensity that produces from the optimum or normal structure of same type.The ratio of these expression intensities has shown that the multiple in gene expression changes between test specimen and the control sample.

Selection can be based on statistical test, described statistical test produce with the original origin of tumour position correlation factor between the differentially expressed relevant sequencing table of conspicuousness evidence of each gene.The example of such check comprises ANOVA and Kruskal-Wallis.Ordering can be as the weight in the model, and described modelling is used for the summation of these weightings, be up to the evidence that cutoff is interpreted as helping a class surpasses another kind of.Evidence also can be used to regulate weight before document was described.

In the present invention, selected 10 marks, they show marked difference expression evidence in 6 tumor types.Selection course comprises, the special collection of statistical test, mean variance optimization, and expertise.In an alternate embodiment, can make the feature extracting method robotization, select and test badge with learning method by supervision.Along with the generation of database, selection that can repeating label is to be created in any given database positioning the highest possible diagnosis accuracy.

Preferred embodiment is to gather by differentiating stable contrast, and this set is scaled to zero variance between all samples, each measured value of standardization.This contrast sets definition be any single endogenous transcript, or the set of endogenous transcript, and the systematic error during it is measured influences, and does not know it to be independent of this error and change.Institute is underlined with the specific factor adjusting of sample, and the described factor produces zero variance for any descriptive statistic of contrast set (for example mean value or intermediate value) or for direct measurement.Perhaps, if the prerequisite that only relevant with systematic error contrast changes is false, yet the error in classification that produces when carrying out standardization is littler, and the contrast set will be still according to described use.Non-endogenous peak value (spike) contrast also is useful, but is not preferred.

After mark is selected, the variable of these selections is used for sorter (classifier), described classifier design is used to produce high as far as possible classify accuracy.Can use the learning algorithm of supervision, it is designed for the output set of related input measurement value set and predicted value (predictor), predicts tissue of origin to set up model from 10 input values.Problem can be described as: given training data { (x ₁, y) ..., (x _n, y) } and produce sorter h:x → y, it is mapped to sample x ∈ x its tissue of origin mark y ∈ y.That analyzes before described prediction is based on is included in example in the database, thereby forms the training set.

Based on the relation of input variable and known output valve, the learning algorithm of supervision should find to make the minimized parameter of error in classification of expection.These parameters can be used for then from fresh sample input value prediction tissue of origin.The example of these algorithms comprises linear classification model, quadratic classifier, based on method, neural network and the prototype method of tree for example k-nearest neighbor classifier or learning vector quantization algorithm.

A specific embodiments of building the model of 10 standardization marks is LDA methods, wherein uses default parameter, as described in Venables and Ripley (2002).This method has wherein provided the mean value of y class mark 0 and 1 based on the Fisher linear discriminant analysis ${\stackrel{\→}{\mathrm{\μ}}}_{y=0},{\stackrel{\→}{\mathrm{\μ}}}_{y=1}$ With covariance Σ _Y=0, Σ _Y=1, we seek

Linear combination, it will have mean value $\stackrel{\→}{\mathrm{\ω}}\·{\stackrel{\→}{\mathrm{\μ}}}_{y=i}$ And variance ${\stackrel{\→}{\mathrm{\ω}}}^T{\mathrm{\Σ}}_{y=i}\stackrel{\→}{\mathrm{\ω}},$ They will make the ratio maximization of the variance in variance between the classification and the classification:

LDA can be summarized in the multiclass discriminatory analysis, wherein y has N kind possibility state, rather than only 2 kinds.The value that comprises from the database of selection marquee is estimated class mean value and variance.In preferred embodiments, with the equal prior probability of the every kind of tumor type that carries out following processing, weighting covariance matrix.By the model prediction male patient, in this model, the prior probability of each female sex organ tumor group (prior) is 0.Similarly, by the model prediction female patient, in this model, genital orgnas,male's prior probability is 0.In the present invention, for women's prostate check, prior probability is 0, and checks for the male sex's breast and ovary, and prior probability is 0.In addition, have sample with classification mark same background by model testing, wherein for this particular category mark, prior probability is 0.

The problems referred to above can be regarded the maximization of the Rayleigh coefficient of handling as generalized eigenvalue problem as.The distance of the centre of moment by calculating each sample and selected subspace, the subspace that will reduce are used for classification.Can this model of match by maximum likelihood, and use the BayesShi theorem to calculate posterior probability.

A kind of alternative method can comprise, finds that n-dimensional feature space (the wherein number of the variable that is to use of n) to the reflection of a group categories mark, will be referred to described feature space is separated into the zone, will classification-designatedly arrive each zone then.Distance dependent between the scoring of these arest neighbors type algorithm and the decision border, and must not translate into class probability.

If too many alternative variables is arranged, and wherein many are random noises, and then Variables Selection and model emitted the danger that is fit to (over-fit) this problem.Therefore, at the sequencing table of various cutoffs through being commonly used for input value, with the number of dominated variable.Searching algorithm for example genetic algorithm also can be used for the choice variable subclass, because their inspection cost functions.The annealing that can attempt simulating is limited in the risk that local minimum is caught cost function.However, must verify these programs with the sample that is independent of selection and modeling process.

Also can use the latent variable method.Estimate the learning algorithm of any non-supervision in lower-dimensional topology space (manifold) from higher dimensional space, can be used to find that can there be the littler latent variable set of many matches well association between the input variable and they.Although the estimation of the validity that reduces is subjective, the algorithm application of supervision can be closed in the variables set that reduces, to estimate classify accuracy.Thereby, can also can make up from the sorter of latent variable structure from one group of variable with the latent variable significant correlation.An example can comprise, uses the variable relevant with essential component from essential component analysis, as the input value of the disaggregated model of any supervision.

These algorithms can implemented in the software code arbitrarily, and described software code has following method: input variable, train sample with function, and output to control desk based on this model measurement sample with the result.R, Octave, C, C++, Fortran, Java, Perl and Python have available library under the open-source permission, to be used to realize many functions of listing above.Commerce is for example wrapped, and S+ and Matlab also pack with many such methods.

Described code is carried out following step according to following order, uses the R version 2.2.1 (http://www.r-project.org) in MASS (Venables etc. (2002)) library has been installed.Term LDA refers to the lda function in MASS name space (name space).

1) is kept on the hard drives for the CT value of all available training set samples 10 marker gene and 2 contrasts.

2), deduct the specific mean value of sample of contrast, 10 marker gene values of standardization from each mark for each sample.

3) training dataset is made up of the metastasis at known origin position, and wherein each sample has at least one to the specific target indicia of the tissue of origin of mark, and it has the standardization CT value less than 5.

4) training data of LDA from (3) makes up 4 2LDA model sets.In each set, a model is that the male sex is specific, and has and be set at 0 breast and ovary prior odds (odds), and the prostate prior odds that is set at the prior odds of equal value of other classification mark.Another model of every centering is that the women is specific, has to be set at 0 prostate prior odds, and the breast and the ovary prior odds that are set at the prior odds of finding in other classification mark of equal value.

A. first set is for use in testing for the CUP sample of finding in the colon, and the prior odds of colon is set at 0, and all other non-reproduction classification flag settings are prior odds of equal value.

B. second model set is specific to the CUP that finds in ovary, and the prior odds of ovary is set at 0, and all other non-reproduction classification flag settings are prior odds of equal value.

C. the 3rd set is used for the CUP that finds at lung, and the prior odds of lung is set at 0.All other non-reproduction classification flag settings are prior odds of equal value.

D. this universal model is used for all other background tissues.Set all prior oddses, exception is the specific classification mark of the reproduction that sets in 4 of equal valuely.

For specimen, we move the R program, and it carries out following content.

1) reads in test data set.

2) the specific mean value of sample of 2 contrasts of generation.

3), deduct the specific mean value of sample from each mark for each sample.

4) replace from the CT of the arbitrary standardsization of 40 thick CT generations with 12.

5) for each sample in the test set, test following content.

If a. the mean value of 2 contrasts is greater than 34, then this identified as samples is designated as ' CTR_FAILURE ' with 0 posterior probability.

B. check the background of colon, ovary or lung.If the coupling of discovery is then also checked sex.The model of background and sex-specific is used for assess sample then.

If c. find breast, pancreas, lung SCC or prostate mark as a setting, then provide the mark of ' FAILURE_ineligible_ sample ', and posterior probability all is set at 0 for this sample.

D. the universal model of sex is used for all other samples.

With result's format, and be written to file.

The present invention includes the gene expression combination that obtains by this method.

The also available a lot of modes of gene expression overview are showed.The most frequently used is that original fluorescence intensity or rate matrix are arranged as tree derivation, and test specimen is shown in tabulation wherein, and the line display gene.Array data like this is so that it is closer to each other to have a gene of similar expression overview.Each expression of gene is than showing with color.For example, the ratio less than 1 (downward modulation) appears at the blue portion of spectrum, and appears at the red part of spectrum greater than 1 ratio (rise).The computer software programs that commercialization can get can be used for showing this data, comprise " GeneSpring " (SiliconGenetics, Inc.) and " Discovery " and " Infer " (Partek, Inc.).

From the metastatic tumo(u)r of the primary tumo(u)r of primary tumor or known origin, collect the measurement of the abundance of distinct rna species.These readings and clinography include but not limited to, the initial position of patient's age, sex, primary tumor and metastasis site (if being suitable for) are used to produce relational database.Described database is used to select the rna transcription thing and the clinical factor, and they can be as the token variable in the former initiation source of predicting metastatic tumour.

Measuring under the situation of protein level with definite gene expression, any method known in the art all is suitable for, and prerequisite is that it causes sufficient specificity and sensitivity.For example, by antibody or the antibody fragment of combination to protein-specific, and the amount of the albumen of measurement antibody-combination, can measure protein level.Can be with radioactive, fluorescence or other detectable reagent labelled antibody, to promote detection.The method that detects includes but not limited to enzyme linked immunosorbent assay (ELISA) (ELISA) and immunoblot assay.

The gene of being regulated that uses has in the method for the invention been described in an embodiment.Compare with those patients with Different Origin cancer, differentially expressed gene is raised or is reduced in the patient with specific origin cancer.Last mediation downward modulation is relative term, refers to that the gene expression amount found is with respect to the detectable difference of some baselines (surpassing the influence of the noise of the system that is used for measuring it).In this case, determine baseline based on algorithm.Then, use identical measuring method, with respect to the target gene in baseline values rise or the downward modulation diseased cells.In the present context, the change of ill finger condition, its interruption or upset the correct performance of body function, or have the correct potential of performance that upsets body function, as cell uncontrolled propagation took place.When some aspects of a people's genotype or phenotype were consistent with existing of disease, promptly diagnosable he suffered from disease.But, diagnose or the behavior of prognosis, can comprise definite disease/state issues, for example determine the possibility of recurrence, the type of treatment and treatment monitoring.In the treatment monitoring, by the gene expression that contrast is passed in time, determine whether the gene expression overview changes, or be varied to and the more corresponding to pattern of normal structure, make clinical judgment about given therapeutic process effect.

Gene can be divided into groups, make the reliable basis of correlated judgment (for example diagnosis, prognosis or treatment are selected) clinically thereby the information about gene sets in the group that obtains provides.These gene sets constitute combination of the present invention.The same with most of diagnostic flags, often need to use the minimized number mark that is enough to make correct medical judgment.This can prevent the delay of the treatment during further analyzing, and the unworthy utilization of time and resource.

A method of establishing the gene expression combination is by the use optimization algorithm, as is widely used in the mean variance algorithm of establishing stock portfolio.This method is described in detail in 20030194734.In fact, this method requires to establish one group of input value (stock in the financial application, here refer to expression by ionization meter), this input value makes the variability of rreturn value minimize the rreturn value (for example signal of Xing Chenging) that the optimization people use its acceptance simultaneously.A lot of business software programs can be used for carrying out this operation." Wagner associating average-variance optimization is used (Wagner Associates Mean-VarianceOptimization Application) " that be called " Wagner software " in whole instructions is preferred.This software uses the function of " Wagner unites average-variance optimization library (Wagner Associates Mean-Variance OptimizationLibrary) " to measure effective border, and the best of breed on the Markowitz meaning is preferred.Markowitz(1952)。Use such software requirement conversion microarray data, when being used for the financial analysis purpose of its expection with this software of box lunch, data can be handled as input value in the mode of using stock rreturn value and risk measured value.

The method of selection combination also comprises the application of heuristic rule.Preferably, these rules are based on the understanding that biology and being used to produces the technology of clinical effectiveness and formulate.More preferably, they can be used for exporting from optimization method.For example, the mean variance method of combination selection can be used for the microarray data of the gene of a large amount of differential expressions among the cancer experimenter.The output valve that obtains from this method will be one group of optimized gene, and it can comprise the gene that some are expressed in peripheral blood and in the illing tissue.If the sample that is used for method of testing is available from peripheral blood, and in cancer some gene of differential expression also can be in peripheral blood differential expression, so just can applies heuristic rules, wherein select combination from those the efficiency frontier of having got rid of peripheral blood differential expression.Certainly, by for example this rule of utilization in the preselected process of data, can before forming efficiency frontier, use this rule.

Also can use must be not relevant with the biology of being discussed other heuristic rule.For example, people can use the combination of having only specified percentage can be by the rule of the group representative of special genes or gene.The software that commercialization can get as Wagner software, provides the heuristics of these types easily.This also is useful, and for example, the factor beyond accuracy and degree of accuracy (as the license fee of expection) is to the what one hopes or wishes for that comprises one or more genes when influential.

Gene expression overview of the present invention also can be united use with other useful in cancer diagnosis, prognosis or treatment monitoring non-genetic diagnosis method.For example, in some cases, be that the diagnosis capability of above-mentioned method based on gene expression and data from conventional mark such as haemocyanin mark (for example, cancer antigen 27.29 (" CA 27.29 ")) are combined valuably.There are a series of this marks, comprise analyte, as CA27.29.In a kind of such method,, then it is carried out enzyme immunoassay (EIA), to measure one of above-mentioned serum marker from by periodically blood sampling the patient who treats.When the concentration of mark is pointed out tumor recurrence or treatment failure, obtain the sample source that to carry out gene expression analysis.When having suspicious, can carry out fine needle aspiration (FNA), as mentioned above the gene expression of cells overview from the piece collection is analyzed then.Perhaps, can be from gathering tissue sample with the adjacent areas of organizing of previous taking-up tumour.When other test gained result was indeterminate, this method was useful especially.

Kit prepared in accordance with the present invention comprises the format mensuration that is used for determining the gene expression overview.These can comprise measures required all or some material, as the medium of reagent and instructions and mensuration biomarker.

Article of the present invention comprise and being used for the treatment of, diagnosis, prognosis and estimate the representation (representation) of the gene expression overview of disease in other mode.These overview representations are handled the medium that can be read automatically by machine to a kind of, as computer-readable medium (magnetic, optics etc.).Article also can be included in the instruction of estimating the gene expression overview in this medium.For example, article can comprise CD ROM, and it has the computer instruction of the gene expression overview that is used to contrast the said gene combination.Article are recorded in gene expression overview wherein with also can comprising digitizing, so that it can be used for comparing with the gene expression data of patient's sample.Perhaps, the representation record that overview can be different.Graphic recording a kind of form that comes to this.Partek for example above-mentioned, the clustering algorithm of integrating in " DISCOVERY " of Inc. and " INFER " software, the manifesting of these data of assistance that can be best.

Goods of the present invention dissimilar are medium or the formative mensuration that is used to disclose the gene expression overview.These can comprise, for example, microarray, wherein sequence complement or probe stationary on matrix, the combination with it of the sequence of indicating target gene, thus produce readable mensuration about its existence.Perhaps, article of the present invention can be made into kit, produce with hybridization, amplification and the signal that is used to indicate the target gene expression that is used to detect cancer.

Provide the following examples to be used for explaining rather than restriction the present invention.All lists of references that this paper quotes are all incorporated by reference in this article.

Embodiment 1

Material and method

Pancreas cancer marker gene is found.

Use Trizol, from pancreas tumour, normal pancreas, lung, colon, breast and ovary tissue isolation of RNA.Then, RNA is used to produce RNA amplification, mark (Lipshutz etc. (1999)), it is hybridized on Affymetrix U133A array subsequently.Analyze data in 2 kinds of modes then.

In first method, filter this data set, have those genes that have signal (present calls) at least for 2 times between whole data set only to be retained in.This filters remaining 14,547 genes.Record 2,736 genes relative with normal pancreas in the pancreas cancer overexpression, the p value is less than 0.05.45 genes in 2,736 genes are also than at least 2 times of the maximum intensity overexpressions of finding in lung and colon.At last, found 6 probe set, it is than at least 2 times of the maximum intensity overexpressions of finding in lung, colon, breast and ovary tissue.

In the second approach, filter this data set, have those genes that only have signal for 2 times in breast, colon, lung and the ovary tissue only to be retained in.This filters remaining 4,654 genes.Find 160 genes in 4,654 genes, in pancreas tissue (normal and cancer), have at least 2 times and have signal.At last, selected 8 probe set, they show maximum differential and express between pancreas cancer and normal structure.

Tissue sample.

From many commercial vendor, obtaining, totally 260 FFPE shift and former tissue.The sample of test comprises: the transfer of 30 breast, the transfer of 30 colorectums, the transfer of 56 lungs, the transfer of 49 ovarian metastasis, 43 pancreases, former of 18 prostates and 2 prostates shift and 32 other origins (6 stomaches, 6 kidneys, 3 larynxs, 2 livers, 1 oesophagus, 1 pharynx, 1 bile duct, 1 pleura, 3 bladders, 5 melanoma, 3 lymthomas).

RNA extracts.

Based on described method of high pure rna paraffin kit handbook (Roche) and reagent, through following improvement, from paraffin organization slice separation RNA.According to the size (2-5mm=9X10 μ m, 6-8mm=6X10 μ m, 8-〉=10mm=3X10 μ m) of the metastasis of embedding,, and place RNA enzyme/DNA enzyme 1.5ml Eppendorf pipe with paraffin-embedded tissue sample section.The following dewaxing of will cutting into slices: vortex 10-20 was after second, room temperature in 1ml dimethylbenzene incubation 2-5 minute.Centrifuge tube takes out supernatant, and repeats the step that dewaxes then.After taking out supernatant, add 1ml ethanol, sample vortex 1 minute is centrifugal, and take out supernatant.This process is repeated in addition once.Take out remaining ethanol, in 55 ℃ of baking ovens drying precipitated 5-10 minute, and be suspended in 100 μ l again and organize lysis buffer, 16 μ l, 10% SDS and 80 μ l Proteinase Ks.The vortex sample, and in being set at the hot mixed instrument (thermomixer) of 400rpm 55 ℃ of incubations 2 hours.Add 325 μ l binding buffer liquid and 325 μ l ethanol to each sample, mixing is centrifugal then, and supernatant is added on the filtrator post.With filtrator post and collection tube centrifugal 1 minute together at 8000rpm, and abandon and flow through thing (flow through).Carry out a series of continuous washing (500 μ l lavation buffer solution I → 500 μ l lavation buffer solution II → 300 μ l lavation buffer solution II), wherein every kind of solution is added in the post, centrifugal and abandon and flow through thing.Then with post maximum velocity centrifugation 2 minutes, place fresh 1.5ml pipe, and add 90 μ l elution buffers.Room temperature incubation 1 minute, after 8000rpm is centrifugal 1 minute, obtain RNA subsequently.By adding 10 μ l DNA enzyme incubation buffering liquids, 2 μ l DNA enzyme I, and 37 ℃ of incubations 30 minutes, the DNA enzyme was handled sample.After adding 20 μ l organize lysis buffer, 18 μ l, 10% SDS and 40 μ l Proteinase Ks, deactivation DNA enzyme.Once more, 325 μ l binding buffer liquid and 325 μ l ethanol are added in each sample, mixing is centrifugal then, and supernatant is added on the filtrator post.Carry out continuous washing and the wash-out of RNA as mentioned above, exception is to use 50 μ l elution buffer eluted rnas.Pollute in order to eliminate the glass fibre that carries from post,, and supernatant moved into fresh 1.5ml Eppendorf pipe RNA centrifugal 2 minutes at full speed.By the OD260/280 reading that obtains from spectrophotometer, quantitative sample, and with diluted sample to 50ng/ μ l.The RNA that separates is standby in-80 ℃ of water that are deposited in no RNA enzyme.

TaqMan primer and probe design.

Suitably mRNA canonical sequence registration number and Oligo6.0 are used for exploitation

CUP measures (lung mark: HS, lung-associated protein B (HUMPSPBA), thyroid gland transcription factor 1 (TTF1), desmoglein 3 (DSG3), colorectum mark: cadherin 17 (CDH17), breast mark: the ets transcription factor (PDEF) that mammary gland globin (MG), Prostato-are derived, ovary mark: wilms knurl 1 (WT1), pancreas mark: prostate stem cell antigen (PSCA), coagulation factors V (F5), prostate mark kallikrein 3 (KLK3)) and run one's home and measure β actin, Hydroxymethylbilane synthase (PBGD).The primer and the hydrolysis probes of each mensuration in table 2, have been listed.By the mensuration around design extron-introne splice site, get rid of the genomic DNA amplification.Use BHQ1-TT as inner quencher dyestuff, mark hydrolysis probes with FAM as the report dyestuff and at 3 ' nucleotide place at 5 ' nucleotide place.

Quantitative real-time polymerase chain reaction.

On ABI Prism 7900HT sequence detection system (Applied Biosystems), in 384 orifice plates, carry out the quantitative of gene-specific RNA.For each thermal cycler operation, amplify calibrating device and typical curve.The calibrating device of each mark by in from the vector rna of kidney of rats with 1X10 ⁵The target gene in-vitro transcription thing of copy dilution is formed.Run one's home mark typical curve by in from the vector rna of kidney of rats with 1X10 ⁷, 1X10 ⁵And 1X10 ³The target gene in-vitro transcription thing of copy serial dilution is formed.Measure the no target contrast that also comprises in service at each, to guarantee not having environmental pollution.All samples and contrast be operation in duplicate all.Use common laboratory reagent, carry out qRTPCR in 10 μ l reactants, described reactant contains: RT-PCR damping fluid (50nM N-two (hydroxyethyl) glycocoll/KOH pH 8.2,115nM KAc, 8% glycerine, 2.5mM MgCl ₂, 3.5mM MnSO ₄0.5mM every kind of dCTP, dATP, dGTP and dTTP), adjuvant (2mM Tris-Cl pH 8,0.2mM bovine albumin, 150mM trehalose, 0.002% Tween 20), enzymatic mixture (2U Tth (Roche), 0.4mg/ μ l Ab TP6-25), primer and probe mixture (0.2 μ M probe, 0.5 μ M primer).Defer to following loop parameter: 1 95 ℃ of circulations of 1 minute; 1 55 ℃ of circulations of 2 minutes; Oblique ascension (Ramp) 5%; 1 70 ℃ of circulations of 2 minutes; With 40 95 ℃ of 15 seconds, 58 ℃ circulations of 30 seconds.After finishing the PCR reaction, in ABI 7900HT Prism software, set baseline and threshold value, and the Ct value of calculating is outputed among the Microsoft Excel.

One step is to two-step reaction.

The primer of 100ng random hexamer or gene specific is used in each reaction, carries out first chain and synthesizes.In the first step, with 11.5 μ l potpourris-1 (the total RNA of primer and 1ug) be heated to 65 ℃ 5 minutes, then in cooled on ice.With 8.5 μ l potpourris-2 (1x damping fluid, 0.01mMDTT, every kind of dNTP ' s of 0.5mM, 0.25U/ μ l

10U/ μ l Superscript III) adds potpourri-1,50 ℃ of incubations 60 minutes, then 95 ℃ of incubations 5 minutes.CDNA is deposited in-20 ℃ standby.Carry out the qRTPCR in second step of two-step reaction as mentioned above, loop parameter is as follows: 1 95 ℃ of circulations of 1 minute; 40 95 ℃ of 15 seconds, 58 ℃ circulations of 30 seconds.Accurately as in the previous paragraph, carry out the qRTPCR of single step reaction.The two all carries out one step and two-step reaction on 100ng template (RNA/cDNA).After finishing the PCR reaction, in ABI 7900HT Prism software, set baseline and threshold value, and the Ct value of calculating is outputed among the Microsoft Excel.

The generation of thermal map.

For each sample, by getting the average Ct of each CUP mark, and, calculate Δ Ct (Δ Ct=Ct (CUP mark)-Ct (Ave.HK mark)) from the mean value that average Ct deducts the mark of running one's home.Measure the minimum delta Ct of each tissue of origin tag set (lung, breast, prostate, colon, ovary and pancreas) of each sample.The tissue of origin scoring that will have total minimum delta Ct is 1, and all other tissue of origin scorings are 0.According to pathological diagnosis, the sorting data.With the feasibility data computation Partek Pro that modifies, and produce intensity map.

The result.

The discovery of new pancreas origin tumour and cancerous state mark.

At first, 5 pancreas mark candidate things have been analyzed: prostate stem cell antigen (PSCA), serpin, clade A member 1 (SERPINA1), cytokeratin 7 (KRT7), matrix metalloproteinase 11 (MMP11) and MUC-4 (MUC4) (Varadhachary etc. (2004); Fukushima etc. (2004); Argani etc. (2001); Jones etc. (2004); Prasad etc. (2005); With (2004) such as Moniaux), wherein use dna microarray and one group of 13 ductal pancreatic adenocarcinoma, 5 normal pancreas tissues and 98 samples from breast, colorectum, lung and ovarian neoplasm.Only PSCA shows the moderate sensitivity degree (having detected 6/13 or 46% pancreas tumour) (Fig. 4 A) high specific (91/98 or 93% quilt is correctly differentiated and is non-pancreas origin).On the contrary, KRT7, SERPINA1, MMP11 and MUC4 show respectively and are respectively 66%, 91%, 82% and 81% 38%, 31%, 85% and 31% sensitivity in specificity.These data with shift in 27 pancreases origin and 39 non-pancreases origins shift to the underlined qRTPCR that carries out ten minutes consistently, exception is MMP11, it shows worse sensitivity and specificity in qRTPCR and transfer.In a word, send out microarray data structural and can use qRTPCR to differentiate that FFPE shifts the good indicant as the ability of pancreas origin that freeze suddenly, former, but other mark may be useful to optimum performance with marking.

Because ductal pancreatic adenocarcinoma is from the development of pipe epithelial cell, described pipe epithelial cell only accounts for the little number percent (wherein acinar cells and islet cells are in the great majority) of all pancreatic cells, and because the cancer of pancreas tissue contains the normal adjacent tissue (Prasad etc. (2005) of significant quantity; With (2005) such as Ishikawa), so be difficult to differentiate the pancreas cancer mark (that is, in cancer, raising) that also this organ can be distinguished from organ.For the purposes in the CUP group, such differentiation is essential.First kind of querying method (seeing material and method) returns 6 probe set: coagulation factors V (F5), is similar to protein-bonded albumen FLJ22041 (FKBP10), β 6 integrins (ITGB6), transglutaminase 2 (TGM2), heterogeneous nuclear ribonucleoprotein A0 (HNRP0) and the BAX δ (BAX) of inferring of FK506.Second kind of querying method (seeing material and method) returns 8 probe set: 2 probe set of a kind of agnoprotein (SCD) of mRNA (p73), the MGC:10264 of F5, TGM2, pairing-sample homeodomain transcription factor 1 (PITX1), three-piece albumen isoform mRNA (TRIO), p73H and close protein 18.F5 and TGM2 are present in 2 kinds of Query Results, and as if in the two, F5 is the most promising (Fig. 4 B).

Use FFPE to organize optimization specimen preparation and qRTPCR.

Then, before check mark group performance, use the fixing optimization RNA that organizes to separate and the qRTPCR method.At first, analyzed the Proteinase K incubation time has been decreased to 3 hours effect from 16 hours.To not effect of output.But when using shorter Proteinase K step, some sample shows longer RNA fragment (Fig. 5).For example, when RNA is when separating, in electrophoresis pattern, do not observe difference from 1 age piece (C22).But, when RNA is when separating,, when using shorter protease K digesting, observe more most more high molecular RNA as by the assessment of the peak on the shoulder from 5 age pieces (C23).When other sample of processing, this trend can keep usually, the source organ no matter FFPE shifts.In a word, shorten the protease K digesting time and do not sacrifice RNA output, and may assist separation RNA longer, lower degraded.

Then, 3 kinds of different reverse transcription methods have been contrasted: use the reverse transcription of random hexamer, use qPCR (2 step) subsequently, use the reverse transcription of gene-specific primer, use qPCR (2 step) and use a step qRTPCR of gene-specific primer subsequently.From 11 transfer isolation of RNA, and cross the Ct value (Fig. 6) that 3 kinds of methods contrast beta-actin, HS's protein B (HUMSPB) and thyroid gland transcription factors (TTF).All there is the significant difference of statistics (p＜0.001) in all contrasts.For all 3 genes, use the reverse transcription of random hexamer, use qPCR (reaction of 2 steps) to produce the highest Ct value subsequently, and use the reverse transcription of gene-specific primer, use qPCR (reaction of 2 steps) to produce the 1 step reaction Ct value that (but statistics is significant) is lower slightly subsequently than correspondence.But, use 2 step RTPCR of gene-specific primer to have longer reverse transcription step.When with the HUMSPB of each sample and TTF Ct value during to the beta-actin value standardization of correspondence, the standardized Ct value between 3 kinds of methods does not have difference.In a word, optimization RTPCR reaction conditions can produce lower Ct value, this can the older paraffin mass (Cronin etc. (2004)) of assistant analysis, and uses a step RTPCR reaction of gene-specific primer can produce the comparable Ct value of Ct value with generation in corresponding 2 steps reaction.

The diagnosis performance that CUP qRTPCR measures.

Then, on shifting, 239 FFPE carry out 12qRTPCR reaction (10 marks and 2 housekeeping genes).Be displayed in Table 2 and measured used mark.The lung mark is HS's lung-associated protein B (HUMPSPB), thyroid gland transcription factor 1 (TTF1) and desmoglein 3 (DSG3).The colorectum mark is cadherin 17 (CDH17).The breast mark is the Ets transcription factor (PDEF) that mammary gland globin (MG) and Prostato-are derived.The ovary mark is Wilms knurl 1 (WT1).The pancreas mark is prostate stem cell antigen (PSCA) and coagulation factors V (F5), and the prostate mark is kallikrein 3 (KLK3).Describe about gene, see Table 31.

Table 2. primer and probe sequence, registration number and amplicon length

Target	SEQ ID NO	Sequence (5 '-3 ')	Describe	SEQ ID NO
					SP-B	59	cacagccccgacctttgatga	Forward primer		11
		ggtcccagagcccgtctca	Reverse primer								12
							agctgtccagctgcaaaggaaaagcc	Probe *	13
		cacagccccgacctttgatgagaactcagctgtccagctgcaaaggaaaagc caagtgagacgggctctgggacc	Amplicon	14
					TTF1	60	ccaacccagacccgcgc	Forward primer		15
		cgcccatgccgctcatgttca	Reverse primer								16
							cccgccatctcccgcttcatg	Probe *	17
		ccaacccagacccgcgcttccccgccatctcccgcttcatgggcccggcgagc ggcatgaacatgagcggcatgggcg	Amplicon							18
					DSG3	61	gcagagaaggagaagataactcaa	Forward primer			19
		actccagagattcggtaggtga	Reverse primer							20
							attgccaagattacttcagattacca	Probe	*		21
		gcagagaaggagaagataactcaaaaagaaacccaattgccaagattacttc agattaccaagcaacccagaaaatcacctaccgaatctctggagt	Amplicon	22
					CDH17	62	tccctcggcagtggaagctta	Forward primer		23
		tcctcaaactctgtgtgcctggta	Reverse primer								24
							ccaaaatcaatggtactcatgcccgactg	Probe *	25
		tccctcggcagtggaagcttacaaaacgactgggaagtttccaaaatcaatggt actcatgcccgactgtctaccaggcacacagagtttgagga	Amplicon	26
					MG	63	agttgctgatggtcctcatgc	Forward primer	27
		cacttgtggattgattgtcttgga	Reverse primer	28
							ccctctcccagcactgctacgca	Probe *	28
		agttgctgatggtcctcatgctggcggccctctcccagcactgctacgcaggctct ggctgccccttattggagaatgtgatttccaagacaatcaatccacaagtg	Amplicon							30
					PDEF	64	cgcccacctggacatctgga	Forward primer	31
		cactggtcgaggcacagtagtga	Reverse primer	32
							gtcagcggcctggatgaaagagcgg	Probe *	33
		cgcccacctggacatctggaagtcagcggcctggatgaaagagcggacttca cctggggcgattcactactgtgcctcgaccagtg	Amplicon	34
					WT1	65	gcggagcccaatacagaatacac	Forward primer		35
		cggggctactccaggcaca	Reverse primer	36
							tcagaggcattcaggatgtgcgacg	Probe *	37
		gcggagcccaatacagaatacacacgcacggtgtcttcagaggcattcagga tgtgcgacgtgtgcctggagtagccccg	Amplicon	38
					PSCA	66	ctgttgatggcaggcttggc	Forward primer	39
		ttgctcacctgggctttgca	Reverse primer							40
							gcagccaggcactgccctgct	Probe *	41
		ctgttgatggcaggcttggccctgcagccaggcactgccctgctgtgctactcct gcaaagcccaggtgagcaa	Amplicon	42
					F5	67	tgaagaaatatcctgggattattca	Forward primer	43
		tatgtggtatcttctggaatatcatca	Reverse primer	44
							acaaagggaaacagatattgaagactc	Probe *	45
		tgaagaaatatcctgggattattcagaatttgtacaaagggaaacagatattgaa gactctgatgatattccagaagataccacata	Amplicon	46
					KLK3	68	cccccagtgggtcctcaca	Forward primer	47
		aggatgaaacaagctgtgccga	Reverse primer	48
							caggaacaaaagcgtgatcttgctgg	Probe *	49
		cccccagtgggtcctcacagctgcccactgcatcaggaacaaaagcgtgatct	Amplicon							50

		tgctgggtcggcacagcttgtttcatcct
					β Actin	69	gccctgaggcactcttcca	Forward primer	51
		cggatgtccacgtcacacttca	Reverse primer	52
							cttccttcctgggcatggagtcctg	Probe *	53
		gccctgaggcactcttccagccttccttcctgggcatggagtcctgtggcatccac gaaactaccttcaactccatcatgaagtgtgacgtggacatccg	Amplicon	54
					PBGD	70	ccacacacagcctactttccaa	Forward primer		55
		tacccacgcgaatcactctca	Reverse primer	56
							aacggcaatgcggctgcaacggcggaa	Probe *	57
		ccacacacagcctactttccaagcggagccatgtctggtaacggcaatgcggc tgcaacggcggaagaaaacagcccaaagatgagagtgattcgcgtgggta	Amplicon	58

^*Probe is 5 ' FAM-3 ' BHQ1-TT

Analyze the standardized Ct value in the thermal map, disclosed high specific, the medium specificity of colon, lung and ovary and the lower a little specificity of pancreas mark of breast and prostate mark.The qRTPCR data of combination standardization and calculating are refining, have improved the performance of mark group.Obtain the result from the standardized qRTPCR data and the algorithm of combination, and the accuracy of definite qRTPCR mensuration.

Discuss.

In this embodiment, on primary tumor, use expression profile analysis, the candidate's mark that is used to shift with discriminating based on microarray.Primary tumor can be used to find the fact of the origin tumor marker that shifts, and is consistent with several nearest discoveries.For example, Weigelt and colleague show that the gene expression overview of former hair-cream room tumour keeps in remote the transfer.Weigelt etc. (2003).Italiano and colleague find, and be similar in 80 former colorectum tumours and 80 associated transitions by the EGFR state of IHC assessment.Italiano etc. (2005).5 inconsistencies that show the EGFR state only in 80.Italiano etc. (2005).Backus and colleague have identified the mark of inferring that is used to detect Metastasis in Breast Cancer, wherein use the complete genomic gene expression analysis of breast and other tissue, and confirm mammary gland globin and the transfer of CK 19 to work clinically in 90% sensitivity and the 94% specific detection breast warning lymph node.Backus etc. (2005).

Use that former tissue carry out based on studies confirm that of microarray the specificity and the sensitivity of known mark.As a result, except F5, use underlined all to here research tissue have high specific.Argani etc. (2001; Backus etc. (2005); Cunha etc. (2005); Borgono etc. (2004); McCarthy etc. (2003); Hwang etc. (2004); Fleming etc. (2000); Nakamura etc. (2002); With (1997) such as Khoor.Nearest research uses IHC to determine, PSCA is overexpression in prostate cancer shifts.Lam etc. (2005).Dennis etc. (2002) confirm that also PSCA can be used as pancreas and prostatic origin tumor marker.As shown in this paper, find the strongly expressed of PSCA in some prostata tissue at rna level, but owing in mensuration, comprise PSA, so can separate prostate and pancreas cancer now.A new discovery of this research is to use complementation (with the PSCA complementation) mark of F5 as the pancreas tissue of origin.In microarray data collection (containing former tissue) and qRTPCR data set (containing FFPE shifts), F5 and PSCA complementation (Fig. 4 and table 3).

Table 3 feasibility data

Researchist has in the past used IHC or microarray to generate CUP mensuration.Su etc. (2001); Ramaswamy etc. (2001); With (2004) such as Bloom.Recently, SAGE is combined with little qRTPCR mark group.Dennis (2002); With (2003) such as Buckhaults.This research has been made up first based on the expression profile analysis of microarray and the qRTPCR of group and has been measured.Use the microarray research of former tissue identified some but not all with studied the identical tissue of origin mark of mark that identified by SAGE in the past.Some researchs are verified, have medium consistance between the profile analysis data based on SAGE and dna microarray, and this correlativity is for having the more gene raising of high expression level.(2005) such as van Ruissen; And Kim (2003).For example, Dennis and colleague have identified PSA, MG, PSCA and HUMSPB, and Buckhaults and colleague (Dennis etc. (2002)) have identified PDEF.It is preferred using qRTPCR to carry out CUP mensuration, because it is strong technology, and may have the feature performance benefit better than IHC.Al-Mulla etc. (2005); With (2005) such as Haas.As shown here, by in single step reaction, using gene-specific primer, improved the qRTPCR rules.This has confirmed the application of gene-specific primer in a step qRTPCR reaction that contains the FFPE tissue first.Other researchist has carried out 2 step qRTPCR (synthetic cDNA, qPCR subsequently in a reaction), or has used the gene-specific primer of random hexamer or brachymemma.Abrahamsen etc. (2003); Specht etc. (2001); Godfrey etc. (2000); Cronin etc. (2004); With (2004) such as Mikhitarian.

Embodiment 2

The total RNA of CUP FFPE separates rules

(high purity reagent box catalog number (Cat.No.) 3270289)

Purpose:

From the total RNA of FFPE separate tissue

Program:

The preparation of working solution

1. the Proteinase K in the kit (PK)

Lyophilized products is dissolved in the 4.5ml elution buffer.Aliquot and be preserved in-20 ℃ was stablized 12 months.

PK-4x250mg (catalog number (Cat.No.) 3115852)

Lyophilized products is dissolved in 12.5ml elution buffer (1x TE damping fluid (pH7.4-7).Aliquot and be preserved in-20 ℃.

2. lavation buffer solution I

The 60ml absolute ethyl alcohol is added lavation buffer solution I, in the room temperature preservation.

3. lavation buffer solution II

The 200ml absolute ethyl alcohol is added lavation buffer solution II, in the room temperature preservation.

4.DNA enzyme I

Lyophilized products is dissolved in 400 μ l elution buffers.Aliquot and be preserved in-20 ℃ was stablized 12 months.

With paraffin mass section about 30-45 minute, 12 (12 blocks of x2 pipe=24 pipes)

Should for extracting, handle RNA immediately from the section that piece downcuts

1. on microtome, use of the piece of tissue cutting 6X10 micron thickness section (big or small 3-4x5-10mm) of clean sharp cutter from pruning.

Annotate: new piece-abandon the wax section, up to obtaining histotomy.Preceding 3 histotomies of the piece that uses-abandon

2. the tissue that will cut immediately places 1.5mL microcentrifugal tube (microfuge tube), and covers tight lid, so that moisture minimizes.

3. recommend to get the number of section based on the tumour size shown in the table 4.

Table 4

Dewaxed about 30-45 minute

1. 1.0ml dimethylbenzene is added in each sample, violent vortex 10-20 second, and room temperature incubation 2-5 minute.Centrifugal 2 minutes at full speed.Take out supernatant carefully.

Annotate: if tissue shows floatingly, then extra centrifugal 2 minutes.

2. repeating step 1.

3. full speed is centrifugal 2 minutes.Take out supernatant.

4. adding 1ml absolute ethyl alcohol, and violent vortex 1 minute.Centrifugal 2 minutes at full speed.Take out supernatant.

5. repeating step 4.

6. on paper handkerchief, blot pipe tout court, to remove the ethanol residue.

7. dry tissue precipitates 5-10 minute in 55 ℃ of baking ovens.

Annotate: crucial is, removes ethanol fully, and the finish-drying precipitation, and remaining ethanol can suppress PK digestion.

Annotate:, then heated 20-30 minute in room temperature if PK is at-20C.

RNA extracted about 2.5-3 hour

1. add 100 μ l and organize lysis buffer, 16 μ l10%SDS and 80 μ l Proteinase K working solutions in a tissue precipitation, vortex and shook down at 400rpm several times tout court, 55 ℃ of incubations 2 hours.

2. add 325 μ l binding buffer liquid and 325 μ l absolute ethyl alcohols.By on move down liquid, mix gently.

3. with lysate centrifugal 2 minutes at full speed.

4. combination filter pipe and collection tube (12 pipes) are in lysate supernatant shift-in filtrator.

5. at 8000rpm centrifugal 30 seconds, and abandon and flow through thing.

Annotate: merge RNA with other 2 kinds of tissue precipitation preparations if desired, then can repeating step 4-5.

6. repeat at 8000rpm centrifugal 30 seconds, with device for drying and filtering.

7. add 500 μ l lavation buffer solution I working solutions to post, and, abandon and flow through thing in the centrifugal 15-30 of 8000rpm second.

8. add 500 μ l lavation buffer solution II working solutions.In the centrifugal 15-30 of 8000rpm second, abandon and flow through thing.

9. add 300 μ l lavation buffer solution II working solutions,, abandon and flow through thing in the centrifugal 15-30 of 8000rpm second.

10. the high-purity filtrator of maximum velocity centrifugation 2 minutes.

11. high-purity filter tube is placed fresh 1.5ml pipe, and adds 90 μ l elution buffers.Room temperature incubation 1-2 minute.At 8000rpm centrifugal 1 minute.

DNA enzyme I handled about 1.5 hours

12. in eluate, add 10 μ l 10x DNA enzyme incubation buffering liquids and 1.0 μ l DNA enzyme I working solutions, and mix.In 37 ℃ of incubations 45 minutes (or 2.0 μ l DNA enzymes I30 minute).

Organize lysis buffer 13. add 20 μ l, 18 μ l, 10% SDS and 40 μ l Proteinase K working solutions.Brief vortex.At 55 ℃ of incubations 30 minutes (30-60 minute .).

14. add 325 μ l binding buffer liquid and 325 μ l absolute ethyl alcohols.Mix, and move liquid and go into the fresh high-purity filter tube that has collection tube (12 pipes).

15. centrifugal 30 seconds, and abandon and flow through thing at 8000rpm.

16. repetition centrifugal 30 seconds at 8000rpm is with device for drying and filtering.

17. add 500 μ l lavation buffer solution I working solutions to post, at 8000rpm centrifugal 15 seconds, abandon and flow through thing.

18. add 500 μ l lavation buffer solution II working solutions.At 8000rpm centrifugal 15 seconds, abandon and flow through thing.

19. add 300 μ l lavation buffer solution II working solutions, at 8000rpm centrifugal 15 seconds, abandon and flow through thing.

20. the high-purity filtrator of maximum velocity centrifugation 2 minutes.

21. high-purity filter tube is placed fresh 1.5ml pipe, adds 50 μ l elution buffers; Room temperature incubation 1-2 minute.At 8000rpm centrifugal 1 minute, to collect the RNA of wash-out.

22. centrifugal eluate is 2 minutes at full speed, and supernatant is transferred to new pipe, not disturbance is at the glass fibre of bottom.

23. get 260/280 OD reading, and be diluted to 50ng/ μ l.Be deposited in-80 ℃.

CUP ASR measures rules (ABI 7900)

Purpose: use qRTPCR to determine the tissue of origin of CUP sample

Contrast is provided with:

1. positive control (reference table 5 and Fig. 7 plate are provided with in plate C)

The serial dilution of table 5 IVT-with 5 μ l 1X10 ⁸470 μ l H are gone in dilution ₂O+25 μ l10000 rRNA

The IVT contrast	CE/μl	Sample	Water	Bkgd rRNA
					The B actin	100E+05	50	425	25
CDH17	100E+05	50	425	25
					DSG3	100E+05	50	425	25
F5	100E+05	50	425	25
					Hump	100E+05	50	425	25
MG	100E+05	50	425	25
					PBGD	100E+05	50	425	25
PDEF	100E+05	50	425	25
					PSCA	100E+05	50	425	25
TTF1	100E+05	50	425	25
					WT1	100E+05	50	425	25

1E6. table 5. is with 50,000CE/ μ l rRNA is diluted to 500CE/ μ l-5 μ l50,000CE/ μ l+495 μ l H ₂O

Each strip pipe aliquot 10 μ l (2 plates); Potpourri is placed-80 ℃ standby.

2. typical curve (reference table 6 and Fig. 7 plate are provided with in plate C)

Step 1: the complete setting as shown in table 6 of typical curve.

The IVT contrast	CE/μl	Sample	Water	Bkgd rRNA
					B actin-1	100E+07	50	425	25
B actin-2	100E+06	50	425	25
					B actin-3	100E+05	50	425	25
B actin-4	100E+04	50	425	25
					B actin-5	100E+03	50	425	25
PBGD-1	100E+07	50	425	25
					PBGD-2	100E+06	50	425	25
PBGD-3	100E+05	50	425	25
					PBGD-4	100E+04	50	425	25
PBGD-5	100E+03	50	425	25

Table 7. stock solution-1X10 ⁸IVT.With 50,000CE/ μ l rRNA is diluted to 500CE/ μ l-5 μ l 50,000 CE/ μ l+495 μ l H ₂O

Each strip pipe aliquot 10 μ l (2 plates); Potpourri is placed-80 ℃ standby.

Enzymatic mixture:

1. main potpourri: enzyme (Tth)/antibody (TP6-25) sees Table 7.

Table 7

Reagent	2x
		Enzyme Tth (5U/ μ l)	600.00
Antibody: TP6-25 (1mg/ml)	600.00
		Water	300.00
Amount to	1500.00

Aliquot 500 μ l/ pipe, and freezing at-20 ℃.

CUP master's potpourri:

1.2.5X CUP master's potpourri (table 8-11):

Table 8

ml	The 5x adjuvant	2.5x concentration
			0.50	1M Tris-Cl pH 8	5mM
1.25	The 40mg/ml bovine albumin	500μg/ml
			37.50	1M trehalose stoste	375mM
2.5	20％v Tween 20	0.50％
			7.00	ddH 2O
48.75

Allow reagent to mix fully〉15 minutes

Table 9

ml	The 5x adjuvant	2.5x concentration
			12.50	1M N-two (hydroxyethyl) glycocoll/potassium hydroxide pH8.2	125mM
5.75	The 5M potassium acetate	287.5mM
			20.00	Glycerine (V x D=M-〉19.6x1.26=24.6g)	20％
1.25	The 500mM magnesium chloride	6.25mM
			1.75	The 500mM manganese chloride	8.75mM
5.00	ddH 2O
			46.25

Allow reagent to mix fully〉15 minutes; Merge said mixture to sterile chamber-add following

Table 10

ml	The 5x adjuvant	2.5x concentration
			1.25	100mM dATP	1.25mM
1.25	100mM dCTP	1.25mM
			1.25	100mM dTTP	1.25mM
1.25	100mM dGTP	1.25mM
			100.00

Allow reagent to mix fully〉15 minutes; Aliquot 1.8ml/ pipe, and freezing at-20 ℃

Table 11

Primer/probe	Stoste (μ M)	FC(μM)	μl
				Forward primer	100	10	100.0
Reverse primer	100	10	100.0
				Probe (5 ' FAM/3 ' BHQ1-TT)	100	4	40.0
DI water			760.0
				Amount to			1000.0

Primer and probe mixture:

Aliquot 250 μ l/ pipe, and freezing at-20 ℃

Reaction mixture:

1.CUP main potpourri (CMM): (reference table 12-14 and Fig. 7 plate are provided with in plate A)

Table 12

Reagent	FC	X1(10μl)	450
				2.5x CUP master's potpourri	1X	4.00	1800
ROX	1x	0.20	90
				2x TthAb potpourri	2U	1.00	450
Water		2.3	1035
				Amount to		7.50	3375

Preferably, each plate each run has only 356 reactions: 12 samples use 2 positives and 2 negative controls (4x12=48) of 12 marks (288 reactions, each mark is duplicate)+duplicate (20)+each mark of 10 typical curves contrast

Water is regulated sample volume-4.3 μ l sample MAx; Fully mix

Table 13

Reagent	FC	X1(10μl)	34
				Primer 10 μ M/ probes 4 μ M	0.5μM/0.2μM	0.50	17
CMM	1x	7.50	255
				Amount to		8.00	272

2.ToO mark: fully mix

Table 14

Reagent	FC	X1(10μl)	44
				Primer 10 μ M/ probes 4 μ M	0.5μM/0.2μM	0.50	22
CMM	1x	7.50	330
				Amount to		8.00	352

3. beta-actin and PBGD mark: fully mix

The sample scheme:

Table 15

Sample	Sample ID	Concentration	Water=50ng/ μ the l that adds
				A1
A2
				A3
A4
				A5
A6
				A7
A8
				A9
A10
				A11
A12

1.CUP sample: 12 samples are in 96 orifice plates: A1-A12 (reference table 16 and Fig. 7 plate are provided with in plate B); Aliquot 50 μ l 50ng/ μ l (2 μ l/rxn)

Load plate:

1.384 orifice plate setting: (be provided with reference to figure 7 plates in plate D)

2 μ l samples and 8 μ l CMM are loaded upper plate (sample=50ng/ μ l)

4 μ l samples and 6 μ l CMM are loaded upper plate (sample=25ng/ μ l)

Seal pad, and mark.At 2000rpm centrifugal 1 minute.

ABI 7900HT is provided with: place ABI 7900.Option program " CUP 384 ", and click beginning.

Table 16

Thermal cycle conditions
	95C x 60s
55C x 2m
	Oblique ascension 5%
70C x 2m
	40 following circulations
95C x 15s
	58C x 30s

ROX opens

Analyze data, extract Ct, and insertion algorithm.

Embodiment 3

The CUP algorithm

Based on the tissue of origin of initial selection, the standardized Δ Ct of the actin of HPT, MGB, PDEF, PSA, SP-B, TFF, DSG, WT1, PSCA and F5 value is inserted 6 set.Deduct constant 9.00,11.00,7.50,5.00,10.00,9.50,6.50,8.00,9.00 and 8.00 from each Δ Ct respectively.Then, for each sample, selection is from each the minimum CT value in 6 set (reckling among reckling among reckling, PSA, (SP-B, TFF or the DSG) among the HPT, (MGB or PDEF), WT1 and (PSCA or the F5)), as the representative variable of this group.

Use linear discriminant, these variablees and metastasis site are used for graded samples.(Venables etc. (2002) in R (2.0.1 version) makes up 2 different models from training data, and one is the male sex, and one is the women should to use MASS built-in function ' 1da '.Use then the sex model ' prediction ' function, calculate the posterior probability of each ToO.

The variable that in male sex's model, uses be HPT, PSA, (' SP-B ', ' TFF ', ' DSG3 ') in reckling and metastasis site in the reckling, (' PSCA ', ' F5 ').The metastasis site classification has with colon, lung, ovary and all organizes corresponding 4 levels.With regard to women's model, variable be reckling among HPT, (' the MGB ', ' PDEF '), (' SP-B ', ' TFF ', ' DSG3 ') in reckling and metastasis site among reckling, WT1, (' the PSCA ', ' F5 ').

Example R code:

#Training the male model

dat.m<-CUP2.MIN.NORM[，c

(′HPT′，′PSA′，′SP.B.TTF.DSG3′，′PSCA.F5′，′Class′，′background′)]

CUP.lda.m<-lda(Class-.，dat.m，prior＝c(0，0.09，0，23，0.43，0，0.16，0.02)/sum(c(0，0.09，0.23，0.43，0，0.16，0.02)))

#Training the female model

dat.f<-CUP2.MIN.NORM[，c

(′HPT′，′MFB.PDEF′，′SP.B.TTF.DSG3′，′WT1′，′PSCA.F5′，′Class′，′background′)]CUP.lda.f<-1da(Class～.，dat.f，prior＝c(D.03，0.09，0.23，0.43，0.04，0.16，0)/sum(c(0.03，0.09，0.23，0.43，0.04，0.16，0)))

#if unknownsample(i)is_male

predict(CUP.lda.m，CUP2.MIN.NORM.TEST[i，J)

#if unknown sample(i)is female

predict(CUP.lda.f，CUP2.MIN.NORM.TEST[i，l)

In order to move this code, the Frame that is called CUP2.MIN.NORM need contain training data, and it has the minimum value that aforesaid each tissue of origin set is calculated.

The classification correspondence tissue of origin, and the background correspondence above-mentioned metastasis site.

Test figure can be included among the CUP2.MIN.NORM.TEST, and uses anticipation function, can test the specific sample at the i place of being expert at.In addition, test figure must be the form identical with the training set, and has the minimum value adjusting that also is applied to it.

Embodiment 4

The sample that CUP differentiates

Contrast 48 CUP that differentiate with unresolved sample, to determine the correlativity with true CUP sample.The method of using is in those methods described in the embodiment 1-3.The result who obtains is shown in table 17.Tested 11 samples of unresolved CUP, 8 samples are diagnosed, 3 belong to other classification.

Table 17

Embodiment 5

CUP measures boundary

Fig. 8 has described the result who uses the described method of embodiment 1-3 to obtain, with the boundary of determining that CUP measures.Tested the mensuration performance of certain RNA concentration range, and found that it is effective that CUP is determined in the 100-12.5ng RNA scope.

Embodiment 6

QRTPCR measures

Material and method.The freezing tissue sample that is used for microarray analysis.Totally 700 former freezing people's tissues are used for gene expression microarray profile analysis.Obtain sample from many academic institutions, comprise University of Washington (St.Louis, MO), Erasmus medical center (Rotterdam, Holland) and storehouse company of establishment, comprise Genomics Collaborative, Inc (Cambridge, MA), Asterand (Detroit, MI), Oncomatrix (LaJolla, CA) and ClinomicsBiosciences (Pittsfield, MA).For each sample, also collect patient demographics, clinical and pathology information.Comment on the histopathology feature of each sample, to confirm diagnosis and sample estimates preservation and tumour content.

RNA extracts and Affymetrix GeneChip hybridization.To contain freezing cancer sample greater than 70% tumour cell, optimum and normal specimens is dissected, and with mechanical homogenizer (UltraTurrex T8, Germany) in Trizol reagent (Invitrogen, Carlsbad, CA) middle homogenate.By following standard Trizol rules from the freezing tissue isolation of RNA (Invitrogen, Carlsbad, CA), homogenization tissue in Trizol reagent.After centrifugal, collect the top layer liquid phase, and with the total RNA of-20 ℃ isopropanol precipitating.With 75% ethanol washing RNA precipitation, water-soluble, and be deposited in-80 ℃ standby.

(AgilentTechnologies, Palo Alto CA), check the quality of RNA with Agilent 2100 biological analyser RNA 6000 Nano Assay.According to manufacturer's rules of standard, prepare the cRNA of mark, and (Affymetrix, Santa Clara are hybridized CA) with the high density oligonucleotide array Hu133A genetic chip that contains 22,000 probes set altogether.Use Affymetrix rules and scanner, scanning array.For analysis subsequently, gene is separately regarded in each probe set as.Use Affymetrix genetic chip analysis software MAS 5.0, calculate each expression of gene value.All chips satisfy 3 quality control standards: the number percent of array " existence " signal is greater than 35%, and conversion factor is less than 12 when total target strength of being converted to 600, and the average background level is less than 150.

The mark candidate thing is selected.Selection about tissue of origin (ToO) the mark candidate thing of lung, colon, breast, ovary and prostata tissue, in covering totally 682 RNA samples from breast, colon, lung, ovary, prostatic normal, optimum and cancerous tissue, the expression of measuring probe set.Based on the number of statistics inquiry, select tissue-specific mark candidate thing.

In order to produce the pancreas material standed for, use the gene expression overview of 13 former ductal pancreatic adenocarcinomas, 5 normal pancreases and 98 lungs, colon, breast and oophoroma samples to select the cancer of pancreas mark.Carry out 2 kinds of inquiries.In first kind of inquiry, generate the data set that contains 14547 genes, described gene has " existence " signal at least 2 times in the pancreas sample.Identify by T-check (p＜0.05) and be accredited as and normal phase totally 2736 genes than overexpression in the pancreas cancer.The minimum expression that is chosen in the pancreas cancer of 11 hundredths is at least 2 times a gene of the mxm. in colon and lung cancer, thereby produces 45 probes set.As final step, selecting high expressed is 6 genes of at least 2 times of the high expressed in colon, lung, breast and oophoroma.In second kind of inquiry, generate the data set that contains 4654 probe set, described probe sets is combined in maximum 2 " existence " signals in all breast, colon, lung and the ovary sample.Be chosen in totally 160 genes that have at least 2 " existence " signals in the normal and cancer sample of pancreas.In 160 genes,, select 10 genes contrasting them behind the expression between pancreas and the normal structure.The result who merges 2 pancreas inquiries.

Except the gene expression profile analysis, from document, select the minority mark.The result who merges all inquiries tabulates with the short of ToO mark candidate thing that produces every kind of types of organization.Estimate the sensitivity and the specificity of every kind of mark.To show the mark of the ability of the best origin dividing tissue that passes through them, recommended being used for based on the redundant and complementary RT-PCR check of mark,

The FFPE metastatic carcinoma of known origin and CUP tissue.Obtain totally 386 FFPE metastatic carcinomas (III-IV phase) and former gland cancer of 24FFPE prostate of known origin from many commercial vendor, comprise Proteogenex (Los Angeles, CA), Genomics Collaborative, Inc. (Cambridge, MA), Asterand (Detroit, MI), Ardais (Lexington, MA) and Oncomatrix (La Jolla, CA).(Albany NY) obtains known former the independent of 48 metastatic carcinomas with the CUP tissue and gathers from Albany medical college.For each sample, collect patient demographics, clinical and pathology information.Comment on the histopathology feature of each sample, preserve and tumour content to confirm diagnosis and sample estimates.With regard to shifting sample, the Histological evaluation based on patient's clinical history and metastatic carcinoma are compared with corresponding primary carcinoma establishes the diagnosis and the ToO of metastatic carcinoma clearly.

From FFPE sample separation RNA.As being described in high pure rna paraffin kit handbook (Roche), has following improvement from paraffin organization slice separation RNA.According to the size of the metastasis of embedding, with paraffin-embedded tissue sample section (2-5mm=9X10 μ m, 6-8mm=6X10 μ m, 8-〉=10mm=3X10 μ m).As described in the kit handbook, the dewaxing of will cutting into slices, dry tissue precipitation is 5-10 minute in 55 ℃ of baking ovens, and is suspended in 100 μ l again and organizes lysis buffer, 16 μ l10%SDS and 80 μ l Proteinase Ks.The vortex sample, and in being set at the hot mixed instrument of 400rpm 55 ℃ of incubations 2 hours.According to high pure rna paraffin kit handbook, carry out sample processing subsequently.By OD 260/280 reading that obtains from spectrophotometer, quantitative sample, and with diluted sample to 50ng/ μ l.The RNA that separates is standby in-80 ℃ of water that are deposited in no RNA enzyme.

The qRTPCR that is used for mark candidate thing prescreen.(Invitrogen, Carlsbad CA), use Superscript II reverse transcriptase, with the 1 μ g total RNA of random hexamer reverse transcription from each sample according to manufacturer's instructions.Use Primer Express software (Applied Biosystems, Foster City, CA), the genetic marker material standed for of design test and the primer of crt gene ACTB and MGB-probe, perhaps use ABI Assay-on-Demand (Applied Biosystems, Foster City, CA).Test the primer of all inside (in house) designs and probe greater than 90% best amplification efficiency.Carry out the RT-PCR amplification in the 20ml reaction mixture, described reaction mixture contains 200ng template cDNA, 2x Universal PC R master potpourri (10ml) (Applied Biosystems, FosterCity, CA), 500nM forward and reverse primer and 250nM probe.(Applied Biosystems, Foster City react on CA) at ABI PRISM7900HT sequence detection system.Cycling condition is: AmpErase UNG activate 50 ℃ 2 minutes, polymerase activates 95 ℃ of 10 minutes and 50 95 ℃ of 15 seconds and circulations of 60 seconds of annealing temperature (60 ℃).In each is measured, for target gene and crt gene comprise " no template " contrast and template cDNA in duplicate.The relative expression of each target gene is expressed as Δ Ct, and it equals the Ct that target gene Ct deducts crt gene (ACTB).

An optimized step qRTPCR.Suitably mRNA canonical sequence registration number is used for exploitation in conjunction with Oligo6.0

CUP measures (lung mark: HS, lung-associated protein B (HUMPSPBA), thyroid gland transcription factor 1 (TTF1), desmoglein 3 (DSG3), colorectum mark: cadherin 17 (CDH17), breast mark: mammary gland globin (MG), the ets transcription factor (PDEF) that Prostato-is derived, ovary mark: wilms knurl 1 (WT1), pancreas mark: prostate stem cell antigen (PSCA), coagulation factors V (F5), prostate mark kallikrein 3 (KLK3)) and run one's home and measure β actin (beta-actin), Hydroxymethylbilane synthase (PBGD).The primer and the hydrolysis probes (SEQ ID NO:11-58) of the gene specific of an optimized step qRT-PCR mensuration in table 2, have been listed.By the mensuration around design extron-introne splice site, get rid of the genomic DNA amplification.Use BHQ1-TT as inner quencher dyestuff, mark hydrolysis probes with FAM as the report dyestuff and at 3 ' nucleotide place at 5 ' nucleotide place.

On ABI Prism 7900HT sequence detection system (Applied Biosystems), in 384 orifice plates, carry out the quantitative of gene-specific RNA.For each thermal cycler operation, amplify calibrating device and typical curve.The calibrating device of each mark by in from the vector rna of kidney of rats with 1X10 ⁵The target gene in-vitro transcription thing of copy dilution is formed.Run one's home mark typical curve by in from the vector rna of kidney of rats with 1X10 ⁷, 1X10 ⁵And 1X10 ³The target gene in-vitro transcription thing of copy serial dilution is formed.Measure the no target contrast that also comprises in service at each, to guarantee not having environmental pollution.All samples and contrast be operation in duplicate all.Use common laboratory reagent, carry out qRTPCR in 10 μ l reactants, described reactant contains: RT-PCR damping fluid (50nM N-two (hydroxyethyl) glycocoll/KOH pH8.2,115nM KAc, 8% glycerine, 2.5mM MgCl ₂, 3.5mM MnSO ₄0.5mM every kind of dCTP, dATP, dGTP and dTTP), adjuvant (2mM Tris-Cl pH 8,0.2mM bovine albumin, the 150mM trehalose, 0.002%Tween20), enzymatic mixture (2U Tth (Roche), 0.4mg/ μ l Ab TP6-25), primer and probe mixture (0.2 μ M probe, 0.5 μ M primer).Defer to following loop parameter: 1 95 ℃ of circulations of 1 minute; 1 55 ℃ of circulations of 2 minutes; Oblique ascension 5%; 1 70 ℃ of circulations of 2 minutes; With 40 95 ℃ of 15 seconds, 58 ℃ circulations of 30 seconds.After finishing the PCR reaction, in ABI 7900HT Prism software, set baseline and threshold value, and the Ct value of calculating is outputed among the Microsoft Excel.

One step is to two-step reaction.In order to contrast two steps and one step RT-PCR reaction, the primer of 100ng random hexamer or gene specific is used in each reaction, and first chain that carries out two-step reaction is synthetic.In the first step, with 11.5 μ l potpourris-1 (the total RNA of primer and 1ug) be heated to 65 ℃ 5 minutes, then in cooled on ice.With 8.5 μ l potpourris-2 (1x damping fluid, 0.01mMDTT, every kind of dNTP ' s of 0.5mM, 0.25U/ μ l

10U/ μ l Superscript III) adds potpourri-1,50 ℃ of incubations 60 minutes, then 95 ℃ of incubations 5 minutes.CDNA is deposited in-20 ℃ standby.Carry out the qRTPCR in second step of two-step reaction as mentioned above, loop parameter is as follows: 1 95 ℃ of circulations of 1 minute; 40 95 ℃ of 15 seconds, 58 ℃ circulations of 30 seconds.Fully as in the previous paragraph, carry out the qRTPCR of single step reaction.The two all carries out one step and two-step reaction on 100ng template (RNA/cDNA).After finishing the PCR reaction, in ABI 7900HT Prism software, set baseline and threshold value, and the Ct value of calculating is outputed among the Microsoft Excel.

Algorithm development.Use MASS (Venables and the Ripley) built-in function ' lda ' in the R language (2.1.1 version), make up linearity and distinguish the formula function.The model that uses depends on the tissue of extraction metastasis and patient's sex.When running into lung, colon or ovarian metastasis position, for the classification of metastasis site equivalence, the classification prior odds is set at 0.In addition, the prior odds of breast in the male patient and ovary classification is set at 0, and in female patient, the prior odds of prostate classification is set at 0.Other prior odds of all that use in model is of equal value.In addition, the classification of each sample is based on the highest posterior probability of determining by the model of each classification.For the estimation model performance, carry out the leave-one-out cross validation.In addition, with data set randomly in two, be divided into training and test set, keep the proportionate relationship between each classification simultaneously.Repeating this separates 3 times at random.

The result.The purpose of this research is that exploitation is used to predict that the qRTPCR of metastatic carcinoma tissue of origin measures.Experimental work is by 2 most of compositions.First part comprises tissue-specific mark candidate thing recommendation, and they are in the selection (Fig. 9 A) of 10 marks of structural checking of FFPE metastatic carcinoma and mensuration.Second part comprises that qRTPCR measures optimization, measures on another group FFPE metastatic carcinoma subsequently, sets up prediction algorithm, its cross validation and the checking (Fig. 9 B) in the independent sample set.

Sample characteristic.Be used for the gene expression profile analysis and types of organization's specific gene is differentiated from the RNA of totally 700 freezing former tissue samples.Sample comprises 545 primary carcinoma (29 lungs, 13 pancreases, 315 breast, 128 colorectums, 38 prostates, 22 ovaries), 37 benign lesions (1 lung, 4 colorectums, 6 breast, 26 prostates) and 118 (36 lungs, 5 pancreases, 36 colorectums, 14 breast, 3 prostates, 24 ovaries) normal structures.

375 known origin metastatic carcinomas (III-IV phase) and former gland cancer sample of 26 prostates have been used under study for action altogether.Metastatic carcinoma is derived from lung, pancreas, colorectum, ovary, prostate and other cancer." other " sample classification is made up of the metastasis that is derived from lung, pancreas, colon, breast, ovary and prostate tissue in addition.Patient's feature is summarised in the table 18.

Table 18

Shift the set of CUP sample

Sum 401 48

Mean age 57.8 ± 11 ^*62.13 ± 11.7

Sex women 241 20

The male sex 160 28

Tissue of origin

Lung 65 9

Pancreas 63 2

Colorectum 61 4

Breast 63 5

Ovary 82 2

Prostate 27 2

Kidney 88

Stomach 70

Other ^*25 5

Unknown primary carcinoma 11

Histopathologic diagnosis

Gland cancer, medium/as well to break up 306 27

Gland cancer, bad differentiation 49 4

Squamous cell carcinoma 16 5

The cancer 16 10 of bad differentiation

Small cell carcinoma 3

Melanoma 5

Lymthoma 3

Hepatocellular carcinoma 2

Celiothelioma 1

Other ^{* *}14 2

Metastasis site

Lymph node 73 1

Brain 17 14

Lung 20 7

Liver 75 11

Pelvic region (ovary, bladder, fallopian tubal) 53 2

Belly (nethike embrane (nethike embrane, mesenterium, colon, peritonaeum) 91 5

Other (skin, thyroid gland, the wall of the chest, navel) 44 8

Unknown 2

Former (prostate) 26

^*26 patient's age the unknowns

^*Oesophagus, bladder, pleura, liver and gall capsule, bile duct, larynx, pharynx, non-Hodgkin's lymphoma

^{* *}Cellule, celiothelioma, liver cell, melanoma, lymthoma

Sample is divided into 2 set: the checking that is used for tissue-specificity difference expression of verification mark material standed for is gathered (205 sample) and is used to check the training of an optimized step qRTPCR program and training prediction algorithm to gather (260 sample).The set of first 205 samples comprises 25 lungs, 41 pancreases, 31 colorectums, 33 breast, 33 ovaries, 1 prostate, 23 other metastasis of cancer and 18 prostate primary carcinoma.Second set is made up of 260 samples, comprises 56 lungs, 43 pancreases, 30 colorectums, 30 breast, 49 ovaries, 32 other metastasis of cancer and 20 former prostate cancers.64 samples in 2 set comprise 16 lungs, 21 pancreases, 15 other transfers and 12 prostate primary carcinoma, from same patient.

The independent sample set that obtains from Albany medical college comprises 33 CUP samples, and it is former that wherein 22 are proposed, and 15 is known origin metastatic carcinoma.For be proposed former CUP, provide diagnosis based on morphological feature and/or with the result of one group of IHC label check.Patient demographics, clinical and pathological characteristics have been displayed in Table 18.

The mark candidate thing is selected.Analyze the gene expression overview of 5 kinds of former types of organizations (lung, colon, breast, ovary, prostate), cause being used for the recommendation of 13 tissue specificity mark candidate things of qRTPCR check.In the former carcinoma in situ research, identified best material standed for.Argani etc. (2001); Backus etc. (2005); Cunha etc. (2005); Borgono etc. (2004); McCarthy etc. (2003); Hwang etc. (2004); Fleming etc. (2000); Nakamura etc. (2002); With (1997) such as Khoor.Except analyzing microarray data, from document, select 2 marks, comprise complementary squamous cell lung carcinoma mark DSG3 and breast mark PDEF.Backus etc. (2005).Microarray data has confirmed the high sensitivity and the specificity of these marks.

A special method is used to differentiate the specific mark of pancreas.At first, 5 pancreas mark candidate things have been analyzed: prostate stem cell antigen (PSCA), serpin, clade A member 1 (SERPINA1), cytokeratin 7 (KRT7), matrix metalloproteinase 11 (MMP11) and MUC-4 (MUC4) (Varadhachary etc. (2004); Argani etc. (2001); Jones etc. (2004); Prasad etc. (2005); With (2004) such as Moniaux), wherein use dna microarray and one group of 13 ductal pancreatic adenocarcinoma, 5 normal pancreas tissues and 98 samples from breast, colorectum, lung and ovarian neoplasm.Only PSCA shows the moderate sensitivity degree (having detected 6/13 or 46% pancreas tumour) high specific (91/98 or 93% quilt is correctly differentiated and is non-pancreas origin).On the contrary, KRT7, SERPINA1, MMP11 and MUC4 show respectively and are respectively 66%, 91%, 82% and 81% 38%, 31%, 85% and 31% sensitivity in specificity.These data with shift in 27 pancreases origin and 39 non-pancreases origins shift to the underlined qRTPCR that carries out ten minutes consistently, exception is MMP11, it shows worse sensitivity and specificity in qRTPCR and transfer.In a word, send out microarray data structural and can use qRTPCR to differentiate that FFPE shifts the good indicant as the ability of pancreas origin that freeze suddenly, former, but other mark may be useful to optimum performance with marking.

Ductal pancreatic adenocarcinoma is from the development of pipe epithelial cell, and described pipe epithelial cell only accounts for the little number percent (wherein acinar cells and islet cells are in the great majority) of all pancreatic cells in normal pancreas.In addition, the cancer of pancreas tissue contains the normal adjacent tissue of significant quantity.Prasad etc. (2005); With (2005) such as Ishikawa.Therefore, be the gene that ratio in cancer of pancreas raises, enrichment candidate pancreas mark in normal pancreatic cell.First kind of querying method returns 6 probe set: coagulation factors V (F5), is similar to protein-bonded albumen FLJ22041 (FKBP10), β 6 integrins (ITGB6), transglutaminase 2 (TGM2), heterogeneous nuclear ribonucleoprotein A0 (HNRP0) and the BAX δ (BAX) of inferring of FK506.Second kind of querying method (details is seen material and method part) returns 8 probe set: 2 probe set of a kind of agnoprotein (SCD) of mRNA (p73), the MGC:10264 of F5, TGM2, pairing-sample homeodomain transcription factor 1 (PITX1), three-piece albumen isoform mRNA (TRIO), p73H and close protein 18.

Select totally 23 tissue-specific mark candidate things, be used for further carrying out the RT-PCR checking at metastatic carcinoma FFPE tissue by qRT-PCR.The mark candidate thing is tested on 205 FFPE metastatic carcinomas from lung, pancreas, colon, breast, ovary, prostate and prostate primary carcinoma.

Table 19 provides the gene symbol of the tissue specificity mark of selecting for the RT-PCR checking, and has also summed up the result of the check of carrying out with these marks.

Table 19

Based on they cross reactivity, shift low expression or redundancy in the tissue in correspondence, in the mark of 23 tests, get rid of 13.Select 10 marks to be used for last mensuration form.The lung mark is HS's lung-associated protein B (HUMPSPB), thyroid gland transcription factor 1 (TTF1) and desmoglein 3 (DSG3).The pancreas mark is prostate stem cell antigen (PSCA) and coagulation factors V (F5), and the prostate mark is kallikrein 3 (KLK3).The colorectum mark is cadherin 17 (CDH17).The breast mark is the Ets transcription factor (PDEF) that mammary gland globin (MG) and Prostato-are derived.The ovary mark is Wilms knurl 1 (WT1).Average relative expression value in Figure 10, having shown in the different transfer tissues of being marked at of selection.

Use FFPE to organize optimization specimen preparation and qRT-PCR.Then, before check mark group performance, use the fixing optimization RNA that organizes to separate and the qRTPCR method.At first, analyzed the Proteinase K incubation time has been decreased to 3 hours effect from 16 hours.To not effect of output.But, when using shorter Proteinase K step, some sample show longer RNA fragment (Figure 11 A, B).For example, when RNA is when separating, in electrophoresis pattern, do not observe difference from 1 age piece (C22).But, when RNA is when separating,, when using shorter protease K digesting, observe more most more high molecular RNA as by the assessment of the peak on the shoulder from 5 age pieces (C23).When other sample of processing, this trend can keep usually, the source organ no matter FFPE shifts.In a word, shorten the protease K digesting time and do not sacrifice RNA output, and may assist separation RNA longer, lower degraded.

Then, 3 kinds of different reverse transcription methods have been contrasted: use the reverse transcription of random hexamer, use qPCR (2 step) subsequently, use the reverse transcription of gene-specific primer, use qPCR (2 step) and use a step qRTPCR of gene-specific primer subsequently.Shift isolation of RNA from 11, and cross 3 kinds of methods contrast beta-actins, HUMSPB (Figure 11 C, D) and the Ct value of TTF.The result of all contrasts shows the significant difference of statistics (p＜0.001).For 2 genes, use the reverse transcription of random hexamer, use qPCR (reaction of 2 steps) to produce the highest Ct value subsequently, and use the reverse transcription of gene-specific primer, use qPCR (reaction of 2 steps) to produce the 1 step reaction Ct value that (but statistics is significant) is lower slightly subsequently than correspondence.But, use 2 step RTPCR of gene-specific primer to have longer reverse transcription step.When with the HUMSPB Ct value of each sample during to the beta-actin value standardization of correspondence, the standardized Ct value between 3 kinds of methods does not have difference.In a word, optimization RTPCR reaction conditions can produce lower Ct value, this can the older paraffin mass (Cronin etc. (2004)) of assistant analysis, and uses a step RTPCR reaction of gene-specific primer can produce the comparable Ct value of Ct value with generation in corresponding 2 steps reaction.

The diagnosis performance that optimized qRTPCR measures.In the new set of 260FFPE metastasis, carry out 12qRTPCR reaction (10 marks and 2 housekeeping genes).21 samples produce the high Ct value of housekeeping gene, so used only 239 samples in thermal imagine analysis.Analyze the standardized Ct value in the thermal map, disclosed the high specific of breast and prostate mark, the lower a little specificity (Figure 12) of the medium specificity of colon, lung and ovary and pancreas mark.The qRTPCR data of combination standardization and calculating are refining, have improved the performance of mark group.

Use is to the standardized expression values of the mean value of the expression of 2 housekeeping genes, qRTPCR data and algorithm by combination standardization, developed prediction and shifted the algorithm of tissue of origin, and, measured the accuracy that qRTPCR measures by carrying out leave-one-out cross validation check (LOOCV).For 6 types of organizations that comprise in measuring, estimate the number (when the sample error prediction is other tumor type that comprises in mensuration (for example, pancreas being predicted as colon)) of false positive signal respectively and sample is not predicted as the number of times of measuring those types (other) that comprise in the types of organization.The result of LOOCV is table 20 illustrate.

Table 20

Be 204 in the sample of 260 tests, the overall accuracy with 78% has correctly been predicted tissue of origin.Quite the false positive signal of vast scale is because the mark cross reactivity in similarly organizing on the histology.For example, be derived from that 3 kinds of squamous cell transfer cancers of pharynx, larynx and oesophagus are mispredicted to be lung, this is because the DSG3 expression in these tissues.CDH17 (comprises stomach and pancreas) beyond colon GI cancer positive expression causes in the stomach of 6 tests 43 mis-classifications in the pancreas metastasis of cancer with 43 tests to become colon.

Except the LOOCV check, data are divided into 3 independent right training and test set at random.Each separation contains 50% the every class sample of having an appointment.When 50/50 separated, in 3 independent right training and test set, measuring total classification accuracy was 77%, 71% and 75%, thereby has confirmed the mensuration stability.

At last, another that tested the 48FFPE metastatic carcinoma independently gathered, the CUP sample that it comprises known former metastatic carcinoma, the CUP sample of the tissue of origin diagnosis that provides by pathological evaluation (comprising IHC) is provided and keeps CUP in test I HC test back.For every class sample is estimated the tissue of origin prediction accuracy respectively.Table 21 has been summed up measurement result.

Table 21

	Test	Correctly	Accuracy
				Known transfer	15	11	73.3
The CUP that analyzes	22	17	77.3
				The CUP of Fen Xiing not	11

Except only making an exception in a small amount, the tissue of origin prediction is consistent with known former that assesses by clinical/pathological evaluation (comprising IHC) or tissue of origin diagnosis.Be similar to training set, mensuration can not be distinguished the squamous cell carcinoma that is derived from separate sources, and is lung with their error predictions.

Measure and also be 8 tissue of origin diagnosis of making inferring in 11 samples that keep CUP in standard diagnostics test back.One in the CUP case is to make us interested especially.Male patient with prostate cancer history is diagnosed out the metastatic carcinoma in lung and pleura.Shifting the blood-serum P SA test that tissue carries out and the IHC of use PSA antibody is negative, so virologist's diagnosis is the CUP that tends to stomach and intestine tumor.Described mensuration strong (posterior probability 0.99) prediction tissue of origin is a colon.

Discuss.In this research, the expression profile analysis based on microarray on the primary tumor is used to differentiate the candidate's mark that uses into metastasis.Primary tumor can be used to find the fact of the origin tumor marker that shifts, and is consistent with several nearest discoveries.For example, Weigelt and colleague show that the gene expression overview of former hair-cream room tumour keeps in remote the transfer.Weigelt etc. (2003).Backus and colleague have identified the mark of inferring that is used to detect Metastasis in Breast Cancer, wherein use the complete genomic gene expression analysis of breast and other tissue, and confirm mammary gland globin and the transfer of CK19 to work clinically in 90% sensitivity and the 94% specific detection breast warning lymph node.Backus etc. (2005).

In measuring performance history, select to concentrate on 6 cancer types, be included in lung the most general among the CUP, pancreas and colon (Ghosh etc. (2005); With Pavlidis etc. (2005)) and treatment may be to the most useful breast of patient, ovary and prostate.Ghosh etc. (2005).But, other types of organization and mark also can the adding group in, only otherwise the overall accuracy that infringement is measured get final product, and if be suitable for the logic (logistics) that does not hinder RTPCR to react.

Specificity that studies confirm that known mark and sensitivity based on former tissue of the use of microarray.As a result, most of tissue-specific marks have high degree of specificity to the tissue of research here.Nearest use IHC discovers, PSCA is overexpression in prostate cancer shifts.Lam etc. (2005).Dennis etc. (2002) confirm that also PSCA can be used as pancreas and prostatic origin tumor marker.At rna level, in some prostata tissue, there is the strongly expressed of PSCA, still, can separate prostate and pancreas cancer now by PSA owing in mensuration, comprising.The new discovery of this research is to use complementation (with the PSCA complementation) mark of F5 as the pancreas tissue of origin.In microarray data collection (containing former tissue) and qRTPCR data set (containing FFPE shifts), find F5 and PSCA complementation.

Researchist has in the past used IHC (Brown etc. (1997); DeYoung etc. (2000); With (2005a) such as Dennis) or microarray produced CUP mensuration.Su etc. (2001); Ramaswamy etc. (2001); With (2004) such as Bloom.Recently, SAGE is combined with little qRTPCR mark group.Dennis (2002); With (2003) such as Buckhaults.This research has been made up first based on the expression profile analysis of microarray and the qRTPCR of group and has been measured.Use the microarray research of former tissue identified some but not all with studied the identical tissue of origin mark of mark that identified by SAGE in the past.This discovery is not curious, because research is verified, has medium consistance between the profile analysis data based on SAGE and dna microarray, and this correlativity is for having the more gene raising of high expression level.(2005) such as van Ruissen; And Kim (2003).For example, Dennis and colleague have identified PSA, MG, PSCA and HUMSPB, and Buckhaults and colleague (Buckhaults etc. (2003)) have identified PDEF.

It is preferred using qRTPCR to carry out CUP mensuration, because it is strong technology, and may have the feature performance benefit better than IHC.Al-Mulla etc. (2005); With (2005) such as Haas.In addition, as shown here, by in single step reaction, using gene-specific primer, improved the qRTPCR rules.This has confirmed the application of gene-specific primer in a step qRTPCR reaction that contains the FFPE tissue first.Other researchist has carried out 2 step qRTPCR (synthetic cDNA, qPCR subsequently in a reaction), or has used the gene-specific primer of random hexamer or brachymemma.Abrahamsen etc. (2003); Specht etc. (2001); Godfrey etc. (2000); Cronin etc. (2004); With (2004) such as Mikhitarian.

In a word, 78% overall accuracy of the mensuration of 6 types of organizations advantageously is equal to other research.Brown etc. (1997); DeYoung etc. (2000); Dennis etc. (2005a); Su etc. (2001); Ramaswamy etc. (2001); With (2004) such as Bloom.

Embodiment 7

In this research, by from MVO, selecting, and use sorter to predict the tissue origin and the cancerous state of 5 kinds of main cancer types (comprising breast, colon, lung, ovary and prostate), set up the sorter that uses the genetic marker combination.Use Affymetrix people U133AGeneChip, analyzed 378 primary carcinoma, 23 benign proliferative epithelial lesions and 103 normally freeze people's tissue sample suddenly.Also analyzed the leucocyte sample, to deduct by the gene expression of the potential covering of coexpression in the leucocyte background cell.Developed new bioinformatics method, selected the genetic marker combination of tissue of origin and cancerous state based on MVO.Data acknowledgement, the group of 26 genes can be as the sorter of accurately predicting tissue of origin and cancerous state in 5 kinds of cancer types.Thereby the gene expression overview of the genetic marker by measuring appropriate peanut can obtain many cancer classifications method.

Table 22 has shown the mark of differentiating for the tissue origin of appointment.Describe about gene, see Table 31.

Table 22

Tissue	SEQ ID NO：	Title
			Lung	59	SP-B
	60	TTF1
				61	DSG3
Pancreas	66	PSCA
				67	F5
	71	ITGB6
				72	TGM2
	84	HNRPA0
			Colon	85	HPT1
	77	FABP1
				78	CDX1
	79	GUCY2C
			Prostate
	86	PSA
				80	hKLK2
Breast	63	MGB1
				81	PIP
	64	PDEF
			Ovary	82	HE4
	83	PAX8
				65	WT1

The sample set comprises that amounting to 299 shifts colon, breast, pancreas, ovary, prostate, lung and other cancer and former prostate cancer sample.Execution is based on the QC of the expression of Histological evaluation, RNA yield and crt gene beta-actin.Other sample classification comprises the transfer that is derived from stomach (5), kidney (6), bile duct/gall-bladder (4), liver (2), head and neck (4), ileum (1) cancer and a kind of celiothelioma.Table 23 has been summed up the result.

Table 23

Types of organization	Collect	Histology QC	RNA separates QC	ACTB is by QC
					Lung	41	37	36	25
Pancreas	63	57	49	41
					Colon	45	42	42	31
Breast	40	35	35	34
					Ovary	37	36	35	33
Prostate	27	27	25	19
					Other	46	34	29	23
Amount to	299	268	251	205

Test above-mentioned sample, cause tag set is contracted in the table 24 those, the results are shown in Table 25.

Table 24

Table 25

Cancer	Sample #	Mark	Correctly	Sensitivity %	Mistake	Specificity %
							Lung
	25/180	SP-B	13/25	52	0/180	100
									TTF	12/25	48	1/180	99
		DSG3	5/25	20	0/180	100
							Pancreas	41/164	PSCA	24/41	59	6/164	96
		F5	6/41	15	4/164	98
							Colon	31/174	HPT1	22/31	71	2/174	99
Breast	33/172	MGB	23/33	70	3/172	98
									PDEF	16/33	48	1/172	99
Prostate	19/186	PSA	19/19	100	0/186	100
									PDEF	19/19	100	2/186	99
Ovary	33/172	WT1	24/33	71	1/172	99
							Amount to	205

The result shows that in 205 paraffin-embedded metastatic tumours, 166 samples (81%) have conclusive measurement result, table 26.

Table 26

	Material standed for	Correctly	Mistake	No	Accuracy (%)
						Lung	SP-B+TFF+DSG3	19	0	6	76
Pancreas	PSCA+F5	27	1	13	66
						Colon	HPT1		24	2	5	78
Prostate	PSA		19	0	0								100
						Breast	MGB+PDEF	23	3	7	70
Ovary	WT1		23	2	8							70
						Other		20	3		87
Amount to		155	11	39	76

In false positive results, many error source self-organizations and embryology are similarly organized table 27.

Table 27

Sample ID	Diagnosis	Prediction
			OV_26	Ovary	Breast
Br_24	Breast	Colon
			Br_37	Breast	Colon
CRC_25	Colon	Ovary
			Pn_59	Pancreas	Colon
Cont_27	Stomach	Pancreas
			Cont_34	Stomach	Colon
Cont_35	Stomach	Colon
			Cont_43	Bile duct	Pancreas
Cont_44	Bile duct	Pancreas
			Cong_25	Liver	Pancreas

For model development is considered following parameter:

Separate the mark of the women and male sex set, and be respectively the masculinity and femininity patient and calculate the CUP probability.Male sex's set comprises: SP_B, TTF1, DSG3, PSCA, F5, PSA, HPT1; Women's set comprises: SP_B, TTF1, DSG3, PSCA, F5, HPT1, MGB, PDEF, WT1.Getting rid of background from measurement result expresses: lung: SP_B, TTF1, DSG3; Ovary: WT1; And colon: HPT1.

The CUP model is adjusted to CUP morbidity rate (%): lung 23, pancreas 16, colorectum 9, breast 3, ovary 4, prostate 2, other 43.Male patient's breast and ovary morbidity rate are adjusted to 0%, and the prostate morbidity rate of female patient is adjusted to 0%.

Take following step: mark is placed on the similar scale; By selecting minimum value, the number of variable is reduced to 8 from 12 from the set of each tissue specificity; Save (leave out) 1 sample; Set up model from remaining sample; The sample that test is saved; Repeat to and tested 100% sample, save about 50% sample (each organizes about 50%) at random; Set up model from remaining sample; Test about 50% sample; With being separated, 3 different random carry out repetition.

Classify accuracy is adjusted to the cancer types morbidity rate, and to produce the result that table 28 is summed up, raw data is shown in table 29.

Table 28

	Breast	Colon	Lung	Other	Ovary	Pancreas	Prostate	Amount to	Regulate
										Correctly	23	29	22	19	24	35	19	171
Not test	3	2	2		2	3	0	12
										Mistake	7	0	1	4	7	3	0	22
Morbidity rate	0.03	0.09	0.23	0.43	0.04	0.16	0.02
Test/total %	91	94	92	100	94	93	100	94	95
										Correctly/total %	70	94	88	83	73	85	100	89	89
Do not test %	9	6	8	n/a	6	7	0	6	5
Correctly	23	25	19	20	20	24	19	150
										Do not test %	7	6	5		10	15	0	43
Mistake	3	0	1	3	3	2	0	12
										Morbidity rate	0.03	0/09	0.23	0.43	0.04	0.16	0.02
										Test/total %	79	81	80	100	70	63	100	79	83
Correctly/total %	70	81	76	87	61	59	100	73	76
										Correctly/test %	88	100	95	87	87	92	100	93	91
Do not test %	21	19	20	n/a	30	37	0	21	17

Embodiment 8

With the promising gene mark of intervening the not clear metastatic carcinoma CUP in former position that surveys tissue of origin Note research

The specific purposes of this research are, the ability of the tissue of origin of metastatic carcinoma in measuring that the prediction of 10-genetic marker is former and sending out cancer (CUP) patient not clear.

Basic purpose: the feasibility from the genetic analysis of biopsy core sample is carried out in confirmation in the continuous patient of CUP.

Second purpose: the result that 10-genetic marker RT-PCR is measured is associated with the deagnostic test that carries out in M.D.Anderson Cancer center (MDACC).

The 3rd purpose: the morbidity rate of 6 kinds of cancer types that will be by measuring prediction is associated with the morbidity rate of deriving from document and MDACC experience.

Methods described herein are used to carry out 700 freezing primary carcinoma and optimum and microarray gene expression analysis normal specimens, and identify the specific genetic marker material standed for of lung, pancreas, colon, breast, prostate and oophoroma.Be derived from lung, pancreas, colon, breast, ovary and prostatic metastatic carcinoma (III-IV phase) and be derived from other cancer types that is used for specificity contrast transfer 205 formalin fixed, paraffin-embedded (FFPE) sample, by RT-PCR test cdna mark candidate thing.Other metastatic carcinoma type comprises stomach, nephrocyte, liver cell, bile duct/gall-bladder and head and neck cancer.The result allows to select the 10-genetic marker, the tissue of origin of its prediction metastatic carcinoma, and produce 76% overall accuracy.The average CV of the duplicate measurements of RT-PCR experiment is 1.5%, and this calculates and get based on 4 repeat number strong points.Beta-actin (ACTB) is used as housekeeping gene, and its intermediate value is expressed the expression (CV=5.6%) that is similar in the transfer sample of Different Origin.

The specific purposes of this research are that checking 10-genetic marker is compared the ability of metastatic carcinoma tissue of origin among the prediction CUP patient with the comprehensive diagnostic inspection.

Patient's appropriateness

The patient must be at least 18 years old, and ECOG performance state is 0-2.The patient of gland cancer or bad differentiation cancer is diagnosed out in acceptance.Gland cancer patient's group comprises the tumour of good, medium and bad differentiation.

The patient has satisfied the standard of CUP: do not detect former after estimating fully, the described evaluation fully is defined as intrusion research historical completely and physical examination, the inspection of detailed experiments chamber, imaging research and symptom or symptom guidance.Only allow untreated patient to be used for research.

If the patient with chemotherapy or radiation therapy mistake, then then allows to participate in studying as (before the treatment) tissue before the piece that writes down if can obtain in 10 year sections.

The patient provides the written Informed Consent Form/power of attorney that participates in this research.

Research and design

Allow the patient go out CUP after diagnosing under study for action, they have experienced the core pin or the excisional biopsy of accessible metastatic lesion.It is inappropriate only carrying out the bioptic patient of FNA.Preceding 60 patients that occur continuously that comprise standard and agree research are satisfied in registration.If for their diagnostic purpose of treatment, the biopsy that need repeat at MDACC, if the patient agree, then obtain other tissue and be used for research.All participants are registered in the rules of institutional Protocol Data Management System (PDMS).

According to the MDACC standard, the patient of all registrations is carried out complete deagnostic test, comprise clinical and the pathology assessment.The pathology part of deagnostic test may comprise immunohistochemistry (IHC) mensuration of usage flag, and described mark comprises other mark of the appointment that CK-7, CK-20, TTF-1 and virologist think.This is the part of routine inspection that presents all patients of CUP.

Tissue sample is collected

Research comprises the metastatic carcinoma sample of collecting from CUP patient formalin fixed, paraffin-embedded.

6 10 μ m sections are used for RNA and separate, and littler tissue sample will need 9 10 μ m sections.In another section with h and E (HE) dyeing, confirmation is used for the histopathologic diagnosis and the tumour content of each sample of RNA separation.Tumor sample should have the tumour content greater than 30% in the HE section.

Clinical data is offered Veridex anonymously, and comprise patient age, sex, the description of the tumor histology that checks by light microscopy, tumour grade (differentiation), metastasis site, sample collection date, the deagnostic test that carries out about single patient.

Tissue processing and RT-PCR test

Use above-mentioned rules, extract total RNA from each tissue sample.Only will in the tissue of standard volume, produce the sample that surpasses the total RNA of 1 μ m and be used for subsequently RT-PCR test.Sample with lower RNA yield is regarded degraded as, and excludes in experiment subsequently.According to the Veridex program of standard, carry out the RNA integrality contrast of expressing based on running one's home, have the sample of the RNA of degraded with eliminating.

The RT-PCR that comprises the group of 10 genes and 1-2 crt gene measures, and is used to analyze the RNA sample.Use above-mentioned rules, finish reverse transcription and PCR and measure.

Reckoner is shown relative expression's value of gene of each test of Δ Ct, and is used for the tissue of origin prediction, and the Ct that described Δ Ct equals target gene deducts the Ct of crt gene.

Sample size and data interpretation

Owing to the character of seeking and visiting of exploratory development, studied 60 patients' limited sample size.Up to now, after tested 22 patients.1 patient's sample does not produce the RNA of enough RT-PCR tests, and 3 QC contrasts of not assessing by the RT-PCR that uses crt gene.Totally 18 patients are used to measure the probability of patient's metastasis.

The probability that statistical models is used to measure following 7 class metastatic carcinoma tissue of origin: lung, pancreas, colon, breast, prostate, ovary and not test (other).For each sample, from the probability of the every class of linear classification Model Calculation.In table 30, summed up measurement result.

The probability of patient's metastasis that obtains from these 6 positions (colon, pancreas, lung, prostate, ovary, breast) (having known former) is about 76%.This digital source is from the incidence of disease that provides various cancers and propagate the document of potentiality and at the undisclosed data from the M.D.Anderson of tumour registration.For the sample of test, the morbidity rate at 6 positions is 67% (test specimens of 12/18), and this is very consistent with former observation.

Table 30

Although the explanation and the embodiment ratio of purpose have described aforementioned invention in greater detail by being used to understand clearly, described description and embodiment not should be understood to limitation of the scope of the invention.

Table 31

Title	SEQ ID NO	Registration number	Describe
				CDH17	62	NM_004063	Cadherin	17
CDX1	78	NM_001804	Homeobox transcription factor 1
				DSG3	61/3	NM_001944	Desmoglein	3
F5	67/6	NM_000130	Coagulation factors V
				FABP1
	71	NM_001443	Fatty acid binding protein 1, liver
				GUCY2C	79	NM_004963	GUCY2C
HE4	82	NM_006103	The oophoroma mark of inferring
				KLK2	80	BC005196	Kallikrein	2, prostatic
HNRPA0
		84	NM_006805	Heterogeneous nuclear ribonucleoprotein A0
HPT1					85/4	U07969	Intestines peptide-relevant transport protein
	ITGB6
		71	NM_000888	Integrin, β 6
	KLK3				68	NM_001648	Kallikrein	3
MGB1		63/7	NM_002411	Mammary gland globin 1
	PAX8				83	BC001060	Paired box gene 8
PBGD		70	NM_000190	Hydroxymethylbilane synthase
	PDEF				64/8	NM_012391	The domain that contains the Ets transcription factor
PIP		81	NM_002652	The albumen of prolactin-induce
	PSA				86/9	U17040	The antigen precursor of prostate specific
PSCA		66/5	NM_005672	Prostate stem cell antigen
	SP-B				59/1	NM_198843	Curosurf-associated protein B
TGM2		72	NM_004613	Transglutaminase				2
	TTF1				60/2	NM_003317	Be similar to thyroid gland transcription factor 1
WT1		65/10	NM_024426	Wilms knurl				1

Beta-actin	69	NM_001101	Beta-actin
				KRT6F	87	L42612	Keratin	6 isoform K6f
p73H	88	AB010153	The p53-associated protein
				SFTPC	89	NM_003018	Surfactant, lung-associated protein C
KLK10	90	NM_002776	KLK10
				CLDN18	91	NM_016369	Close protein 18
TR10	92	BD280579	Tumor Necrosis Factor Receptors
				B305D	93
B726	94
				GABA-pi	95	BC109105	γ-An Jidingsuan A acceptor, pi
StAR	96	NM_0100724 3	Steroids generates acute regulation protein
				EMX2	97	NM_004098	Air gate homologue 2 (fruit bat (Drosophila))
NGEP	98	AY617079	The long variant of NGEP
				NPY	99	NM_000905	Neuropeptide tyrosine
SERPINA1	100	NM_000295	The serpin peptidase inhibitors, clade A member 1
				KRT7	101	NM_005556	Keratin	7
MMP11	102	NM_005940	Matrix metal peptase 11 (molten stromatin enzyme 3)
				MUC4	103	NM_018406	Be correlated with in MUC-4 cell-surface
FLJ22041
		104	AK025694
BAX					105	NM_138763	The X protein transcript variant Δ that BCL2-is relevant
	PITX1	106	NM_002653	Pairing-sample homeodomain transcription factor 1
MGC： 10264					107	BC005807	Stearyl-coenzyme A desaturase (Δ-9-desaturase)

List of references

U.S. Patent Application Publication and patent

5242974 5700637 20030194733

5350840 5786148 20030198970

5384261 6004755 20030215803

5405783 6136182 20030215835

5412087 6218114 20030219760

5424186 6218122 20030219767

5429807 6225051 20030232350

5436327 6232073 20030235820

5445934 6261766 20040005563

5472672 6271002 20040009154

5527681 6339148 20040009489

5529756 20010029020 20040018969

5532128 20020055627 20040029114

5545531 20020068288 20040076955

5554501 20020168647 20040126808

5556752 20030044859 20040146862

5561071 20030087818 20040219572

5571639 20030104448 20040219575

5593839 20030124128 20050037010

5599695 20030124579 20050059008

5624711 20030138793 20060094035

5658734 20030190656

Open and the patent of foreign patent

WO1998040403 WO2001073032 WO2004030615

WO1998056953 WO2002046467 WO2004031412′

WO2000006589 WO2002073204 WO2004063355

WO2000055320 WO2002101357 WO2004077060

WO2001031342 WO2004018999 WO2005005601

Periodical

Abrahamsen et al.(2003)Towards quantitative mRNA analysis in paraffin-embeddedtissues using real-time reverse transcriptase-polymerase chain reaction J MolDiag 5：34-41

Al-Mulla et al.(2005)BRCA1 gene expression in breast cancer：a correlative studybetween real-time RT-PCR and immunohistochemistry J Histochem Cytochem53：621-629

Argani et al.(2001)Discovery of new Markers of cancer through serial analysis ofgene expression：prostate stem cell antigenis overexpressed in pancreaticadenocarcinoma Cancer Res 61：4320-4324

Autiero et al.(2002)Intragenic amplification and formation of extrachromosomal smallcircular DNA molecules from the PIP gene on chromosome 7 in primary breastcarcinomas Int J Cancer 99：370-377

Backus et al.(2005)Identification and characte rization of optimal gene expressionMarkers for detection of breast cancer metastasis J Mol Diagn 7：327-336

Bentov et al.(2003)The WT1 Wilms′tumor suppressor gene：a novel target forinsulin-like growth factor-l action Endoc rinol 144：4276-4279

Bera et al.(2004)NGEP，a gene encoding a membrane protein detected only inprostate cancer and normal prostate Proc Natl Acad Sci USA 101：3059-3064

Bibikova et al(2004)Quantitative gene expression profiling in formalin-fixed，paraffin-embedded tissues using universal bead arrays Am j Pathol165：1799-1807

Bloom et al.(2004)Multi-platform，multi-site，microarray-based human tumorclassification Am J Pathol 164：9-16

Borchers et al.(1997)Heart-type fatty acid binding protein-involvement in growthinhibition and differentiation Prostaglandins Leukot Essent Fatty Acids 57：77-84

Borgono et al.(2004)Human tissue kallikreins：physiologic roles and applications incancer Mol Cancer Res 2：257-280

Brookes(1999)The essence of SNPs Gene 23：177-186

Brown et al.(1997)Immunohistochemicalidentification of tumor Markers in metastaticadenocarcinoma.A diagnostic adjunct in the determination of primary site Am JClin Pathol 107：12-19

Buckhaults et al.(2003)Identifying tumor origin using a gene expression-basedclassification map Cancer Res 63：4144-4149

Chan et al.(1985)Human liver fatty acid binding protein cDNA and amino acidseq uence.Functional and evolutionaryimplications J Biol Chem260:2629-2632

Chen et al.(1986)Human liver fatty acid binding protein gene is located onchromosome 2 Somat CellMol Genet 12：303-306

Cheung et al.(2003)Detection of the PAX8-PPAR gamma fusion oncogene in bothfollicular thyroid carcinomas and adenomas J Clin Endoc rinol Metab 88：354-357

Clark et al.(1999)The potential role for prolactin-inducible protein(PIP)as a Markerof human breast cancer micrometastasis Br J Cancer 81：1002-1008

Cronin et al.(2004)Measurement of gene expression in archlval paraffin-embeddedtissue Am J Pathol 164：35-42

Cunha et al.(2006)Tissue-specificity of prostate specific antigens：Comparativeanalysis of transc ript levels in prostate and non-prostatic tissues Cancer Lett236：229-238

Dennis et al.(2002)Identification from public data of molecular Markers ofadenocarcinoma characteristic of the site of origin Can Res 62：5999-6005

Dennis et al.(2005a)Hunting the primary：novel strategies for defining the origin oftumors J Pathol 205：236-247

Dennis et al.(2005b)Markers of adenocarcinoma characteristic of the site of origin：development of a diagnostic algo rithm Clin Can Res 11：3766-3772

DeYoung et al.(2000)Immunohistologic evaluation of metastatic carcinomas ofunknown origin：an algorithmic approach Semin Diagn Pathol 17：184-193

Di Palma et al.(2003)The paired domain-containin9 factor Pax8 and thehomeodomain-containing factor TTF-1 directly interact and synergistically activatetranscri ption Biol Chem 278：3395-3402

Dwight et al.(2003)Involvement of the PAX8peroxisome proliferator-activatedreceptor gamma rearrangement in follicular thyroid tumors J Clin Endoc rinolMetab 88：4440-4445

Feldman et al.(2003)PDEF expression in human breast cancer is correiated withinvasive potential and altered gene expression Cancer Res 63：4626-4631

Fleming et al.(2000)Mammaglobin，a breast-specific gene，and its utility as a Markerfor breast cancer Ann N Y Acad Sci 923：78-89

Fukushima et al.(2004)Characterization of gene expression in mucinous cysticneoplasms of the pancreas using oligonucleotide microarrays Oncogene 23：9042-9051

Ghosh et al(2005)Management of patients with metastatic cancer of unknownprimary Curr Probl Surg 42：12-66

Giordano et al.(2001)Organ-specific molecular classification of primary lung，colon，and ovarian adenocarcinomas using gene expression profiles Am JPathol.159：1231-1238

Glasser et al(1988)cDNA，deduced polypeptide structure and chromosomalassignment of human pulmonary surfactant proteolipid，SPL(pVal)J Biol Chem263：9-12

Godfrey et al.(2000)Quantitative mRNA expression analysis from formalin-fixed，paraffin-embedded tissues using 5′nuclease quantitative reversetranscription-polymerase chain reaction J Mol Diag 2：84-91

Goldstein et al.(2002)WT1 Immunoreactivity in uterine papillary serous carcinomasis different from ovarian serous carcinomas Am J Clin Pathol 117：541-545

Gradi et al.(1995)The human steroldogenic acute regulatory(StAR)gene isexpressed in the urogenital system and encodes a mitochondrial polypeptideBiochim Biophys Acta 1258：228-233

Greco et al.(2004)Carcinoma of unknown primary site：sequential treatment withpaclitaxel/carboplatin/etoposide and gemcitabine/irinotecan:A Minnie Pearlcancer research network phase ll trial The Oncologist 9：644-652

Haas et al.(2005)Combined application of RT-PCR and immunohistochemistry onparaffin embedded sentinel lymph nodes of prostate cancer patients Pathol ResPract 200：763-770

Hwang et al.(2004)Wilms tumor gene product：sensitive and contextually specificMarker of serous carcinomas of ovarlan surface epithelial origln ApplImmunohistochem Mol Morphol 12：122-126

lshikawa et al.(2005)Expe rimental trial for diagnosis of pancreatic ductal carcinomabased on gene expression profiles of pancreatic ductal cells Cancer Sci 96：387-393

ltaliano et al.(2005)Epidermal growth factor receptor(EGFR)status in primarycolorectaltumors correlates with EGFR expression in related metastatic sites：biological and clinical implications Ann Oncol 16：1503-1507

Jones et al；(2004)Comprehensive analysis of matrix metalloproteinase and tissueinhibitor expression in pancreatic cancer：increased expression of matrixmetalloproteinase-7 predicts poor survival Clin Cancer Res 10：2832-2845

Jones et al.(2005)Thyroid transc ription factor 1 expression in small cell carcinoma ofthe urinary bladder：an immunohistochemical profile of 44 cases Hum Pathol36：718-723

Khoor et al.(1997)Expression of surfactant protein B precursor and surfactant proteinB mRNA in adenocarclnoma of the lung Mod Pathol 10：62-67

Kim(2003)Comparison of oligonucleotide-microarray and serial analysis of geneexpression(SAGE)in transcript profiling analysis of megakaryocytes derived fromCD34+cells Exp Mol Med 35：460-466

Kim et al.(2003)Steroidogenic acute regulatoryprotein expression in the normalhuman brain and intracranial tumors Brain Res 978：245-249

Lam et al.(2005)Prostate stem cellantigen is overexpressed in prostate cancermetastases Clin Can Res 11：2591-2596

Lembersky et al.(1996)Metastases of unknown primary site Med Clin North Am.80：153-171

Lewis et al.(2001)Unlocking the archive-gene expression in paraffin-embeddedtissue J Pathol 195：66-71

Lipshutz et al.(1999)High density synthetic oligonucleotide arrays Nature Genetics21：S20-24

Lowe et al.(1985)Human liver fattyacid binding protein.Isolation of a fulllengthcDNA and comparative sequence analyses of orthologous and paralogousproteins J Biol Chem 260：3413-3417

Ma et al.(2006)Molecular classification of human cancers using a 92-gene real-timequantitative polymerase chain reaction assay Arch Pathol Lab med 130：465-473

Magklara et al.(2002)Characterization of androgenreceptor and nuclear receptorco-regulator expression in human breast cancer celll ines exhibiting differentialregulation of kallikreins 2 and 3 Int J Cancer 100：507-514

Markowitz(1952)Portfolio Selection J Finance 7：77-91

Marques et al.(2002)Expression of PAX8-PPAR gamma 1 rearrangements in bothfollicular thyroid carcinomas and adenomas J Clin Endocrinol Metab87：3947-3952

Masuda et al.(1999)Analysis of chemical modification of RNA from formalin-fixedsamples and optimization of molecular biology applications for such samples NuclAcids Res 27：4436-4443

McCarthy et al.(2003)Novel Markers of pancreatic adenocarcinoma in fine-needleaspiration：mesothelin and prostate：stem cell antigen labeling increases accuracyin cytologically borde rline cases Appl Immunohistochem Mol Morphol 11：238-243

Mikhita rian et al.(2004)Enhanced detection of RNA from paraffin-embedded tissueusing a panel of truncated gene-specific primers for reverse transcriptionBioTechniques 36：1-4

Mintzer et al.(2004)Cancer of unknown primary：changing approaches，amultidis ciplinary case presentation from the Joan Karnell Cancer Center ofPennsylvania Hospital The Oncologist 9：330-338

Moniaux et al.(2004)Multiple roles of mucins in pancreatic cancer，a lethal andchallenging malignancy Br J Cancer 91：1633-1638

Murphy et al.(1987)lsolation and sequencing of a cDNA clone for a prolactin-inducible protein(PIP).Regulation of PIP gene expression in the human breastcancer cellline，T-47D J Biol Chem 262：15236-15241

Myal et al.(1991)The prolactin-inducible protein(PIPGCDFP-15)gene：cloning，structure and regulation JMol Cell Endocrinol 80：165-175

Nakamura et al.(2002)Expression of thyroid transcription factor-1in normal andneoplastic lung tissues Mod Pathol 15：1058-1067

Noonan et al.(2001)Characterization of the homeodomain gene EMX2：sequenceconservation，expression analysis，and a search for mutations in endomet rialcancers Genomics 76：37-44

Oettgen et al.(2000)PDEF，a novel prostate epithelium-specific Ets transc riptionfactor，interacts with the androgen receptor and activates prostate-specific antigengene expression J Biol Chem 275：1216-1225

Oji et al.(2003)Overexpression of the Wilms′tumor gene WT1 in head and necksquamous cell carcinoma Cancer Sci 94：523-529

Pavlidis et al.(2003)Diagnostic and therapeutic management of cancer of anunknown primary Eur J Can 39：990-2005

Pilot-Mathias et al.(1989)Structure and organization of the gene encoding humanpulmonary surfactant proteolipid SP-B DNA 8：75-86

Pilozzi et al.(2004)CDX1 expression is reduced in colorectal carcinoma and isassociated with promoter hypermethylation J Pathol 204：289-295

Poleev et al.(1992)PAX8，a human paired boxgene：isolation and expression indeveloping thyroid，kidney and Wilms′tumors Development 116：611-623

Prasad et al.(2005)Gene expression profiles in pancreatic intraepithelial neoplasiareflect the effects of Hedgehog signaling on pancreatic ductal epithelial cellsCancer Res 65：1619-1626

Ramaswamy(2004)Translating cancer genomics into clinical oncology N Engl J Med350：1814-1816

Ramaswamy et al.(2001)Multiclass cancer diagnosis using tumor gene expressionsignatures Proc Natl Acad Sci USA 98：15149-15154

Rauscher(1993)The WT1 Wilms tumor gene product：a developmentally regulatedtranscription factor in the kidney that functions as a tumor suppressor FASEB J7：896-903

Reinholz et al.(2005)Evaluation of a panel of tumorMarkers for molecular detectionof circulating cancer cells in women with suspected breast cancer Clin CancerRes 11：3722

Schlag et al.(1994)Cancer of unknown p rimary site Ann Chir Gynaecol 83：8-12

Senoo et al.(1998)A second p53-related protein，p73L，with high homology to p73Biochem Biophys Res Comm 248：603-607

Specht et al.(2001)Quantitative gene expression analysis in microdissected archivalformalin-fixed and paraffin-embedded tumor tissue Amer J Pathol 158：419-429

Su et al.(2001)Molecular classification of human carcinomas by use of geneexpression signatures Cancer Res 61：7388-7393

Takahashi et al.(1995)Cloning and characte rization of multiple human genes andcDNAs encoding highly related typell keratin 6 isoforms J Biol Chem 270：18581-18592

Takamura et al.(2004)Reduced expression of liver-intestine cadherin is associatedwith progression and lymph node metastasis of human colorectal carcinomaCancer Lett 212：253-259

Tothill et al.(2005)An expression-based site of origin diagnostic method designed forclinical application to cancer of unknown origin Can Res 65：4031-4040

van Ruissen et al.(2005)Evaluation of the simila rity of gene expression dataestimated with SAGE and Affymetrix GeneChips BMC Genomics 6：91

Varadhachary et al.(2004)Diagnostic strategies for unknown primary cancer Cancer100：1776-1785

Venables et al.(2002)Modem Applied Statistics with S.Fourth edition.Springer

Wallace et al.(2005)Accurate Molecular detection of non-small cell lung cancermetastases in mediastinal lymph nodes sampled by endoscopicultrasound-guided needle aspiration Cest 127：430-437

Wan et al.(2003)Desmosomal proteins，including desmoglein 3，serve as novelnegative Markers for epidermal stem cell-containing population of keratinocytes JCell Sci 116：4239-4248

Watson et al.(1996)Mammaglobin，a mammary-specific member of the uteroglobingene family，is overexpressed in human breast cancer Cancer Res 56：860-865

Watson et al.(1998)Structure and transc riptional regulation of the humanmammaglobin gene，a breast cancer associated member of the uteroglobin genefamily localized to chromosome 11q13 Oncogene 16：817-824

Weigelt et al.(2003)Gene expression profiles of primary breasttumors maintained indistant metastases Proc Natl Acad Sci USA 100：15901-15905

Zapata-Benavides et al.(2002)Downregulation of Wilms′tumor 1 protein inhibitsbreast cancer proliferation Biochem Biophys Res Commun 295：784-790

Claims (33)

1. differentiate the method for the origin of the not clear metastasis of origin, it comprises following step

A. obtain to contain the sample of transitional cell;

B. measure different Cancer-Related biomarkers with at least 2 kinds;

C. will advance in the algorithm wherein said algorithm from the data combination of biomarker

I. the relative described biomarker of reference standardization; With

Ii adopts cutoff, and the sensitivity and the specificity of every kind of biomarker of its optimization make the morbidity rate weighted sum of cancer select tissue of origin;

D. based on the maximum probability of measuring by algorithm, determine origin, or determine that this cancer is not to be derived from the particular cancer set; With

E. randomly measure the specific biomarker of one or more other different carcinoma, and be other biomarker repeating step c) and d).

2. the process of claim 1 wherein that described marker gene is selected from: from following corresponding group at least a:

I.SP-B, TTF, DSG3, KRT6F, p73H or SFTPC;

Ii.F5, PSCA, ITGB6, KLK10, CLDN18, TR10 or FKBP10; Or

Iii.CDH17, CDX1 or FABP1.

3. the method for claim 2, wherein said marker gene is SP-B, TTF, DSG3, KRT6F, p73H or SFTPC.

4. the method for claim 3, wherein said marker gene is SP-B, TTF and DSG3.

5. the method for claim 4, wherein said marker gene also comprises, or is replaced by, KRT6F, p73H and/or SFTPC.

6. the method for claim 2, wherein said marker gene is F5, PSCA, ITGB6, KLK10, CLDN18, TR10 or FKBP10.

7. the method for claim 6, wherein said marker gene is F5 and PSCA.

8. the method for claim 7, wherein said marker gene also comprises, or is replaced by, ITGB6, KLK10, CLDN18, TR10 and/or FKBP10.

9. the process of claim 1 wherein that described marker gene is CDH17, CDX1 or FABP1.

10. the method for claim 9, wherein said marker gene is CDH17.

11. the method for claim 10, wherein said marker gene also comprises, or is replaced by, CDX1 and/or FABP1.

12. the method for one of claim 1-11 is wherein used at least one among the SEQ ID NO:11-58, measures gene expression.

13. the method for claim 2, wherein said marker gene further is selected from, and is selected from following at least a sex-specific marker

I. under male patient's situation, KLK3, KLK2, NGEP or NPY; Or

Ii. under the situation of female patient, PDEF, MGB, PIP, B305D, B726 or GABA-Pi; And/or WT1, PAX8, STAR or EMX2.

14. the method for claim 13, wherein said marker gene is KLK2.

15. the method for claim 14, wherein said marker gene is KLK3.

16. the method for claim 15, wherein said marker gene also comprises, or is replaced by, NGEP and/or NPY.

17. the method for claim 13, wherein said marker gene are PDEF, MGB, PIP, B305D, B726 or GABA-Pi.

18. the method for claim 17, wherein said marker gene are PDEF and MGB.

19. the method for claim 18, wherein said marker gene also comprises, or is replaced by, PIP, B305D, B726 or GABA-Pi.

20. the method for claim 13, wherein said marker gene are WT1, PAX8, STAR or EMX2.

21. the method for claim 20, wherein said marker gene is WT1.

22. the method for claim 21, wherein said marker gene also comprises, or is replaced by, PAX8, STAR or EMX2.

23. the method for one of claim 13-22 is wherein used at least one among the SEQ ID NO:11-58, measures gene expression.

24. the method for claim 1 or 2 also comprises, and obtains additional clinical information, comprises metastasis site, to determine the origin of cancer.

25. obtain the method for the biomarker set of cancer the best, this method comprises following step: use the metastasis of known origin, measure its biomarker, and contrast the biomarker of described biomarker and the not clear metastasis of origin.

26. the method for orientation treatment is provided, and it is by determining the origin of the not clear metastasis of origin and differentiating that its suitable treatment realizes according to one of claim 1-3.

27. the method for prognosis is provided, and it is by determining the origin of the not clear metastasis of origin and differentiating that its corresponding prognosis realizes according to one of claim 1-3.

28. find the method for biomarker, it comprises, and measures mark expression of gene level in the specific metastasis, the biomarker of measurement markers gene, determining that it expresses, right to analysis requires the expression of 1 marker gene, and determines that whether described marker gene is effectively specific to the origin tumour.

29. composition, it contains the sequence of at least one separation that is selected from SEQ ID NO:11-58.

30. be used to carry out the kit according to the mensuration of one of claim 1-3, it contains biological markers detection reagent.

31. be used to carry out the microarray or the genetic chip of the method for one of claim 1-3.

32. diagnosis/prognosis combination, it comprises the isolated nucleic acid sequences of one of claim 2-11 or the 13-22 described assortment of genes, their complement or its part, wherein said combination be enough to measure or the characterising biological sample in gene expression, with compare from the cell of different carcinoma or normal structure, described biological sample has transitional cell.

33., also comprise at least one expression of gene of measurement constitutive expression in sample according to the method for one of claim 2-11 or 13-22.

Images