WO2019057919A9 - Methods and kits for the prognostic of lung adenocarcinoma - Google Patents

Abstract

The invention relates to the field of cancer prognosis and more particularly, to methods for predicting the outcome of a lung adenocarcinoma patient based on the expression levels of several proteins and to specific prognostic signatures and algorithms, as well as to methods for selecting those patients who would benefit from adjuvant treatment after tumor resection, and kits for implementing those methods. More specifically, said prognostic/predictive signature is based on the protein expression levels of nuclear QKI (QKI-N) and membrane SLC2A1 (SLC2A1-MB) in a biological sample of a subject having lung adenocarcinoma.

Description

Methods and kits for the prognostic of lung adenocarcinoma

FIELD OF THE INVENTION

The invention relates to the field of cancer prognosis and more particularly, to methods for predicting the outcome of a lung adenocarcinoma patient based on the expression levels of several proteins and to specific prognostic signatures and algorithms, as well as to methods for selecting those patients who would benefit from adjuvant treatment after tumor resection, and kits for implementing those methods. More specifically, said prognostic/predictive signature is based on the protein expression levels of nuclear QKI (QKI-N) and membrane SLC2A1 (SLC2A1-MB) in a biological sample of a subject having lung adenocarcinoma.

STATE OF THE ART

Lung cancer remains the leading cause of cancer deaths worldwide with an overall 5 years survival of 10-15%. This low survival rate is mainly associated with the fact that about two thirds of lung cancer patients are diagnosed at advanced stages for which curative treatment is not feasible. These data could change very soon since the National Lung Screening Trial, a randomised study of over 53,000 high-risk individuals randomized between screening with low dose spiral computed tomography (LDCT) versus chest X-ray, has demonstrated that the LDCT screening is truly efficient in detecting more early stage lung cancers. After resection of these early stage tumors, they have shown a 20% relative mortality reduction in the LDCT group respect to the CXR one in seven years of follow up (Aberle DR et al, 2011). These data have led to the implementation of policies of coverage and reimbursement of low-dose CT (LDCT) lung cancer screening by the U.S. Centers for Medicare and Medicaid Services and other Health Providers.

Nowadays, surgical resection represents the best option for a cure of early stage Non-small cell lung cancer (NSCLC) patients (I-II); however no more than 50%> of these cases survive longer than 5 years after operation (Siegel RL et al, 2017). Several randomized clinical trials have demonstrated that adjuvant chemotherapy improved the survival of stage II to IIIA NSCLC patients, but the benefit in stage I is still contentious (Arriagada R et al, 2004; Winton T et al, 2005; Douillard JY et al, 2006; Strauss GM et al, 2008). In stage IA patients, adjuvant chemotherapy is not recommended since the LACE meta-analysis, which combines data from 5 large adjuvant cisplatin-based chemotherapy trials, suggested worse outcome in stage IA patients treated with adjuvant therapy (Pignon JP et al 2008). Moreover, in stage IB patients, adjuvant chemotherapy is recommended only in special circumstances (Pisters KM et al, 2007). All these data suggest that among stage I patients exist a subpopulation with poor prognosis which may potentially benefit from adjuvant treatment, whereas other stage I or II patients might not benefit or even could be harmed by postsurgical therapy. Considering the increasing number of early stage lung cancer patients in the coming years due to the implementation of screening programs worldwide, new information that helps the clinicians in the management of early stage patients after surgery is needed.

Prognostic markers are tumor or patient characteristics that may be measured and evaluated to predict the course of a disease. Nowadays, the tumor-node-metastasis (TNM) staging system is the most powerful method for predicting survival of patients with lung cancer. Many other clinical and molecular markers of potential prognostic significance in NSCLC have been reported in the literature (Coate LE et al, 2009; Ferte C et al 2010; Lin J and Beer DG, 2012; Zhu CQ and Tsao MS, 2014); however, nowadays none of them have been incorporated to the clinical practice .

In the past few years, attention has shifted to genomic technologies, particularly DNA microarray technology, that simultaneously measures mRNA expression levels for tens of thousands of genes. The development of this technology has provided the opportunity to discover groups of genes whose combined expression could predict disease outcome in a more accurate way. However, the complex technology required and the large number of genes contained in these profiles makes very difficult its applicability to clinical practice.

In response to these concerns, several recent efforts have focused on developing quantitative polymerase chain reaction (q-PCR)-based molecular signatures. It is unlikely that every gene in the molecular profiles obtained by microarray analysis has equal relevance with respect to prognosis. Ideally, a small number of genes could predict survival with the same precision as microarray-based gene signatures. These genes could be analyzed by q-PCR, the gold standard assay for gene expression. Q-PCR has significant advantages to microarray-based assays, including widespread availability, cost and reproducibility. Several q-PCR-based molecular signatures with prognostic implications have been described (Lau SK et al, 2007; Bianchi F et al, 2007; Skrzypski M et al 2008; Raz DJ et al 2008).

One of the major logistic problems in order to implement q-PCR-based signatures in daily clinical management is the need of good quality RNA extracted from fresh frozen resected tissue. The different variables of time, preservation protocol and sample transport logistics may have a profound effect in the quality of RNA of a fresh or frozen sample. This becomes an important limiting factor in order to build a robust, reproducible and feasible technical RNA-based tool of molecular classification.

Different studies have tried to overcome these problems using RNA extracted from formalin fixed, paraffin embedded (FFPE) samples, which require less sophistication in the procurement protocol. Kratz et al defined an 11 -gene prognostic profile for non-squamous NSCLC (Kratz JR et al, 2012). Also, Wistuba et al demonstrated the utility of a 31 genes prognostic signature (the cell cycle proliferation score) as predictor of survival in early stage lung adenocarcinoma (ADC) (Wistuba II et al, 2013). Both signatures were commercially available as prognostic tests, the first called Pervenio™ Lung RS tests which stratified non squamous NSCLC patients in two risk groups and the second one the Myriad myPlan™ Lung Cancer test that separated lung ADC patients into three risk groups. Both signatures were validated in independent series but none of them has been demonstrated survival improvement. Both companies have started a prospective randomized clinical trial to risk-stratify stage I non-squamous (in the case of Pervenio) or stage I-IIA ADC (in the case of Myplan) NSCLC for adjuvant chemotherapy. However, only the second one is currently recruiting participants to demonstrate the efficacy of the test. In light of the technical challenges posed by use of mRNA-based technology and the minimal overlap among the genes that comprise the RNA-based prognostic signatures, a prediction approach based on the determination of the expression levels of proteins, for instance detected by immunohistochemistry (IHC), would be more accurate, reproducible and easy to apply to clinical management than methods based on RNA.

Patent application No. CN106442991 discloses a method for predicting the prognosis of patients with lung adenocarcinoma and determining the benefit of adjuvant chemotherapy, which comprises determining the expression levels of a set of protein markers, namely c-Src, Cyclin El, TTF1, p65, CHK1 and JNK1, by immunohistochemical methods.

Despite recent advances, there is still a need to find accurate and reproducible methods which are easy to apply in the clinical practice for predicting the outcome of lung cancer (e.g. lung adenocarcinoma and squamous cell carcinoma), and help clinicians in the selection of the more appropriate treatment strategies. Such prognostic methods are particularly needed in early stages of the disease where an adjuvant treatment may only be recommended for those patients having poor prognosis.

BRIEF DESCRIPTION OF THE INVENTION

The inventors have identified several signatures of expressed proteins with prognostic value in patients with ADC. These signatures comprise determining the protein expression levels of nuclear QKI (QKI-N) and membrane SLC2A1 (SLC2A1-MB), and optionally, further comprise determining the protein expression levels of one, two, three, four or five additional markers selected from the group consisting of: nuclear BRCA1 (BRCA1-N), cytoplasmic QKI (QKI-C), nuclear STC1 (STC1-N), nuclear CDC6 (CDC6-N), and nuclear SIRT2 (SIRT2-N).

The identified signatures preferably consist of an algorithm which combines the coefficient- weighted expression of the selected marker proteins analyzed by immunohistochemical methods. The levels of the expression of each protein are preferably measured by H-Score and the algorithm assigns a coefficient to each protein H-Score to get a final combined score that in the context of this invention is called Prognostic Index (PI) value.

These protein scores and the prognostic indicator have been shown to provide independent prognostic information for recurrence free survival (DFS) and overall survival (OS) in patients with lung ADC. Indeed, as shown in Examples 1 to 3, the identified signatures are able to discriminate with high statistical significance a group of ADC patients of early stages of lung cancer who have higher risk of recurrence or worse survival prospects. This group of patients should be monitored more closely and managed more intensely, and may be potential candidates for adjuvant chemotherapy.

The inventors have also shown the prognostic value of these signatures in combination with the cancer clinico-pathological stage (i.e. using the TNM staging system), see Examples 4 and 5.

Also, the signature which has been identified as having the highest prognostic value for ADCs has been shown to be able to predict which patients would likely benefit from postsurgical therapy, in particular in early stage lung adenocarcinoma patients. Indeed, in Example 6 is shown that in stage I ADC patients presenting higher PI who received platinum-based chemotherapy after surgical resection a statistically significant association was observed with longer overall survival. However this association was not observed in patients with lower PI. Accordingly, in a first aspect the present invention relates to an in vitro method for determining the prognosis of a subject having lung adenocarcinoma, wherein said method comprises: a) determining in a biological sample isolated from said subject the protein expression levels of nuclear QKI (QKI-N) and membrane SLC2A1 (SLC2A1-MB), optionally, further determining in said biological sample the protein expression levels of at least one, two, three, four or five additional markers selected from the group consisting of:

-nuclear BRCA1 (BRCA1-N),

-cytoplasmic QKI (QKI-C),

-nuclear STC1 (STC1-N),

-nuclear CDC6 (CDC6-N), and

-nuclear SIRT2 (SIRT2-N); b) calculating a combined score from the protein expression levels of the markers determined in the biological sample as defined in step a); and c) classifying the subject as having good prognosis or poor prognosis based on the combined score.

In another aspect, the invention relates to a method for determining target protein expression levels by immunohistochemistry, wherein said method is suitable for carrying out the determination of the protein expression levels of the markers in the biological sample as defined in step a) of the prognosis method of the invention. In a particular embodiment, said method comprises incubating the biological sample isolated from said subject with an affinity reagent for QKI, an affinity reagent for SLC2A1, and optionally, an affinity reagent for each of the at least one, two, three, or the four additional proteins of the additional protein markers defined in step a) of the method of prognosis of the invention. Additionally, the invention provides an w vitro method for selecting those subjects having lung adenocarcinoma who are expected to benefit from an adjuvant treatment, wherein said method comprises:

a) classifying said subject according to the prognosis method as defined herein;

b) selecting for administration of an adjuvant treatment a subject classified in step a) as having poor prognosis. The invention also relates to an in vitro method for predicting the efficacy of an adjuvant chemotherapy in a subject having lung adenocarcinoma, wherein said method comprises:

a) classifying said subject according to the prognosis method as defined herein; and b) predicting the efficacy of said adjuvant chemotherapy according to the prognosis classification;

wherein classification of the subject in step a) as having poor prognosis is indicative of increased efficacy of the adjuvant chemotherapy. Moreover, the invention provides an in vitro method for selecting an adjuvant treatment for a subject having lung adenocarcinoma, wherein said method comprises:

a) classifying said subject according to the prognosis method as defined herein; and b) selecting an adjuvant treatment according to the prognosis classification;

wherein when the subject is classified in step a) as having poor prognosis then adjuvant chemotherapy is selected as adjuvant treatment.

In a further aspect, the invention relates to a data-processing apparatus comprising means for carrying out one or more of the steps of the methods as described herein. In another aspect, the invention pertains to a computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out one or more of the steps of the methods of the invention. The invention also refers to a computer-readable storage medium having stored thereon such a computer program. In an additional aspect, the present invention provides a method for treating a subject having lung adenocarcinoma comprising administering to said patient a therapeutically effective amount of a treatment, typically an adjuvant treatment, wherein said treatment is selected according to the classification of said patient according to the prognosis method of the invention. Preferably, when said patient is classified as having poor prognosis the administered treatment comprises a platinum anticancer agent.

In another further aspect, the invention relates to a kit suitable for use in a method for the prognosis of a subject having lung adenocarcinoma as defined herein, wherein said kit comprises: - an affinity reagent for QKI; and

- an affinity reagent for SLC2A1;

- optionally, further comprising an affinity reagent for each of the at least one, two, three or the four additional proteins of the additional protein markers defined in step a);

- optionally, further comprising cancer cells to be used as low and/or high expression controls;

- optionally, further comprising instructions for the use of said reagents in determining said protein expression levels in a biological sample isolated from a subject.

In an additional further aspect, the invention relates to the use of a kit in a method for the prognosis of a subject having lung adenocarcinoma as defined herein, wherein said kit comprises:

- a reagent for determining the protein expression levels of QKI; and

- a reagent for determining the protein expression levels of SLC2A1 ;

- optionally, further comprising a reagent for each of the at least one, two, three, or the four additional proteins of the additional protein markers defined in step a) for determining the protein expression levels thereof;

- optionally, further comprising cancer cells to be used as low and/or high expression controls;

- optionally, further comprising instructions for the use of said reagents in determining said proteins expression levels in a biological sample isolated from a subject.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1. Study of the specificity of QKI antibody. A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη.

Figure 2. Study of the specificity of RAE1 antibody. A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη. Figure 3. Study of the specificity of RRM2 antibody. A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη.

Figure 4. Study of the specificity of SLC2A1 antibody. A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη.

Figure 5. Study of the specificity of SRSF1 antibody. A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη

Figure 6. Study of the specificity of STC1 antibody. A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη.

Figure 7. Study of the specificity of LIG1 antibody. A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη.

Figure 8. Study of the specificity of SIRT2 antibody. A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη.

Figure 9. Study of the specificity of CDC6 antibody. A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη. Figure 10. Study of the specificity of RAD51 antibody . A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη.

Figure 11. Study of the specificity of SNRPE antibody . A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη.

Figure 12. Study of the specificity of BRCA1 antibody. A. WB of total and fraction extracts (C: cytoplasmic and N nuclear). B. Relative expression after densitometry quantification of the bands, standardized to GAPDH C. Representative images of signal detection by ICQ in the cell lines. Scale bar: 20μιη. D Images obtained for BRCA1 immunostaining in A549 cell line after siRNA downregulation. Scale bar 20μιη.

Figure 13. Association between PI score obtained from ADC model 1 and survival in ADC patients from a MD Anderson Cancer Center (MDA) cohort. A-B. Kaplan-Meier curves and log rank statistics for disease free survival (A) and overall survival (B) in ADC patients from the training set (MDA cohort) divided in high and low prognostic index (PI) scores.

Figure 14. Association between PI score obtained from ADC model 2 and survival in ADC patients from MDAcohort. A-B. Kaplan-Meier curves and log rank statistics for disease free survival (A) and overall survival (B) in ADC patients from the training set (MDA cohort) divided in high and low prognostic index (PI) scores.

Figure 15. Association between PI score obtained from ADC model 3 and survival in ADC patients from MDA cohort. A-B. Kaplan-Meier curves and log rank statistics for disease free survival (A) and overall survival (B) in ADC patients from the training set (MDA cohort) divided in high and low prognostic index (PI) scores.

Figure 16. Association between PI score obtained from ADC model 4 and survival in ADC patients from MDA cohort. A-B. Kaplan-Meier curves and log rank statistics for disease free survival (A) and overall survival (B) in ADC patients from the training set (MDA cohort) divided in high and low prognostic index (PI) scores.

Figure 17. Association between PI score obtained from SCC model 1 and survival in SCC patients from MDA. A-B. Kaplan-Meier curves and log rank statistics for disease free survival (A) and overall survival (B) in SCC patients from the training set (MDA cohort) divided in high and low prognostic index (PI) scores.

Figure 18. Association between PI score obtained from SCC model 2 and survival in SCC patients from MDA. A-B. Kaplan-Meier curves and log rank statistics for disease free survival (A) and overall survival (B) in SCC patients from the training set (MDA cohort) divided in high and low prognostic index (PI) scores.

Figure 19. Association between PI score obtained from SCC model 3 and survival in SCC patients from MDA. A-B. Kaplan-Meier curves and log rank statistics for disease free survival (A) and overall survival (B) in SCC patients from the training set (MDA cohort) divided in high and low prognostic index (PI) scores.

Figure 20. Association between PI score obtained from SCC model 4 and survival in SCC patients from MDA. A-B. Kaplan-Meier curves and log rank statistics for disease free survival (A) and overall survival (B) in SCC patients from the training set (MDA cohort) divided in high and low prognostic index (PI) scores.

Figure 21. Association between PI score obtained from SCC model 5 and survival in SCC patients from MDA. A-B. Kaplan-Meier curves and log rank statistics for disease free survival (A) and overall survival (B) in SCC patients from the training set (MDA cohort) divided in high and low prognostic index (PI) scores.

Figure 22. ADC model 1 validation in an independent cohort of patients (CIBERES-CUN). Kaplan-Meier curve and log rank statistics for overall survival (A) and disease free survival (B) in ADC patients from the CIBERES-CUN (A) and CUN (B) cohort divided in high and low molecular prognostic index (PI) scores.

Figure 23. SCC model 1 validation in an independent cohort of patients (CIBERES-CUN) Kaplan-Meier curve and log rank statistics for overall survival (A) and disease free survival (B) in SCC patients from the CIBERES-CUN (A) and CUN (B) cohort divided in high and low molecular prognostic index (PI) scores.

Figure 24. Association between PI score and survival in stage I-II ADC patients from MDA . A-B. Kaplan-Meier curves and log rank statistics for disease free survival (A) and overall survival (B) in stage I-II ADC patients from the training set (MDA cohort) divided in high and low prognostic index (PI) scores.

Figure 25. Association between PI score and survival in stage I-II SCC patients from MDA. A-B. Kaplan-Meier curves and log rank statistics for disease free survival (A) and overall survival (B) in stage I-II SCC patients from the training set (MDA cohort) divided in high and low prognostic index (PI) scores.

Figure 26. ADC model validation in an independent cohort of stage I-II patients (CIBERES-CUN). Kaplan-Meier curve and log rank statistics for overall survival (A) and disease free survival (B) in stage I-II ADC patients from the CIBERES-CUN (A) and CUN (B) cohort divided in high and low molecular prognostic index (PI) scores.

Figure 27. SCC model validation in an independent cohort of stage I-II patients (CIBERES-CUN) Kaplan-Meier curve and log rank statistics for overall survival (A) and disease free survival (B) in stage I-II SCC patients from the CIBERES-CUN (A) and CUN (B) cohort divided in high and low molecular prognostic index (PI) scores.

Figure 28. Assessment of the clinical utility of the ADC prognostic signature. A. Prognostic value of the PI score and stage in univariate and bivariate models for ADC patients of MDA cohort. B-C. Kaplan-Meier curve and log rank statistics for disease-free survival (B) and overall survival (C) in ADC patients from the training set (MDAcohort) divided in high and low combined prognostic index (CPI) scores. Figure 29. Assessment of the clinical utility of the SCC prognostic signature. A. Prognostic value of the PI score and stage in univariate and bivariate models for SCC patients of MDA cohort. B-C. Kaplan-Meier curve and log rank statistics for disease-free survival (B) and overall survival (C) in SCC patients from the training set (MDA cohort) divided in high and low combined prognostic index (CPI) scores. Figure 30. CPI validation for ADC in the CIBERES-CUN cohort. A. Kaplan-Meier curve and log rank statistics for overall survival in ADC patients from the validation set (CIBERES- CUN cohort) divided in high and low combined prognostic index (CPI) scores. B. Kaplan-Meier curve and log rank statistics for disease free survival in patients from the CUN cohort divided in high and low combined prognostic index (CPI) scores

Figure 31. CPI validation for SCC in the CIBERES-CUN cohort. A. Kaplan-Meier curve and log rank statistics for disease-free survival and overall survival in SCC patients from the validation set (CIBERES-CUN cohort) divided in high and low combined prognostic index (CPI) scores. B. Kaplan-Meier curve and log rank statistics for disease free survival in patients from the CUN cohort divided in high and low combined prognostic index (CPI) scores

Figure 32. The ADC model is predictive of differential benefit in survival for adjuvant post-operative therapy versus resection alone . Kaplan-Meier curve and log rank statistics for overall survival (A) and disease-free survival (B) in ADC patients from the training set (MD Anderson cohort) after stratification of patients according to molecular prognostic index (PI) and status of adjuvant therapy receiver.

Figure 33. The SCC model is predictive of differential benefit in survival for adjuvant post- operative therapy versus resection alone Kaplan-Meier curve and log rank statistics for overall survival (A) and disease-free survival (B) in SCC patients from the training set (MD Anderson cohort) after stratification of patients according to molecular prognostic index (PI) and status of adjuvant therapy receiver. DETAILED DESCRIPTION OF THE INVENTION

Definitions

The terms "subject", or "individual"' are used herein interchangeably to refer to all the animals classified as mammals and includes but is not limited to domestic and farm animals, primates and humans, for example, human beings, non-human primates, cows, horses, pigs, sheep, goats, dogs, cats, or rodents. Preferably, the subject is a male or female human being of any age or race.

The term "subject suspected of having cancer" as used herein, refers to a subject that presents one or more signs or symptoms indicative of a cancer and is being screened for cancer. A subject suspected of having cancer encompasses for instance an individual who has received a preliminary diagnosis (e.g., an X-ray computed tomography scan showing a mass) but for whom a confirmatory test (e.g., biopsy and/or histology) has not been done or for whom the stage of cancer is not known or not known in enough detail.

The term "prognosis" as used herein refers to predicting disease progression or outcome.

More specifically, "prognostic markers" may refer to patient or tumor characteristics that predict outcome (usually survival) independent of the treatment. Thus, they are usually identified and validated in patients who receive no therapy or surgical therapy only. The goal of identifying prognostic markers is to define patient subpopulations with significantly different anticipated outcomes, which might benefit from different therapies. Good prognostic patients may not require additional treatment beyond the primary surgical resection, while poor prognostic patients may derive improved survival benefit from adjuvant therapy or other closer clinical follow up or therapeutic strategy.

"Predictive markers", on the other hand, may refer to patient or tumor characteristics that predict benefit from specific treatments (either in terms of tumor shrinkage or survival). In other words, the differences in tumor response or survival benefit between treated versus untreated patients will be significantly different in those positive or negative for the predictive marker (Zhu CQ and Tsao MS, 2014).

The term "therapeutically effective amount" as used herein refers to an amount that is effective, upon single or multiple dose administration to a subject (such as a human patient) in the prophylactic or therapeutic treatment of a disease, disorder or pathological condition.

The term "substantially identical" sequence as used herein refers to a sequence which is at least about 90%, preferably at least about 95%, 96%, 97%, 98%, or 99% identical to a reference sequence. Identity percentage between the two sequences can be determined by any means known in the art, for example the Needleman and Wunsch global alignment algorithm.

The term "antibody" as used herein refers to an immunoglobulin or an antigen-binding fragment thereof. Unless otherwise specified, the term includes monoclonal and polyclonal antibodies, as well as recombinant antibodies. The term "antibody" also refers to fragments and derivatives of all of the foregoing, and may further comprise any variants thereof that retain the ability to specifically bind an epitope. Antibodies may include, but are not limited to monoclonal antibodies (mAbs), camelid antibodies, single-chain antibodies (scFvs), Fab fragments, F(ab')2 fragments, disulphide- linked Fvs (sdFv) fragments, anti-idiotypic (anti-Id) antibodies, intra- bodies, synthetic antibodies, and epitope-binding fragments of any of the above. The term "antibody" also refers to a fusion protein that includes a region equivalent to the Fc region of an immunoglobulin.

The term "recombinant antibody" as used herein refers to an antibody produced or expressed using a recombinant expression vector, where the expression vector comprises a nucleic acid encoding the recombinant antibody, such that introduction of the expression vector into an appropriate host cell results in the production or expression of the recombinant antibody. Recombinant antibodies may be chimeric or humanized antibodies, mono- or multi-specific antibodies.

The term "an antibody mimetic" (AbM) as used herein refers to single-domain scaffolds, which have been engineered to bind therapeutic targets with affinity and specificity that match that of natural antibodies. Antibody mimetics have been developed utilizing an immunoglobulin-like fold, for example, fibronectin type III, NCAM and CTLA-4. Other mimetics scaffolds bearing no similarity to immunoglobulin folds have also been obtained. Non-limiting examples of said scaffolds are DARPins, anticalins, affibodies, adnectins, fynomers, etc. (see for instance, Weidle et al. Cancer Genomics & Proteomics. 2013, 10: 1-18; Lofblom, J. et al, Curr. Opin. Biotechnol. 2011, 22: 843-848; Banta, S. et al, Annu. Rev. Biomed. Eng., 2010, 15: 93-113).

The term "marker" or "biomarker" as used herein refers to markers of disease, prognostic or predictive markers which are typically substances found in a bodily sample that can be easily measured. Said bodily sample can be for instance a blood, plasma or feces sample. The term biomarker encompasses biophysical and biochemical determinations, including genetic and serological markers.

The term "combination therapy" as used throughout the specification, is meant to comprise the administration of the referred therapeutic agents to a subject suffering from cancer, in the same or separate pharmaceutical formulations, and at the same time or at different times. If the therapeutic agents are administered at different times they should be administered sufficiently close in time to provide for the combined effect (e.g. potentiating or synergistic response) to occur. The particular combination of therapies to employ in a combination regimen will take into account compatibility of the desired therapeutics and/or procedures and/or the desired therapeutic effect to be achieved. It will be appreciated that the therapies employed may achieve a desired effect for the same disorder (for example, anticancer effects), and/or they may achieve different effects (e.g., control of any adverse effects). Prognostic and predictive methods of the Invention

In a first aspect, the present invention provides an in vitro method for determining the prognosis of a subject having lung adenocarcinoma, wherein said method comprises: a) determining in a biological sample isolated from said subject the protein expression levels of nuclear QKI (QKI-N) and membrane SLC2A1 (SLC2A1-MB), optionally, further determining in said biological sample the protein expression levels of at least one, two, three, four or five additional markers selected from the group consisting of:

-nuclear BRCA1 (BRCA1-N),

-cytoplasmic QKI (QKI-C),

-nuclear STC1 (STC1-N),

-nuclear CDC6 (CDC6-N), and

-nuclear SIRT2 (SIRT2-N); b) calculating a combined score from the protein expression levels of the markers determined in the biological sample as defined in step a); and c) classifying the subject as having good prognosis or poor prognosis based on the combined score. In a related aspect, it refers to an in vitro method for obtaining useful data for the prognosis of a subject having lung adenocarcinoma, said method comprising the steps defined above.

As used herein, the term "lung cancer" is used for cancer that starts in the lung. Non-small cell lung carcinoma (NSCLC) accounts for 80-85% of all cases and includes the two most frequent lung cancer types: adenocarcinomas (ADC) and squamous cell carcinomas (SCC). Small Cell lung cancer (SCLC) comprises 10-15% of lung cancer cases.

Signs or symptoms which may be indicative of lung cancer include for instance one or more of the following: weight loss, loss of appetite, malaise, fever, cough, dyspnea, wheezing, stridor, hoarseness, shortness of breath, weakness, haemoptisis, chest and or back pain, obstructive pneumonia and pleural effusion (see for instance, WHO classification of Tumors of the Lung, Pleura, Thymus and Heart. Edited by WD Travis, E Brambilla, A.P. Burke, A. Marx and A.G. Nicholson (2015). Adenocarcinoma is a malignant epithelial tumor with glandular differentiation, mucin production or pheumocyte marker expression. The tumors show an acinar, papillary, micropapillary, lepidic or solid growth pattern, with either mucin or pneumocyte marker expression(see for instance, WHO classification of Tumors of the Lung, Pleura, Thymus and Heart. Edited by WD Travis, E Brambilla, A.P. Burke, A. Marx and A.G. Nicholson (2015).

In the present invention, the expression "a subject having lung adenocarcinoma" may refer to a subject which has been suspected to have or has been diagnosed with lung cancer and wherein further to histological analysis (e.g. from a pre-surgery biopsy or a biopsy from the resected tumor) it has been determined that said cancer is adenocarcinoma. In a particular embodiment, said subject has been submitted to tumor resection surgery and has not received any neoadjuvant treatment.

The combined score determined by the methods of the present invention may be predictive and/or prognostic.

Disease progression or outcome may be measured using different parameters, including but not limited to, tumor growth, tumor growth delay, increase/decrease of tumor size, increase/decrease in tumor markers, and patient's survival. Preferably, in the context of the present invention, the clinical outcome of a subject, is expressed as overall survival and/or disease-free survival. Survival of cancer patients is generally suitably expressed by Kaplan-Meier curves, named after Edward L. Kaplan and Paul Meier who first described it (Kaplan, Meier: Amer. Statist. Assn. 53:457-481). The Kaplan-Meier estimator is also known as the product limit estimator. It serves for estimating the survival function from life- time data. A plot of the Kaplan-Meier estimate of the survival function is a series of horizontal steps of declining magnitude which, when a large enough sample is taken, approaches the true survival function for that population. The value of the survival function between successive distinct sampled observations is assumed to be constant. With respect to the present invention, the Kaplan-Meier estimator may be used to measure the fraction of patients living for a certain amount of time after beginning a therapy (e.g. after tumor resection). The clinical outcome predicted may be the (overall/disease-free) survival in months/years from the time point of taking the sample. It may be survival for a certain period from taking the sample, such as of six months or more, one year or more, two years or more, three years or more, four years or more, five years or more, six years or more. In each case, "survival" may refer to "overall survival" or "disease free survival".

The term "disease free survival", or "DFS" as used herein, is defined as the interval of time from start of treatment (e.g., date of surgery) to the first measurement of cancer growth. The term "overall survival" or "OS" as used herein, is defined as the interval of time from the start of treatment (e.g., date of surgery) to death from any cause.

The term "poor prognosis" as used herein refers to a high risk of recurrence and/or death. In preferred embodiments, the term "poor prognosis" means a survival (i.e. DFS and/or OS) of six months or less, one year or less, two years or less, three years or less, four years or less, five years or less, six years or less, etc. In one embodiment, the term poor prognosis refers to a DFS and/or OS of less than 5 years.

Step (a) of the method under the first aspect of the invention comprises determining in said biological sample the expression levels of the protein markers defined above at the target subcellular location [i.e., nuclear (N), cytoplasmic (C) and/or membrane (MB)].

The protein quaking (QKI), also referred as Hqk, is an RNA-binding protein which may be expressed in the nucleus or cytoplasm and has been reported to play a central role in myelinization (Aberg K. Proc Natl Acad Sci U S A. 2006, 103(19):7482-7). The canonical sequence of human QKI is referred as SEQ ID NO: l (UniProtKB Accession Number Q96PU8-1 of the entry entry version 140 of 30 Aug 2017, sequence version 1 of 1 Dec 2001):

The term "QKI" as used herein refers to human QKI protein with SEQ ID NO: l and to sequences substantially identical thereto. Preferably, said sequence is SEQ ID NO: 1.

The Solute carrier family 2, facilitated glucose transporter member 1 (SLC2A1), also referred as GLUT-1 or HepG2 glucose transporter, is a facilitative glucose transporter which may be located in the cell membrane or in the cytoplasm. This isoform may be responsible for constitutive or basal glucose uptake. It has been described to have very broad substrate specificity; and to be able to transport a wide range of aldoses including both pentoses and hexoses. The canonical sequence of human SLC2A1 is referred as SEQ ID NO:2 (UniProtKB Accession Number PI 1166-1 of the entr version 210 of 30 Au 2017 se uence version 2 of 3 Oct 2006 :

The term "SLC2A1" as used herein refers to human SLC2A1 protein with SEQ ID NO:2 and to sequences substantially identical thereto. Preferably, said sequence is SEQ ID NO:2.

Breast cancer type 1 susceptibility protein (BRCAl), also referred as RING finger protein 53 or RING-type E3 ubiquitin transferase BRCAl, is an E3 ubiquitin-protein ligase that specifically mediates the formation of 'Lys-6'-linked polyubiquitin chains and plays a central role in DNA repair by facilitating cellular responses to DNA damage. It may be expressed in the nucleus or the cytoplasm. It is unclear whether it also mediates the formation of other types of polyubiquitin chains. The E3 ubiquitin-protein ligase activity is required for its tumor suppressor function. The BRCAl -BARD 1 heterodimer coordinates a diverse range of cellular pathways such as DNA damage repair, ubiquitination and transcriptional regulation to maintain genomic stability. The canonical sequence of human BRCAl is referred as SEQ ID NO:3 (UniProtKB Accession Number P38398-1 of the entry version 228 of 30 Aug 2017, sequence version 2 of 1 Feb 1995):

The term "BRCA1" as used herein refers to human BRCA1 protein with SEQ ID NO:3 and to sequences substantially identical thereto. Preferably, said sequence is SEQ ID NO:3. Stanniocalcin-1 (STC1), is a protein which may be expressed in the nucleus or cytoplasm and has been reported stimulate renal phosphate reabsorption, and could therefore prevent hypercalcemia. The canonical sequence of human STC1 is referred as SEQ ID NO:4 (UniProtKB Accession Number P52823-1 of the entry version 135 of 7 Jun 2017, sequence version 1 of 1 Oct 1996):

The term "STC1" as used herein refers to human STC1 protein with SEQ ID NO:4 and to sequences substantially identical thereto. Preferably, said sequence is SEQ ID NO:4.

Cell division control protein 6 homolog (CDC6), also referred as CDC6-related protein or Cdcl8-related protein, is a protein which may be expressed in the nucleus or cytoplasm and has been reported to be involved in the initiation of DNA replication and Also participate in checkpoint controls that ensure DNA replication is completed before mitosis is initiated. The canonical sequence of human CDC6 is referred as SEQ ID NO:5 (UniProtKB Accession Number Q99741-1 of the entry version 162 of 30 Aug 2017, sequence version 1 of 1 May 1997):

The term "CDC6" as used herein refers to human CDC6 protein with SEQ ID NO:5 and to sequences substantially identical thereto. Preferably, said sequence is SEQ ID NO:5.

NAD-dependent protein deacetylase sirtuin-2 (SIRT2), also referred as Regulatory protein SIR2 homolog 2 or SIR2-like protein 2, is an NAD-dependent protein deacetylase, which deacetylates internal lysines on histone and alpha-tubulin as well as many other proteins such as key transcription factors. It has been reported to participate in the modulation of multiple and diverse biological processes such as cell cycle control, genomic integrity, microtubule dynamics, cell differentiation, metabolic networks, and autophagy. It plays a major role in the control of cell cycle progression and genomic stability. This protein may be expressed in nucleus or the cytoplasm. The canonical sequence of human SIRT2 is referred as SEQ ID NO:6 (UniProtKB Accession Number Q8IXJ6-1 of the entry version 167 of 30 Aug 2017, sequence version 2 of 31 Oct 2003):

The term "SIRT2" as used herein refers to human SIRT2 protein with SEQ ID NO:6 and to sequences substantially identical thereto. Preferably, said sequence is SEQ ID NO:6. In a particular embodiment, the prognosis method of the invention comprises determining in step a) the protein expression levels of nuclear QKI (QKI-N) and membrane SLC2A1 (SLC2A1- MB), and further comprises:

a. l) determining in said biological sample the protein expression levels of at least one, two, three, four, five additional markers selected from the group consisting of: nuclear BRCA1 (BRCAl-N), cytoplasmic QKI (QKI C), nuclear STCl (STCl-N), nuclear CDC6 (CDC6-N), and nuclear SIRT2 (SIRT2-N).

In another particular embodiment, the prognosis method of the invention comprises determining in step a) the protein expression levels of: - QKI-N, SLC2A1-MB and BRCA1-N; or

- QKI-N, SLC2A1-MB and QKI-C; or

- QKI-N, SLC2A1-MB and STC1-N; or

- QKI-N, SLC2A1-MB and CDC6-N.

In a preferred embodiment, the prognosis method of the invention comprises determining in step a) the protein expression levels of:

- QKI-N, SLC2A1-MB, BRCA1-N and QKI-C; or

- QKI-N, SLC2A1 -MB, BRCA1-N, QKI-C and STC1-N; or

- QKI-N, SLC2A1-MB, BRCA1-N, STC1-N, CDC6-N and SIRT2-N; or

- QKI-N, SLC2A1-MB, STC1-N and CDC6-N.

Preferably, the prognosis method of the invention comprises determining in step a) the protein expression levels of QKI-N, SLC2A1-MB, BRCA1-N and QKI-C.

The expression "determining the levels of the marker" or "determining the protein expression levels", as used herein, refers to ascertaining the absolute or relative amount or concentration of the biomarker in the sample. Techniques to assay levels of individual biomarkers from test samples are well known to the skilled technician, and the invention is not limited by the means by which the components are assessed.

Expression levels may be absolute or relative. When the expression levels are normalized, normalization can be performed with respect to different measures in the sample. These procedures are well known to one skilled in the art. Typically, expression levels are normalized with respect to an "endogenous control". An "endogenous control" as used herein relates to a gene expression product whose expression levels do not change or change only in limited amounts in tumor cells with respect to non-tumorigenic cells. "Endogenous control" is usually the expression product from a housekeeping gene and which codes for a protein which is constitutive ly expressed and carries out essential cellular functions. Housekeeping genes that can be used as endogenous control include for example β-2-microglobulin, ubiquitin, 18-S ribosomal protein, cyclophilin, GAPDH, actin and HPRT.

The expression levels of the above-mentioned markers may be determined by any method enabling to determine the specific protein expression levels at the target subcellular localization (i.e., nucleus (N), cytoplasm (C) or membrane (MB). For instance, protein extracts of the specific subcellular location of the target marker may be obtained. Cell lysis, fractionation and protein extraction methods are well known in the art and may be found in protein preparation handbooks (see for instance, Walker JM (2009) The Protein Protocols Handbook. Third Edition. New York (NY): Springer- Verlag New York, LLC). Subsequently, protein determination from the protein extracts can be performed by a suitable method.

In a particular embodiment, optionally in combination with any of the embodiments or features described above or below, protein expression levels are determined by a method comprising: a) incubating nuclear protein extracts with an affinity reagent for QKI,

b) incubating membrane protein extracts with an affinity reagent for SLC2A1; and c) optionally, incubating target protein extracts with an affinity reagent for each of the at least one, two, three, or the four additional proteins defined in step a).

Methods for quantifying protein expression are well known in the art. Suitable methods for determining the levels of a given protein include, without limitation, those described herein below. Preferred methods for determining the protein expression levels in the methods of the present invention are immunoassays. Various types of immunoassays are known to one skilled in the art for the quantitation of proteins of interest. These methods are based on the use of affinity reagents, which may be any antibody or ligand specifically binding to the target protein or to a fragment thereof, wherein said affinity reagent is preferably labeled. For instance, the affinity reagent may be enzymatically labelled, or labeled with a radioactive isotope or with a fluorescent agent.

Affinity reagents may be any antibody or ligand specifically binding to the target protein or to a fragment thereof. Affinity ligands may include proteins, peptides, peptide aptamers, affimers and other target specific protein scaffolds, like antibody-mimetics. Preferred affinity reagents for use in the determination of the protein biomarkers of the invention are antibodies, for example the antibodies defined in Table 3.

Specific antibodies against the protein markers used in the methods of the invention may be produced for example by immunizing a host with a protein of the present invention or a fragment thereof. Likewise, peptides specific against the protein markers used in the methods of the invention may be produced by screening synthetic peptide libraries.

Western blot or immunoblotting techniques allow comparison of relative abundance of proteins separated by an electrophoretic gel (e.g., native proteins by 3-D structure or denatured proteins by the length of the polypept ide). Immunoblotting techniques use antibodies (or other specific ligands in related techniques) to identify target proteins among a number of unrelated protein species. They involve identification of protein target via antigen-antibody (or protein- ligand) specific reactions. Proteins are typically separated by electrophoresis and transferred onto a sheet of polymeric material (generally nitrocellulose, nylon, or polyvinylidene difluoride). Dot and slot blots are simplified procedures in which protein samples are not separated by electrophoresis but immobilized directly onto a membrane.

Traditionally, quantification of proteins in solution has been carried out by immunoassays on a solid support. Said immunoassay may be for example an enzyme-linked immunosorbent assay (ELISA), a fluorescent immunosorbent assay (FIA), a chemiluminescence immunoassay (CIA), or a radioimmunoassay (RIA), an enzyme multiplied immunoassay, a solid phase radioimmunoassay (SPROA), a fluorescence polarization (FP) assay, a fluorescence resonance energy transfer (FRET) assay, a time-resolved fluorescence resonance energy transfer (TR- FRET) assay, a surface plasmon resonance (SPR) assay. Multiplex and any next generation versions of any of the above, such as bead-based flow-cytometry immunoassays (e.g., based on the Luminex xMAP technology) are specifically encompassed. In a particular embodiment, said immunoassay is an ELISA assay or any multiplex version thereof. Other methods that can be used for quantification of proteins in solution are techniques based on mass spectrometry (MS) such as liquid chromatography coupled to mass spectrometry (LC / MS), described for example in US2010/0173786, or tandem LC-MS / MS (WO2012/155019, US2011/0039287, M. Rauh, J Chromatogr B Analyt Technol Biomed Life Sci 2012 February 1, 883-884. 59-67) and multiplex versions of the above techniques, as well as the next generation of such techniques and combinations thereof.

For determining protein expression and location, immunohistochemical and in-situ hybridization analysis are usually preferred. In a preferred embodiment, optionally in combination with one or more of the embodiments or features described above or below, protein quantification and subcellular location is performed by immunohistochemistry.

In a particular embodiment, optionally in combination with any of the embodiments or features described above or below, protein expression levels are determined by a method comprising incubating the biological sample isolated from said subject with:

a) an affinity reagent for QKI,

b) an affinity reagent for SLC2A1 , and

c) optionally, an affinity reagent for each of the at least one, two, three, or the four additional proteins defined in step a).

Immunohistochemistry (IHC) analysis is typically conducted using thin sections of the biological sample immobilised on coated slides. These sections, when derived from paraffin-embedded tissue samples, are deparaffinised and preferably treated so as to retrieve the antigen. The detection can be carried out in individual samples or in tissue microarrays. This procedure, although is subjectively determined by the pathologist, is the standard method of measurement of IHC results, and well known in the art.

In a particularly preferred embodiment of the present invention, the use of this technique entails the determination of the Histological score value.

In the context of the present invention the Histological score (H-Score) value is determined per biological sample according to (i) staining intensity and (ii) the percentage of positive staining tumor cells by using the following formula:

H- Score =∑ (intensity grade x % stained cells)

The staining intensity of tumor cells may be scored in different intensity grades, for example the following 4 grades:

1. no staining or very weak staining (intensity = 0);

2. positive weak staining (intensity= 1 or +);

3. positive moderate staining (intensity=, 2 or ++); and

4. positive strong staining (intensity=3 or +++). The percentage of positive staining cells for each intensity grade may be scored from 0 to 100.

The final score, called "H-Score", is preferably calculated by adding the products of the percentage cells stained with a given intensity grade (0-100) by the corresponding staining intensity grade value (0-3). The following formula may be applied:

H-Score =∑ intensity grade x % stained cells = 1 x (% of cells with weak staining) + 2 x (% of cells with moderate staining) + 3 x (% of cells with strong staining)

The assessment of the staining intensity and the percentage of positive staining tumor cells can be determined by any means known to the skilled person including but not limited to a panel of at least two independent pathologists with no knowledge about clinical data scoring all immunohistochemical stainings. In case, the panel of pathologist were to disagree in the scores it is convenient to expand the panel of independent pathologists to at least 3, 4, or 5.

Once the staining intensity and the percentage of positive staining tumor cells have been determined, the value of the H-Score may be obtained by applying the above formula. The resulting value of the H-Score determines the level of expression of the protein marker.

A preferred method for determining the target protein expression by immunohistochemistry is described herein below.

The prognostic method of the invention can be applied to any type of biological sample from a patient, such as a biopsy sample, tissue, cell or fluid (serum, saliva, semen, sputum, cerebral spinal fluid (CSF), tears, mucus, sweat, milk, brain extracts and the like). In performing the method of the present invention, said biological sample from the cancer patient is preferably a sample containing tumor cells. Tumors or portions thereof may be surgically resected from the patient or obtained by routine biopsy. Preferably, a tumor sample is obtained from the primary tumor. In a particular embodiment, optionally in combination with one or more of the embodiments or features described above or below, said biological sample isolated from the subject is a tumor biopsy sample, preferably obtained from a resected tumor.

These types of samples are routinely used in the clinical practice and a person skilled in the art will know how to identify the most appropriate means for their obtaining and preservation. Once a sample has been obtained, it may be used fresh, it may be frozen or preserved using appropriate means (e.g., as a formal in- fixed, paraffin-embedded tissue sample). Such biological samples can be taken around the time of diagnosis, before, during or after treatment (e.g. surgical resection).

Step (b) of the prognosis method under the first aspect of the invention comprises calculating a combined score. The combined score is a value obtained according to a given mathematical algorithm wherein the expression values of each of the protein markers used in the methods of the invention are variables of said mathematical algorithm.

In a particular embodiment, when calculating the combined score, this is proportional to the expression levels of one or more of QKI-C, SLC2A1-MB, BRCA1-N, SIRT2-N and is inversely proportional to the expression levels of the QKI-N, STC1-N and CDC6-N,

wherein the higher the score, the worst the prognosis and/or the higher the risk of relapse.

Preferably, said combined score is calculated as the sum of the products of the standardized beta coefficients obtained in a regression analysis for each marker and the protein expression values (e.g. expressed as H-Score value). In the context of the invention, the combined score obtained in this way is also named Prognostic Index (PI).

In a particular embodiment, said beta coefficients are positive for markers selected from the group consisting of QKI-C, SLC2A1-MB, BRCA1-N, SIRT2-N and negative for markers selected from the group consisting of QKI-N, STC1-N and CDC6-N. A positive sign is indicative that an increase of the individual marker is associated to poor prognosis, whereas a negative sign means that an increase of the individual marker is associated to good prognosis.

In a preferred embodiment, the combined score is a Prognostic Index obtained by using a formula selected from the group consisting of: a) - 0.004 x H-Score_QKi-N + 0.005 x H-Score_SLc2Ai-MB + 0.006 x H-Score_BRCAi-N + 0,006 x H-Score_QKi-c;

b) - 0.004 x H-Score_QKi-N + 0.004 x H-Score_SLc2Ai-MB + 0.007 x H-Score_BRCAi-N + 0.007 x H-ScoreQKi-c - 0.005 x H-ScoresTci-N;

c) - 0.004 x H-ScoreQKi-N + 0.005 x H-Score_SLc2Ai-MB + 0.003 x H-Score_BRCAi-N - 0.006 x H- ScoresTci-N - 0.005 x H-ScorecDC6 N + 0.006 x H-ScoresiRT2 N; and d) - 0.003 x H-ScoreQKi-N + 0.005x H-ScoresLC2Ai-MB - 0.003 x H-ScoresTci-N - 0.003 x H- Score_CDC6-N

Furthermore, step (c) of the method under the first aspect of the invention comprises classifying the subject as having good prognosis or poor prognosis based on the combined score.

Typically, in step c) said method comprises comparing the combined score in the subject sample with a reference combined score; and an increase of the combined score in the subject sample with regard to said reference combined score is indicative of poor prognosis.

The term "reference combined score" as used herein is a reference value obtained according to a given mathematical algorithm wherein reference expression values of each of the protein markers used in the prognosis method of the invention are variables of said mathematical algorithm.

The term "reference value", as used herein, relates to a predetermined criteria used as a reference for evaluating the values or data obtained from the samples collected from a subject. This "reference value" may also be referred as "cut-off value" or "threshold value". The reference value or reference level can be an absolute value, a relative value, a value that has an upper or a lower limit, a range of values, an average value, a median value, a mean value, a tertile value, or a value as compared to a particular control or baseline value. In a particular embodiment, optionally in combination with one or more of the embodiments or features described above or below, said reference value is the mean value or the tertile value.

A reference value can be based on an individual sample value but is generally based on a large number of samples, including or excluding the sample to be tested. For instance, this reference value may be derived from a collection of tumor tissue samples from a reference lung adenocarcinoma (ADC) patients' population for whom historical information relating to the actual clinical outcome for the corresponding cancer patient is available. Said reference lung ADC patient's population may for instance be from subjects suffering from lung ADC, from patients' suffering from resectable lung ADC (e.g., from stages I to III; or I to Ilia), or from subjects suffering from early stage lung ADC (e.g., from stage I or II). In a particular embodiment, optionally in combination with any of the embodiments described above or below, said combined reference value is determined by a method comprising: a) determining, for each lung ADC patient in a reference population, the protein expression levels of the protein markers as defined in step a) and calculating the combined score as defined in step b) of the method of the invention; b) selecting as provisional reference value an arbitrary combined score from the ones obtained in step a); c) classifying the patients in the reference ADC population in two groups according to the provisional reference value selected combined score obtained in a), wherein:

(i) the first group comprises lung ADC patients that exhibit a combined score that is lower than the arbitrary/provisional reference value; and

(ii) the second group comprises lung ADC patients that exhibit a combined score that is higher than the arbitrary/provisional reference value;

whereby two groups of lung ADC patients are obtained for the specific reference value d) statistically correlating the combined score with the clinical follow up data (disease free survival and overall survival) using log rank test.

In the prognosis method of the invention, the combined score obtained in step b) is considered "decreased" when said combined score is lower than a reference combined score. Preferably, the combined score is considered to be lower than a reference combined score when it is at least 10%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, at least 100%, at least 110%, at least 120%, at least 130%, at least 140%, at least 150%, or more lower than the reference combined score.

Likewise, in the context of the prognosis method of the invention, the combined score obtained in step b) is considered "increased" when said combined score is higher than a reference value. Preferably, the combined score is considered to be higher than a reference combined score when it is at least 10%, at least 15%, at least 20%, at least 25%, at least 30%, at least 35%, at least 40%, at least 45%, at least 50%, at least 55%, at least 60%, at least 65%, at least 70%, at least 75%, at least 80%, at least 85%, at least 90%, at least 95%, at least 100%, at least 110%, at least 120%, at least 130%, at least 140%, at least 150%, or more higher than a reference combined score.

Alternatively or in addition, subjects having more than about 1.1,1.2, 1.3, 1.4, 1.5, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15 or 20 fold levels deviation (i.e., increase or decrease) than the reference combined score as described herein.

Prognosis or outcome prediction in the method of the invention, as it is understood by a person skilled in the art, does not claim to be correct in 100% of the analyzed samples. However, it requires that a statistically significant amount of the analyzed samples are classified correctly. The amount that is statistically significant can be established by a person skilled in the art by means of using different statistical tools; illustrative, non-limiting examples of said statistical tools include determining confidence intervals, determining the p-value, the Chi-Square test discriminating functions, etc. Preferred confidence intervals are at least 90%, at least 95%, at least 97%), at least 98%>, at least 99%>. The p-values are, preferably less than 0.1, less than 0.05, less than 0.01, less than 0.005 or less than 0.0001. The teachings of the present invention preferably allow correctly classifying at least 60%>, at least 70%>, at least 80%>, or at least 90%> of the subjects of a determining group or population analyzed. Irrespective of the histology, tumors can be further stratified according to its stage of development. Nowadays, the tumor-node-metastasis (TNM) staging system is the standard method for clinically predicting survival of patients with lung cancer. This system is based on tumor size, tumor location and involved structures, and presence of nodal and distant metastasis to categorize lung cancer patients in different clinical stages. With the continuous flow of new data and the increasing knowledge of the disease, the staging system requires continuous adjustment. Nowadays the seventh edition of the TNM classification of lung tumors is been used worldwide, although the eight edition has been published; the implementation of the new system has been delayed until January 1, 2018 (Goldstraw P et al, 2007; Goldstraw P et al, 2016). Further details on the TNM method, as well as descriptors and stage groupings according to the 7^th edition are provided in the Examples.

The inventors have also found that prognostic information in lung ADC patients is significantly improved when the Prognostic Index (PI) described above is further combined with a correction coefficient specific for the TNM stage of the particular ADC patient. As referred in Example 4, TNM stage alone was a highly significant prognostic factor in lung ADC patients of both DFS (P<0.001) and OS (P=0.001, Figure 28A). Interestingly, the likelihood ratio significantly increased after adding the molecular information of the Prognostic Index (PI) (P<0.001 both for DFS and OS). This improvement showed that the molecular model complements the TNM stage, adding very valuable prognostic information for the patients.

Accordingly, in a particular embodiment, optionally in combination with any of the embodiments or features described above or below, the prognosis method of the invention comprises further to step a):

Al) determining in said sample the stage of lung adenocarcinoma according to the TNM classification of lung tumors; and

b) calculating a combined score from the protein expression levels of the markers determined in the biological sample as defined in step a) and the stage of TNM classification as defined in step Al); and

c) classifying the subject as having good prognosis or poor prognosis based on the combined score obtained in b).

The prognosis method of this particular embodiment, is characterized by having corrected the combined score calculated on the basis of the levels of expression of the protein markers by the tumor TNM stage, i.e. the calculation of the combined score in step b) comprises correction of the protein marker based combined score with a TNM stage specific correction coefficient.

This improved combined score is herein referred to as Combined Prognostic Index (CPI). Preferably, said subject has resectable lung ADC (e.g., from stage I to IIIA). In a particular embodiment, optionally in combination with any of the embodiments or features described above or below, the combined score in b) is calculated according to the following algorithm:

CPI = 1.090 x PI + n;

where n is a correction coefficient specific for each TNM stage, wherein for IA, n=0; for

IB, n=0.507; for IIA, n=1.099; for IIB, n=1.583; and for IIIA, n=1.450.

It is further noted that the accuracy of the method of the invention can be further increased by determining the presence and/or quantification of other prognostic/predictive molecular markers (Coate LE et al, 2009; Ferte C et al 2010; Lin J and Beer DG, 2012; Zhu CQ and Tsao MS, 2014), and/or clinical signs or symptoms with reported prognostic/predictive value, such as morphological features of the tumor, histological subtypes, radiological traits of the imaging tests (e.g. size, shape, volume, radiological texture, morphological details or other features in a CT scan, X-Ray or SUV or alternative ways to analyze nuclear tracer levels in a PET imaging, etc); clinical characteristics of the patients (e.g. age, sex, race, respiratory function levels, performance status). The potential additional molecular markers to be associated to the present invention can be found in the tumor specimen itself or other cells, body fluids or exhaled breath obtained from the same patient.

The methods of the present invention or any of the steps thereof might be implemented by a computer. Therefore, a further aspect of the invention refers to a computer implemented method, wherein the method is any of the methods disclosed herein or any combination thereof.

It is noted that any computer program capable of implementing any of the methods of the present invention or used to implement any of these methods or any combination thereof, also forms part of the present invention.

An in vitro method for determining the prognosis of a subject having lung adenocarcinoma, wherein said method comprises: a) determining in a biological sample isolated from said subject the protein expression levels of protein markers as described herein above;

b) calculating, using a computer, a combined score from the protein expression levels of the markers determined in the biological sample as defined in step a);

wherein the protein expression levels are calculated as H-Score values, and wherein the H-Score values are determined by adding the products of the percentage of cells stained with a given intensity (0-100) by the staining intensity using the following formula: H- Score =∑ intensity grade x % stained cells; and

c) classifying the subject as having good prognosis or poor prognosis based on the combined score.

This computer program is typically directly loadable into the internal memory of a digital computer, comprising software code portions for performing the steps of comparing the combined score (e.g., obtained from the level of one or more of the target markers as described in the invention), from the one or more biological samples of a subject with a reference value (e.g., reference combined value) and determining the prognosis of said subject or whether it would benefit from adjuvant therapy, when said product is run on a computer.

It is also noted that any device or apparatus comprising means for carrying out the steps of any of the methods of the present invention or any combination thereof, or carrying a computer program capable of, or for implementing any of the methods of the present invention or any combination thereof, is included as forming part of the present specification.

The methods of the invention may also comprise the storing of the method results in a data carrier, preferably wherein said data carrier is a computer readable medium. The present invention further relates to a computer-readable storage medium having stored thereon a computer program of the invention or the results of any of the methods of the invention.

As used herein, "a computer readable medium" can be any apparatus that may include, store, communicate, propagate, or transport the results of the determination of the method of the invention. The medium can be an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system (or apparatus or device) or a propagation medium.

Other methods of the invention

In another aspect, the present invention provides an in vitro method for selecting those subjects having lung adenocarcinoma who are expected to benefit from an adjuvant treatment, wherein said method comprises:

a) classifying said subject according to the prognosis method as described herein;

b) deciding on whether to administer an adjuvant treatment according to the prognosis classification;

wherein a subject classified as having poor prognosis will be selected for administration of an adjuvant treatment.

The term "adjuvant treatment" as used herein refers to a treatment administered after surgical resection of a tumor, typically the primary tumor. Said adjuvant treatment may comprise the administration of any cytotoxic or antiproliferative drug and includes chemotherapy and/or targeted therapies regimens. For instance, said adjuvant treatment may be selected from (but it is not limited to) any one or more of the group consisting of platinum anticancer agents (e.g., cisplatin, oxaliplatin, carboplatin, BBR3464, satraplatin, tetraplatin,ormiplatin, and iproplatin); antimetabolites (e.g., 5-fluorouracil, gemcitabine, cytarabine, capecitabine, decitabine, floxuridine, 6-mercaptopurine, methotrexate, fludarabine, aminopterin, pemetrexed, raltitrexed, cladribine, clofarabine, fludarabine, mercaptopurine, pentostatin, and thioguanine); mitotic inhibitors (e.g., paclitaxel, docetaxel, vinblastine, vincristine, vindesine, and vinorelbine); anthracycline antibiotics (e.g., bleomycin, daunorubicin, doxorubicin, epirubicin, idarubicin, mitomycin, mitoxantrone, pixantrone, and valrubicin); topoisomerase I and/or II inhibitors (e.g., topotecan, SN-38, irinotecan, camptothecine, rubitecan, etoposide, and teniposide); antitumor monoclonal antibodies (e.g., bevacizumab, cetuximan, panitumumab, trastuzumab, rituximab, tositumomab, alemtuzumab, and gemtuzumab); tyrosine kinase inhibitors (e.g., erlotinib, sorafenib, axitinib, bosutinib, cediranib, dasatinib, gefitinib, imatinib, canertinib, lapatinib, lestaurtimb, nilotinib, semaxanib, sunitinib, and vandetanib);. metabolic modulators (e.g., mTOR inhibitors), epigenetic inhibitors (e.g., DNMT inhibitors) and immunotherapy agents (e.g. Pembrolizumab, Nivolumab, Atezolizumab, Avelumab).

Preferably said adjuvant therapy comprises the administration of a platinum anticancer agent as single agent or in a combination therapy. Preferably, the platinum anticancer agent is cisplatin and/or carboplatin.

In a further aspect, the present invention refers to an in vitro method for predicting the efficacy of an adjuvant treatment (e.g., adjuvant chemotherapy) in a subject having lung adenocarcinoma, wherein said method comprises:

a) classifying said subject according to the prognosis method as described herein; and

b) predicting the efficacy of said adjuvant treatment (e.g. adjuvant chemotherapy) according to the prognosis classification; preferably wherein classification of the subject as having poor prognosis is indicative of increased efficacy of the adjuvant treatment (e.g. adjuvant chemotherapy).

Preferably, said adjuvant treatment comprises the administration of a platinum anticancer agent, and a poor prognosis is indicative of increased efficacy of the platinum anticancer agent.

In another further aspect, the present invention refers to an in vitro method for selecting an adjuvant treatment for a subject having lung adenocarcinoma, wherein said method comprises: a) classifying said subject according to the prognosis method as described herein; and b) selecting an adjuvant treatment (preferably an adjuvant chemotherapy treatment) according to the prognosis classification.

Preferably, when the subject is classified as having poor prognosis an adjuvant chemotherapy (preferably, a platinum anticancer agent) is selected.

In an additional aspect, the present invention provides a method for treating a subject having lung adenocarcinoma comprising administering to said patient a therapeutically effective amount of a treatment, typically an adjuvant treatment, wherein said treatment is selected according to the classification of said patient according to the prognosis method of the invention. Preferably, when said patient is classified as having poor prognosis the administered treatment comprises a platinum anticancer agent.

In a preferred embodiment of the methods of the invention, said lung adenocarcinoma is resectable lung adenocarcinoma which generally refers to stages I to III according to the TNM classification of lung tumors, preferably to stages I to Ilia.

In another preferred embodiment of the methods of the invention, said lung adenocarcinoma is early stage lung adenocarcinoma, which generally refers to stage I or II according to the TNM classification of lung tumors.

In an additional preferred embodiment, said lung adenocarcinoma is stage I adenocarcinoma according to the TNM classification of lung tumors. Preferably, TNM stages are defined according to 7th edition of the TNM classification of lung tumors (see Table 12 in the Examples). More specifically, these are defined as follows:

- stage IA is characterized as Tla-b, NO and MO;

- stage IB is characterized as T2a, NO and MO;

- stage IIA is characterized as Tla-b, Nl and MO; or T2a, Nl and MO; or T2b, NO and M0; and

- stage IIB is characterized as T2b, Nl and M0; or T3, NO and M0;

stage IIIA is characterized as T1/T2, N2 and M0; T3, Nl/2 and M0, or T4, N0/N 1 , M0; and

stage IIIB is characterized as T4, N2 and M0; Any T, N3 and M0.

- stage IV is characterized as any T, any N, and Mia; T, any N, and Mlb. Another aspect of the invention relates to prognostic and/or predictive biomarkers of lung adenocarcinoma (ADC) or combinations of a plurality thereof (signatures) as defined in any of the embodiments described herein above.

A preferred embodiment concerns a plurality of biomarkers for predicting disease progression or survival of a subject with lung adenocarcinoma (ADC), the plurality of biomarkers comprising QKI-N, SLC2A1-MB, BRCA1-N and QKI-C. In another preferred embodiment, it refers to a plurality of biomarkers for predicting disease progression or survival of a subject with lung adenocarcinoma (ADC), the plurality of biomarkers comprising QKI-N, SLC2A1-MB, BRCA1-N, QKI-C and STC1-N.

In a further preferred embodiment, it pertains to a plurality of biomarkers for predicting disease progression or survival of a subject with lung adenocarcinoma (ADC), the plurality of biomarkers comprising QKI-N, SLC2A1-MB, BRCA1-N, STC1-N, CDC6-N and SIRT2-N.

In still an additional preferred embodiment, it relates to a plurality of biomarkers for predicting disease progression or survival of a subject with lung adenocarcinoma (ADC), the plurality of biomarkers comprising QKI-N, SLC2A1 -MB, STC 1 -N and CDC6-N.

Method for determining target protein expression levels by immunohistochemistry for conducting the methods of the invention In another aspect, the present invention provides a method for determining target protein expression levels by immunohistochemistry, said method being suitable for carrying out the determination of the protein expression levels of the markers in the biological sample as defined in step a) of the prognosis method of the invention. In a particular embodiment, optionally in combination with any of the embodiments or features described above or below, said method for determining target protein expression levels comprises incubating the biological sample isolated from said subject with an affinity reagent for QKI, an affinity reagent for SLC2A1, and optionally, an affinity reagent for each of the at least one, two, three, four or the five additional proteins of the additional protein markers defined in step a) of the method of prognosis of the invention. Preferably, said sample is a fixed biopsy sample, typically a formalin fixed-paraffin embedded sample.

Optionally, one or more lung cancer cells (e.g. established cell lines) or tissues are used as low and/or high expression controls as has been described herein. Preferably, said cells or tissues are fixed, typically formalin- fixed and paraffin embedded. As low expression controls may be used lung cancer cells or tissues known to have low protein marker expression, and cells or tissues known to have high protein marker expression can be used as high expression controls. For instance, said lung cancer cell lines may be selected from (but are not limited to) any one or more of the group consisting of NCI-H1395, NCI-H23, NCI-H441, A549, NCI-H358, NCI- HI 299, NCI-H460, CALU-1, NCI-H1869, NCI-H520 y HCC15.

For instance, for each of the following proteins the recited cell lines may be used as high or low expression controls:

- SLC2A1 : NCI-H1395 and/or NCI-H1299 as high expression controls NCI-H460 and/or

A549 as low expression controls;

- QKI: Calu-1 and/or NCI-H1299 as high expression controls and NCI-H1395 and/or NCI- HI 869 as low expression controls;

- BRCA1 : Calu-1 and/or NCI-H1395 as high expression controls and NCI-H460 and/or NCI-H1869 as low expression controls;

- STC 1 : HCC 15 and/or NCI-H358 as high expression controls and NCI-H520 and/or NCI- HI 299 as low expression controls; and

- CDC6: HCC 15 and/or NCI-H460 as high expression controls and NCI-H520 and/or NCI- HI 869 as low expression controls.

Prior to incubation with the affinity reagent as defined herein above said sample and/or control cells may be submitted to a deparaffinization and to an antigen retrieval process. Preferably, the antigen retrieval process is conducted at a temperature of 90°C to 100°C during 10 to 30 minutes, preferably at 95°C during 20 minutes.

In a preferred embodiment, optionally in combination with any of the embodiments or features described above or below, the antigen retrieval process comprises the treatment with a citrate buffer when the antigen to be retrieved is selected from the group consisting of BRCA1, SIRT2, SLC2A1 and STC1; and comprises the treatment with a EDTA buffer when the antigen to be retrieved is selected from the group consisting of CDC6 and QKI. The term "citrate buffer" as used herein refers to a buffer system comprising citric acid and/or a salt thereof (i.e., comprising the corresponding conjugate base, namely citrate ion). Illustrative non-limiting examples of a citrate buffer are sodium citrate buffer or potassium citrate buffer. The pH of the citrate buffer may be from pH 5.0 to pH 7.0, preferably from pH 5.5 to pH 6.5, more preferably about pH 6.0.

The term "EDTA buffer" as used herein refers to a buffer system comprising Ethylenediaminetetraacetic acid (EDTA) and/or salt thereof (i.e., comprising the corresponding conjugate base, namely ethylenediaminetetraacetate). Illustrative non-limiting examples of EDTA buffer are disodium EDTA buffer and calcium disodium EDTA buffer. The pH of the EDTA buffer may be from pH 8.0 to pH 10.0, preferably from pH 7.5 to pH 9.5, more preferably about pH 9.0. Other features and embodiments of the method of measuring the levels of a target protein marker used in the methods of the invention are as described herein for other aspects of the invention.

Kit and use of a kit in the methods of the invention

A further aspect, the present invention refers to a kit for determining the levels of one or more of the target markers as described herein in a biological sample (preferably a tumor biopsy sample) isolated from a subject. The kit may also contain instructions indicating how the materials within the kit may be used. The term "kit" or "testing kit" denotes combinations of reagents and adjuvants required for an analysis. Although a test kit consists in most cases of several units, one-piece analysis elements are also available, which must likewise be regarded as testing kits.

In a particular embodiment, said kit is suitable for determining the levels of at least QKI-N and SLC2A1-MB in a biological sample (preferably a tumor biopsy) and comprises:

- an affinity reagent for QKI; and

- an affinity reagent for SLC2A1;

- optionally, further comprising an affinity reagent for each of the one, two, three or the four additional proteins of the additional protein markers defined in step a) of the prognosis method of the invention; - optionally, further comprising cancer cells to be used as low and/or high expression controls

- optionally, further comprising instructions for the use of said reagents in determining said protein expression levels in a biological sample isolated from a subject.

The term "an affinity reagent for" as used herein refers to an affinity reagent capable of specifically binding to the recited target protein. The various affinity reagents may be labelled with the same or different tags. Preferably, these will be labelled with different tags for multiplex analysis.

In a preferred embodiment, said kit comprises an affinity reagent for each of the proteins in the set of markers selected from the group consisting of: - QKI-N, SLC2A1-MB and BRCA1-N;

- QKI-N, SLC2A1-MB and QKI-C;

- QKI-N, SLC2A1-MB and STC1-N; and

- QKI-N, SLC2A1-MB and CDC6-N. In a further preferred embodiment, said kit comprises an affinity reagent for each of the proteins in the set of markers selected from the group consisting of:

- QKI-N, SLC2A1-MB, BRCA1-N and QKI-C;

- QKI-N, SLC2A1-MB, BRCA1-N, QKI-C and STC1-N;

- QKI-N, SLC2A1 -MB, BRCA1-N, STC1-N, CDC6-N and SIRT2-N; and

- QKI-N, SLC2A1-MB, STC1-N and CDC6-N.

Other particular and preferred marker combinations are as defined herein above for the methods of the invention.

Possible immunoassays and affinity reagents have been described herein. In a particular embodiment, said affinity reagent is an antibody, preferably a monoclonal antibody. The affinity reagent may bind to any linear or conformational region (e.g. epitope) specific for the target protein. In preferred embodiments, the affinity reagents are the antibodies defined in Table 3 and antibodies binding to the same antigenic region and/or epitope in the marker protein (see Table I below)

In a particular embodiment, optionally in combination with any of the embodiments or features described above or below, said kit comprises reagents to perform an immunohistochemistry (ICH) assay. For instance, it may contain inter alia: an enzyme-conjugated secondary antibody (e.g. conjugated to horseradish peroxidase or alkaline phosphatase), an enzyme substrate, and a counterstain such as hematoxylin. Kits for ICH are well known in the art and commercially available (http://www.sigmaaldrich.com/life-science/cell-biology/antibodies/antibodies- application/protocols/immunohistochemistry.html#reagents_equipment). In addition, ICH assay kits for use in a method of the invention also contain reagents for quantification of the target protein markers, as described herein above.

Other preferred features and embodiments of the kit of the invention are as described herein throughout the specification. A further aspect of the invention refers to the use of the kit in a method for the prognosis of a subject having lung adenocarcinoma according to the invention, wherein said kit comprises:

- a reagent for determining the protein expression levels of QKI; and

- a reagent for determining the protein expression levels of SLC2A1 ;

- optionally, further comprising a reagent for each of the one, two, three, or the four additional proteins of the additional protein markers defined in step a) for determining the protein expression levels thereof;

- optionally, further comprising cancer cells to be used as low and/or high expression controls;

- optionally, further comprising instructions for the use of said reagents in determining said proteins expression levels in a biological sample isolated from a subject.

These reagents may be useful for determining the expression levels of the target protein marker(s) using any suitable method for determining the expression of the target marker(s) in a given subcellular localization as described herein above. For instance, the determination of the levels of said protein marker(s) may be carried out by a mass spectrometry (MS)-based method, and said kit may comprise said marker unlabelled and/or said marker stably labelled for detection by a mass spectrometry (MS)-based method, preferably wherein the marker is labelled with a tag which comprises one or more stable isotope. Isotopic atoms which may be incorporated into the tag are heavy atoms for example ¹³C, ¹⁵N, ¹⁷0 and/or ³⁴S, which can be distinguished by MS.

Preferably, these reagents are affinity reagents as described herein for use in immunoassay methods as described herein. Preferably, said immunoassay is an immunohistochemistry assay.

Preferably, said kit is as defined in the above aspect of the invention or in any of its preferred embodiments.

It is contemplated that any features described herein can optionally be combined with any of the embodiments of any method, kit, use of a kit, or computer program of the invention; and any embodiment discussed in this specification can be implemented with respect to any of these. It will be understood that particular embodiments described herein are shown by way of illustration and not as limitations of the invention. The principal features of this invention can be employed in various embodiments without departing from the scope of the invention. Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, numerous equivalents to the specific procedures described herein. Such equivalents are considered to be within the scope of this invention and are covered by the claims.

All publications and patent applications are herein incorporated by reference to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference.

The use of the word "a" or "an" may mean "one," but it is also consistent with the meaning of "one or more," "at least one," and "one or more than one". The use of the term "another" may also refer to one or more. The use of the term "or" in the claims is used to mean "and/or" unless explicitly indicated to refer to alternatives only or the alternatives are mutually exclusive.

As used in this specification and claim(s), the words "comprising" (and any form of comprising, such as "comprise" and "comprises"), "having" (and any form of having, such as "have" and "has"), "including" (and any form of including, such as "includes" and "include") or "containing" (and any form of containing, such as "contains" and "contain") are inclusive or open-ended and do not exclude additional, unrecited elements or method steps. The term "comprises" also encompasses and expressly discloses the terms "consists of and "consists essentially of. As used herein, the phrase "consisting essentially of limits the scope of a claim to the specified materials or steps and those that do not materially affect the basic and novel characteristic(s) of the claimed invention. As used herein, the phrase "consisting of excludes any element, step, or ingredient not specified in the claim except for, e.g., impurities ordinarily associated with the element or limitation. The term "or combinations thereof as used herein refers to all permutations and combinations of the listed items preceding the term. For example, "A, B, C, or combinations thereof is intended to include at least one of: A, B, C, AB, AC, BC, or ABC, and if order is important in a particular context, also BA, CA, CB, CBA, BCA, ACB, BAC, or CAB. Continuing with this example, expressly included are combinations that contain repeats of one or more item or term, such as BB, AAA, AB, BBC, AAABCCCC, CBBAAA, CABABB, and so forth. The skilled artisan will understand that typically there is no limit on the number of items or terms in any combination, unless otherwise apparent from the context.

As used herein, words of approximation such as, without limitation, "about", "around", "approximately" refers to a condition that when so modified is understood to not necessarily be absolute or perfect but would be considered close enough to those of ordinary skill in the art to warrant designating the condition as being present. The extent to which the description may vary will depend on how great a change can be instituted and still have one of ordinary skilled in the art recognize the modified feature as still having the required characteristics and capabilities of the unmodified feature. In general, but subject to the preceding discussion, a numerical value herein that is modified by a word of approximation such as "about" may vary from the stated value by ±1, 2, 3, 4, 5, 6, 7, 8, 9, or 10%. Accordingly, the term "about" may mean the indicated value ± 5% of its value, preferably the indicated value ± 2% of its value, most preferably the term "about" means exactly the indicated value (± 0%).

The following examples serve to illustrate the present invention and should not be construed as limiting the scope thereof.

EXAMPLES Example 1: generation of a protein-based signature that correlates with the clinical outcome of the patient

Description of the experiment

The protein expression levels of 12 genes (see Table 3 below) were analyzed in tumor samples from early stage lung cancer patients by IHC and semiquantitative analysis. The prognostic signature was generated performing a manual regression Cox analysis by steps.

Material and methods

Tumor samples (from 239 ADCs and 117 SCC) were collected from consecutive population cohorts surgically treated at MD Anderson Cancer Center (MDA). Inclusion criteria were as follows: patients with NSCLC, complete resection of the primary tumor, no treatment with radiation or chemotherapy before surgery and absence of cancer within the 5 years before surgery. Lung tumors were classified according to the WHO 2004 classification (Travis WD, 2004) and 7th TNM edition was used for the tumors stratification (Goldstraw P, 2009). The study protocol was approved by local ethics committees of each center. Written informed consent was obtained from each patient. Reported recommendations for tumor marker prognostic studies (REMARK) criteria were followed throughout the study (Altman DG, 2012). For survival analysis the follow-up period was restricted to 60 months in all cohorts.

The clinical and pathological data are summarized in Tables 1 and 2.

Data are n(%). ¹M = . . n erson co ort; = n c n vers ty o avarra. = ot available. ⁴WD: well differentiated; MD: moderately differentiated: PD: poorly differentiated.

First, the specificity of each antibody was validated in NSCLC cell lines by Western Blot (WB), IHC and siRNA knock-down technology when necessary. Ten cancer cell lines were cultured until reaching 80% confluence. Half of the cells were used to obtain formalin- fixed paraffin embedded (FFPE) and the other half was used for protein extraction (total, nuclear and cytoplasmic protein fractions), based on cell lysis by RIPA buffer. For each antibody, we confirmed that only specific and size expected molecular- weight bands appeared in the WB. In most cases, one band corresponding to the canonical isoform appeared but if different bands were detected, we correlated them with the molecular- weights of the gene splice isoforms. Second, we studied the level of expression of each protein by IHQ and we tested the correlation of the expression in the cell lines between both techniques. Additionally, we checked the proper localization in the subcellular compartment using both methods. For those antibodies whose expression detected by Western Blot did not match with the expression observed by IHQ in the cell lines, we evaluated the antibody specificity using siRNA knock-down technology. In the case that the first purchased antibody rendered no conclusive results about its specificity we purchased new antibodies from different companies.

Tumor samples from patients were fixed 24 hours in 10% neutral buffered formalin, and embedded in paraffin blocks. Tissue microarray containing three small cores of tumor tissue from each patient were constructed using a manual tissue array er. The expression of the 12 proteins was analyzed using IHC and quantification by two experienced observers. Briefly, 3 μιη sections were dewaxed in xylene and rehydrated through a graded alcohol series. Endogenous peroxidase activity was quenched with 3% hydrogen peroxidase for 10 minutes. Antigen retrieval was carried out in a Lab Vision PT module at 95°C for 20 minutes either with citrate buffer (lOmmol/L, pH 6) or EDTA buffer (1 mmol/L, pH 8). Afterwards, sections were incubated overnight at 4°C with the primary antibodies. Technical conditions of immunohistochemical procedure (source of primary antibodies, type of antigen retrieval buffer and conditions, , and dilutions) are indicated in Table 3. After applying the EnVision+ System- HRP (Dako) for 30 minutes, immunostaining was developed with Liquid DAB⁺Substrate Chromogen System (Dako).For each marker, the extension was scored as percentage of positive cells (0-100%), and the intensity of staining was assessed into three groups (1, weak; 2, moderate; 3, strong staining). The final score, called "H-Score", was calculated by adding the products of the percentage cells stained with a given intensity (0-100) by the staining intensity (0-3). The following formula was applied:

H-Score =∑ intensity grade x % stained cells = 1 x (% of cells with weak staining) + 2 x (% of cells with moderate staining) + 3 x (% of cells with strong staining)

Table 3

Next, we integrated the expression of the candidate markers to obtain a signature that classifies the patient's risk accurately. We studied the subcellular localization for each protein as an independent variable (N, nuclear; C, cytoplasmic; MB, membrane), a correlation study was conducted using Rho Spearman coefficients between all pairs of variables to avoid multicollinearity problems due to high association between potential predictor variables (Rho Spearman < 0.7). We introduced all variables into the Cox analysis and eliminated each variable one by one according to their adjusted p-value (from highest to lowest) avoiding the loss of more than 10% of the initial (full model) Chi square for each variable. . This method was repeated until we obtained 20 different candidate prognostic models (all of them clinically meaningful), stratifying by stage. We then calculated the prognostic index (PI) for each model, as the sum of the products of the B coefficients for each variable (X, Y, Z...) and the H-Score value: (H-Score X · Coefficient B X) + (H-Score Y · Coefficient B Y) + (H-Score Z · Coefficient B Z) + ... . In order to measure the discrimination of the models we used the Harrell^'s C coefficient (CH), with values between 0 and 1, being 1 the best coefficient. In addition, the models were calibrated by Kaplan-Meier curves, stratified into two groups of risk (low and high). We performed this study independently for the two more frequent histological subtypes (ADC and SCC). Afterwards, we chose the best 10 models of each subgroup and we internally validated them by bootstrapping in 100 different bootstrap samples in order to adjust the CH coefficients. We selected the best models for each subgroup according to the following criteria: the principle of parsimony (the simplest explanation -i.e. less number of variables possible- model) and the highest CH corrected coefficient. The proportional hazards assumption was examined by testing interactions between the co-variables of the final model and time. Univariate and multivariate Cox proportional hazards analyses including other clinical and pathological variables were used to assess the prognostic role of the molecular models (Pis). Only those variables with p < 0.25 in the univariate analysis were included in the multivariate analysis.

Results

In order to analyze the level of expression of the selected proteins in FFPE samples by IHC we optimized the methods and tested the specificity of the antibodies. In the western blot, all of them showed a specific band, corresponding with the canonical isoform (Figures 1-11) except for BRCAl antibody that showed two different bands, one corresponding to the canonical isoform at 220 kDa and another one at 37 kDa (Figure 12). The expression levels observed by western blot for each protein correlated with the one observed by IHC. The subcellular location of the staining was the expected one in all the cases. For BRCAl, inhibition with two short interfering sequences were used to demonstrate the specificity of the two isoforms detected in the Western Blot. As shown in Figure 12D, cells transfected with BRCAl siRNAs showed decreased BRCAl expression in nuclei of the cells.

Once we demonstrated the specificity for each antibody, we analyzed the expression of the proteins using IHC and semi-quantification (H- score determination). Manual regression Cox analysis by steps was used to generate the prognostic signatures. The best model for ADC patients considering the principle of parsimony and the highest adjusted coefficient CH was composed of three different proteins (BRCAl, QKI and SLC2A1) divided in four variables (BRCA1-N, QKI-N, QKI-C and SLC2A1-MB). The equation was:

Model 1 ADC: PI= - 0.004 x H-Score_QKi-_N + 0.006 x H-Score_QKi-c + 0.005 x H-Score_SLc₂Ai -MB + 0.006 x H-ScoreBRCAi-N

To obtain a graphic representation of the models, we stratified the PI score in two different subgroups using the median as cut-off and represented the survival curves by Kaplan-Meier method. The low-risk group had significant longer DFS (p=0.004) and OS time (p<0.001) after five years (Figure 13) when compared to the high-risk group of lung ADC patients.

This algorithm was selected for validation in a larger cohort (see Example 2) and in the further studies presented below. Other obtained signatures with good prognostic value in ADC also comprise the determination of the expression levels of QKI-N and SLC2A1-MB proteins. The other generated algorithms are the following:

To obtain a graphic representation of these three models, we stratified the PI score in two different subgroups using the median as cut-off and represented the survival curves by Kaplan- Meier method. The low-risk group had significant longer DFS (model 2: p=0.005; model 3: pO.001 and model 4. pO.001) and OS time (model 2: p=0.001; model 3: pO.001 and model 4: p<0.001) after five years (Figure 14-16) when compared to the high-risk group of lung ADC patients.

In the case of SCC patients the best model was composed of 5 proteins (RAE1, RRM2, SLC2A1, SRSF1 and STC1) divided in six variables (RAE1-C, RRM2-C, SLC2A1-C, SRSF1-N, STC1-C and STC1-N). The equation was:

Also, after calculating the PI for each patient, we stratified the SCC patients in two subgroups (high and low) according the PI score using the first tertile as the cut off. Kaplan-Meier survival curves and log-rank tests showed statistical differences between the two groups on both DFS (p<0.001) and OS (p=0.002) Figure 17.

This algorithm was selected for validation in a larger cohort (see Example 2) and in the further studies presented below. Other obtained signatures with good prognostic value in SCC also comprise the determination of the expression levels of STC1-C and RAE1-C proteins. The other generated algorithms are the following:

To obtain a graphic representation of these five models, we stratified the PI score in two different subgroups using the median (model 4), the first tertile (model 3 and 5) or the second tertile (model 2 ) as cut-off and represented the survival curves by Kaplan-Meier method. The low-risk group had significant longer DFS (model 2: p= p<0.001 model 3 : p=0.063; model 4: p=0.01 land model 5: p=0.001 1) and OS time (model 2: p=0.030; model 3: p=0.054; model 4: p=0.026; model 5: p=0.010) after five years (Figure 18-21) when compared to the high-risk group of lung SCC patients.

We next tested the prognostic ability of the Pis obtained with the selected models for ADC and SCC, respectively, independently from other clinical and pathological parameters by a multivariate analysis using the subgroups of patients with ADC and SCC. The variables with a p<0.2 in the univariate analysis were introduced in the multivariate analysis. The prognostic indexes derived from all the signatures remained as independent risk factors for the outcome. For ADC patients, the PI was significantly prognostic of the five-year outcome for both DFS (Table 4-7, all the models P<0.05) and OS (Table 8-11, all the models P<0.001) in a univariate Cox proportional hazards analysis.. As expected, stage prognosis was highly significant (p<0.001). The PI score after adjustment for the clinical parameters was evaluated in a multivariate Cox regression analysis. Both the molecular prognostic model and the stage remained as independent predictors of five-year outcome for DFS (all the models P<0.001) and OS (all the models P<0.05). All the results from the univariate and multivariate Cox proportional hazards analysis are summarized in Tables 4 -11. Table 4. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from ADC model 1 and other clinicopathological parameters for DFS in ADC from the MDA cohort

Table 5. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from ADC model 2 and other clinicopathological parameters for DFS in ADC from the MDA cohort

Table 6. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from ADC model 3 and other clinicopathological parameters for DFS in ADC from the MDA cohort

Table 7. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from ADC model 4 and other clinicopathological parameters for DFS in ADC from the MDA cohort

Table 8. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from ADC model 1 and other clinicopathological parameters for OS in ADC from the MDA cohort.

Table 9. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from ADC model 2 and other clinicopathological parameters for OS in ADC from the MDA cohort.

Table 10. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from ADC model 3 and other clinicopathological parameters for OS in ADC from the MDA cohort.

Table 11. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from ADC model 4 and other clinicopathological parameters for OS in ADC from the MDA cohort.

For SCC patients, the PI was significantly prognostic of the five-year outcome for both DFS (Table 12-16, P<0.05) and OS (Table 17-21, P<0.05) in a univariate Cox proportional hazards analysis.. As expected, stage prognosis was highly significant (P<0.05 for DFS and OS). The PI score after adjustment for the clinical parameters was evaluated in a multivariate Cox regression analysis. The molecular prognostic model and the stage remained the most significant predictors of five-year outcome for DFS and OS (all the models P<0.05). All the results from the univariate and multivariate Cox proportional hazards analysis are summarized in Tables 12-21.

Table 12. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from SCC model 1 and other clinicopathological parameters for DFS in SCC from the MDA cohort

Table 13. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from SCC model 2 and other clinicopathological parameters for DFS in SCC from the MDA cohort

Table 14. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from SCC model 3 and other clinicopathological parameters for DFS in SCC from the MDA cohort

Table 15. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from SCC model 4 and other clinicopathological parameters for DFS in SCC from the MDA cohort

Table 16. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from SCC model 5 and other clinicopathological parameters for DFS in SCC from the MDA cohort

Table 17. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from SCC model 1 and other clinicopathological parameters for OS in SCC from the MDA cohort

Table 18. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from SCC model 2 and other clinicopathological parameters for OS in SCC from the MDA cohort

Table 19. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from SCC model 3 and other clinicopathological parameters for OS in SCC from the MDA cohort

Table 20. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from SCC model 4 and other clinicopathological parameters for OS in SCC from the MDA cohort

Table 21. Univariate and multivariate Cox proportional hazards analysis of PI score obtained from SCC model 5 and other clinicopathological parameters for OS in SCC from the MDA cohort

The obtained data confirmed the usefulness of the selected histotype-specific protein-based prognostic signatures to stratify the five-year risk of lung cancer recurrence or/and death in patients with either lung ADC or SCC. Example 2: independent validation of the prognostic signature

Description of the experiment

The protein expression levels of all the genes comprised in both signatures selected as the best model developed in Example 1 were analyzed in tumor samples from an independent cohort of early stage lung cancer patients (CUN-CIBERES cohort) and the models were validated.

Material and methods

Tumor samples (from 116 ADCs and 106 SCC) were collected from consecutive population cohorts surgically treated at Clinica Universidad de Navarra (CUN) and CIBERES network. Inclusion criteria were the same as in example 1. The clinical and pathological data are summarized in Tables 1 and 2. The expression of the seven proteins (BRCA1, QKI, SLC2A1, RAE1, RRM2, SRSF1, y STC1) was evaluated using IHC and semiquantitative method (H- score determination) by two experienced observers. Later, for each patient, the immunostaining scores were applied to the prognostic formulas, previously generated in example 1. We calculated the CH and the survival curves with Kaplan-Meier method, which differences were compared using log-rank test as previously described.

Results

The prognostic models were applied in the CIBERES-CUN patients and prognostic indexes were divided into two groups as previously described. In the CIBERES cohort only overall survival data was available, thus the risk of death was chosen as end point. In both models, the performance was not only reproduced, but it was improved for OS (CH for ADC=0.67 and CH for SCC=0.60). After stratification of the patients into two risk groups, Log-rank test significantly discriminated the patients with high or low risk of death (ADC: P=0.001; Figure 22A and SCC: P= 0.068; Figure 23A)

In the CUN cohort, recurrence data was also available. The results were similar to the ones obtained for OS although the number of patients was low (ADC: n=40 and SQ: n=43). In ADC patients, high PI score tended to shorter DFS (P=0.134; Figure 22B) whilst a significant association was observed between high PI score and reduced OS in SCC patient (P=0.027; Figure 23B).

In summary, the two histotype specific protein-based signatures identified in example 1 were validated in an independent cohort of ADC or SCC patients, respectively.

Example 3: risk stratification in stage I-II patients

Description of the experiment

The utility of the Pis was verified in patients with stage I and II tumors, a group of patients where tumor stage remains insufficient to discriminate high and low risk patients.

Material and methods

We performed a sub-analysis in stages I-II patients from MD Anderson for both histologies, ADC and SCC. These subsets of patients were composed of 192 patients with ADC and 93 with SCC. We calculated the CH and the survival curves with Kaplan-Meier method, which differences were compared using log-rank test as previously described.

Results

Both prognostic models, the ADC and the SCC models, showed a significant prognostic ability (Figura 24A-B and 25A-B). Also, the independent prognostic values of both signatures were evaluated in stage I-II patients. Independently to the stage, the PI derived from the prognostic models remained as independent risk factor for the disease-free survival also in this subgroup of patients [ADC model: DFS (P=0.004, HR=2.51 (95% CI, 1.34-4.68) and SCC model: DFS (P=0.011, HR=2.24 (95% CI, 1.20-4.18). For OS, only the ADC model demonstrated independent prognostic value (P=0.001, HR=3.18 (95% CI, 1.62-6.26)]. The results of the univariate and multivariate Cox proportional hazards analysis are summarized in Tables 22-25.

Table 22. Univariate and multivariate Cox proportional hazards analysis of PI score and other clinicopathological parameters for DFS in stages I-II ADC patients from MDA cohort.

Table 23. Univariate and multivariate Cox proportional hazards analysis of PI score and other clinicopathological parameters for OS in in stages I-II ADC patients from MDA cohort

Table 24. Univariate and multivariate Cox proportional hazards analysis of PI score and other clinicopathological parameters for DFS in stages I-II SCC patients from MDA cohort.

Table 25. Univariate and multivariate Cox proportional hazards analysis of PI score and other chnicopathological parameters for OS in stages I-II SCC patients from MDA cohort.

The models were also validated in stage I-II patients from the CUN-CIBERES cohort. The log- rank P values for OS were statistically significant in both histologies (ADC: P<0.001, Figure 26A and SCC: P=0.031, Figure 27A). As before was mentioned, in the CUN cohort, recurrence data was also available. The results were similar to the ones obtained for OS although the number of patients was low (ADC: n=39 and SQ: n=43). In ADC patients, a significant association was observed between high PI score and reduced DFS (P=0.034; Figure 26B) whilst high PI score tended to shorter DFS in SCC patients (P=0.162; Figure 27B). In conclusion, the prognostic ability of both protein-based models were demonstrated and validated in stage I-II lung cancer patients.

Example 4: clinical utility of the prognostic models

Description of the experiment The benefit of combining the pathological stage with the molecular prognostic models was analyzed to assess the medical applicability of the models in MDA cohort.

Material and methods

Tumors were classified according to the 7^th TNM edition for lung tumors stratification (Goldstraw P et al, 2007). Briefly, staging is determined by three components: the primary tumor (T), the lymph nodes (N) and the metastasis (M). The T component is defined by tumor size, tumor location the involved structures or the effects of the tumor growth and has seven categories. The N component is defined by the absence or presence and location of the involved nodes and has five categories and the M component has two different categories defined by the absence or presence and location of the metastasis (Table 26). TNM subsets of similar prognosis are grouped in 7 different stages. Clinical utility of our models was tested by comparing the likelihood ratio of the stage itself to the addition of the molecular model (PI) through a bivariate Cox analysis in the MDA cohort. Then, we developed a new combined prognostic model (CPI) by adding the pathological stage and the molecular data. We calculated the CH coefficient of the new models and calibrated them with survival curves as previously described. Table 26. Definition for T, N, and M descriptors

and the fluid is nonbloody and is not an exudate. Where these elements and clinical judgment dictate that the effusion is not related to the tumor, the effusion should be excluded as a staging element and the patient should be classified as Tl, T2, T3, or T4.

Results

The prognostic capability improved in all cases after addition of the Pis variables. All the prognostic models were complementary to the stage and were able to add additional prognostic information to the patients.

In the case of ADC, stage alone was a highly significant prognostic factor of both DFS (P<0.001) and OS (P=0.001, Figure 28A). However, the likelihood ratio significantly increased after adding the molecular information (PI) (P<0.001 both for DFS and OS). This improvement showed that the molecular model complements the stage, adding very valuable prognostic information for the patients. We next performed a Cox regression analysis to develop a new prognostic model combining stage and the molecular model using patients from stage I to IIIA. This new signature was named combined prognostic index (CPI), and the formula for ADC models is:

CPI = 1.090 x PI + n; where n is a coefficient that changes for each stage (IA, n=0; IB, n=0.507; IIA, n=1.099; IIB, n=1.583 and IIIA, n=1.450).

We applied the CPI model to the MDA patients and the new index was assessed to each patient. Then, the CPI median was used to stratify the risk of outcome into two groups (high/low). The differences according to the five-year survival between two groups were greater in both DFS and OS (P<0.001). Thus, low CPI score was clearly associated with a better outcome (Figure 28B-C). Similar results were observed for the SCC model. After combining both variables: the stage and the molecular models, the likelihood ratio significantly increased after adding the molecular information (PI) (P<0.001 both for DFS and OS) Figure 29 A. We developed a new combined prognostic model, In the case of the SCC model, the formula is: CPI = 1.072 x PI + n; where n is a coefficient that change for each stage (IA, n=0; IB, n= - 0.438; IIA, n= - 0.025; IIB, n=0.243 and IIIA, n=1.560).

We applied the CPI model to the MDA patients and a new index was assessed to each patient. Then we stratified the risk of outcome of SCC patients into two groups (high/low). The differences according to the five-year survival between two groups were highly significant in both DFS and OS (P<0.001). Thus, high CPI score was clearly associated with a worse outcome (Figure 29B-C).

In conclusion, our CPI model stratifies those patients at high risk of recurrence or death more effectively than the conventional TNM classifier. Indeed, the combination of both parameters, the pathological stage and the molecular prognostic index improves the prognostic ability of the conventional stage system.

Example 5: independent validation of the combined prognostic signatures

Description of the experiment

The benefit of the combined prognostic models was validated in an independent cohort of lung cancer patients from CUN-CIBERES.

Material and methods The CUN-CIBERES cohort of patients, previously described in example 2, was used for this study.

Results

As expected, the death risk stratification for the entire cohort was significantly increased when both parameters, the prognostic index and the stage, were analyzed together (ADC model: P<0.001; Figure 30A and SCC model: P=0.003; Figure 31A). As it was mentioned before, in the CIBERES cohort only overall survival data was available. In order to validate the prognostic ability of the combined models for DFS, only the cohort from the CUN was analyzed (n=40). The differences according to the five-year survival between two groups (high CPI/low CPI) were highly significant in the ADC model (P=0.001; Figure 30B) and tends to significance in the SCC model (p=0.056; Figure 31B). The prognostic ability of the combined model (CPI) was validated in an independent cohort of patients from CIBERES-CUN.

Example 6: Predictive value of protein-based models for the benefit of the adjuvant therapy

Description of the experiment

In NSCLC, stage I is the most controversial subgroup in terms of adjuvant treatment effectiveness. At this point, we analyzed whether our model could identify which patients obtain a benefit from a platinum agent based chemotherapy adjuvant treatment. Material and methods

We stratified the MDA cohort into two different subgroups according to the molecular prognostic models (Pis) (low/high according to the median). In each subgroups, we analyzed by log rank and Kaplan-Meier the clinical outcome after stratification of the patients according to presence/absence of adjuvant therapy. This analysis was performed for the two main histologies, ADC and SCC.

Results

In both models (ADC and SCC), we observed an association between patients presenting higher molecular prognostic index and longer OS in patients who received platinum-based chemotherapy after surgery) (ADC: P=0.017; Figure 32A and SCC: P= 0.037; Figure 33A). However this association was not observed in patients with a low PI (ADC: p=0.208; Figure 32A and SCC: 0.335; Figure 33A). For DFS, a trend was also observed for both histologies but significance was not reached (ADC: P=0.055; Figure 32B and SCC: P=0.083; Figure 33B). No association between DFS time and postsurgery treatment was observed in low PI patients (ADC: P=0.403; Figure 32B and SCC: P=0.949; Figure 33B).

In conclusion, our models are able to select those stage I-II patients who could obtain a benefit from an adjuvant therapy, in particular from platinum-based adjuvant chemotherapy treatment, in both histological subtypes. Bibliography

1. Aberle DR, Adams AM, Berg CD, et al. National Lung Screening Trial Research Team. Reduced lung-cancer mortality with low-dose computed tomographic screening.N Engl J Med 2011;365:395-409).

2. Siegel RL, Miller KD, Jemal A. Cancer Statistics, 2017. CA Cancer J Clin. 2017 Jan;67(l):7- 30.

3. Arriagada R, Bergman B, Dunant A, Le Chevalier T, Pignon JP, Vansteenkiste J, et al. Cisplatin-based adjuvant chemotherapy in patients with completely resected non-small-cell lung cancer. N Engl J Med 2004 Jan 22;350(4):351-360.

4. Winton T, Livingston R, Johnson D, Rigas J, Johnston M, Butts C, et al. Vinorelbine plus cisplatin vs. observation in resected non-small-cell lung cancer. N Engl J Med 2005 Jun 23;352(25):2589-2597.

5. Douillard JY, Rosell R, De Lena M, Carpagnano F, Ramlau R, Gonzales-Larriba JL, et al. Adjuvant vinorelbine plus cisplatin versus observation in patients with completely resected stage IB-IIIA non-small-cell lung cancer (Adjuvant Navelbine International Trialist Association [ANITA]): a randomised controlled trial. Lancet Oncol 2006 Sep;7(9):719-727.

6. Strauss GM, Herndon JE,2nd, Maddaus MA, Johnstone DW, Johnson EA, Harpole DH, et al. Adjuvant paclitaxel plus carboplatin compared with observation in stage IB non-small-cell lung cancer: CALGB 9633 with the Cancer and Leukemia Group B, Radiation Therapy Oncology Group, and North Central Cancer Treatment Group Study Groups. J Clin Oncol 2008 Nov 1;26(31):5043- 5051.

7. Pignon JP, Tribodet H, Scagliotti GV, Douillard JY, Shepherd FA, Stephens RJ, et al. Lung adjuvant cisplatin evaluation: a pooled analysis by the LACE Collaborative Group. J Clin Oncol 2008 Jul 20;26(21):3552-3559.

8. (20) Pisters KM, Evans WK, Azzoli CG, Kris MG, Smith CA, Desch CE, et al. Cancer Care Ontario and American Society of Clinical Oncology adjuvant chemotherapy and adjuvant radiation therapy for stages I-IIIA resectable non small-cell lung cancer guideline. J Clin Oncol 2007 Dec 1;25(34):5506-5518.

9. Goldstraw P, Crowley J, Chansky K, Giroux DJ, Groome PA, Rami-Porta R, Postmus PE, Rusch V, Sobin L; International Association for the Study of Lung Cancer International Staging

Committee; Participating Institutions. The IASLC Lung Cancer Staging Project: proposals for the revision of the TNM stage groupings in the forthcoming (seventh) edition of the TNM Classification of malignant tumours. J Thorac Oncol. 2007 Aug;2(8):706-14.

10. Goldstraw P, Chansky K, Crowley J, Rami-Porta R, Asamura H, Eberhardt WE, Nicholson AG, Groome P, Mitchell A, Bolejack V; International Association for the Study of Lung Cancer Staging and Prognostic Factors Committee, Advisory Boards, and Participating Institutions; International Association for the Study of Lung Cancer Staging and Prognostic Factors Committee Advisory Boards and Participating Institutions. The IASLC Lung Cancer Staging Project: Proposals for Revision of the TNM Stage Groupings in the Forthcoming (Eighth) Edition of the TNM Classification for Lung Cancer. J Thorac Oncol. 2016 Jan;l 1(1):39-51

11. Coate LE, John T, Tsao MS, Shepherd FA. Molecular predictive and prognostic markers in non-small-cell lung cancer. Lancet Oncol 2009 Oct; 10(10): 1001-1010.

12. Ferte C, Andre F, Soria JC. Molecular circuits of solid tumors: prognostic and predictive tools for bedside use. Nat Rev Clin Oncol 2010 Jul;7(7):367-380.

13. Lin J, Beer DG. Molecular predictors of prognosis in lung cancer. Ann Surg Oncol 2012 Feb;19(2):669-676.

14. Zhu CQ, Tsao MS. Prognostic markers in lung cancer: is it ready for prime time? Transl Lung Cancer Res. 2014 3(3): 149-58.

15. Lau SK, Boutros PC, Pintilie M, Blackhall FH, Zhu CQ, Strumpf D, et al. Three-gene prognostic classifier for early- stage non small-cell lung cancer. J Clin Oncol 2007 Dec 10;25(35):5562-5569.

16. Bianchi F, Nuciforo P, Vecchi M, Bernard L, Tizzoni L, Marchetti A, et al. Survival prediction of stage I lung adenocarcinomas by expression of 10 genes. J Clin Invest 2007 Nov;117(l l):3436- 3444.

17. Skrzypski M, Jassem E, Taron M, Sanchez JJ, Mendez P, Rzyman W, et al. Three-gene expression signature predicts survival in early-stage squamous cell carcinoma of the lung. Clin

Cancer Res 2008 Aug l;14(15):4794-4799.

18. Raz DJ, Ray MR, Kim JY, He B, Taron M, Skrzypski M, et al. A multigene assay is prognostic of survival in patients with early-stage lung adenocarcinoma. Clin Cancer Res 2008 Sep l;14(17):5565-5570.

19. Kratz JR, He J, Van Den Eeden SK, Zhu ZH, Gao W, Pham PT, et al. A practical molecular assay to predict survival in resected non-squamous, non-small-cell lung cancer: development and international validation studies. Lancet 2012 Mar 3;379(9818):823-832

20. Wistuba II, Behrens C, Lombardi F, Wagner S, Fujimoto J, Raso MG, Spaggiari L, Galetta D, Riley R, Hughes E, Reid J, Sangale Z, Swisher SG, Kalhor N, Moran CA, Gutin A, Lanchbury JS, Barberis M, Kim ES. Validation of a proliferation-based expression signature as prognostic marker in early stage lung adenocarcinoma. Clin Cancer Res. 2013 Nov 15 ; 19(22) : 6261-71.

21. Travis WD, World Health Organization., International Agency for Research on Cancer., et al: Pathology and genetics of tumours of the lung, pleura, thymus and heart. Lyon. Oxford, IARC Press. Oxford University Press (distributor), 2004

22. Goldstraw P, International Association for the Study of Lung Cancer.: Staging manual in thoracic oncology. Orange Park, FL, Editorial Rx Press, 2009 23. Altman DG, McShane LM, Sauerbrei W, et al: Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK): explanation and elaboration. PLoS Med 9:el001216, 2012

Claims

1. An in vitro method for determining the prognosis of a subject having lung adenocarcinoma, wherein said method comprises: a) determining in a biological sample isolated from said subject the protein expression levels of nuclear QKI (QKI-N) and membrane SLC2A1 (SLC2A1-MB), optionally, further determining in said biological sample the protein expression levels of at least one, two, three, four or five additional markers selected from the group consisting of:

-nuclear BRCA1 (BRCA1-N),

-cytoplasmic QKI (QKI-C),

-nuclear STC1 (STC1-N),

-nuclear CDC6 (CDC6-N), and

-nuclear SIRT2 (SIRT2-N); b) calculating a combined score from the protein expression levels of the markers determined in the biological sample as defined in step a); and c) classifying the subject as having good prognosis or poor prognosis based on the combined score.

2. The prognosis method according to claim 1, wherein in step c) said method comprises comparing the combined score in the subject sample with a reference combined score; and wherein an increase of the combined score in the subject sample with regard to said reference combined score is indicative of poor prognosis.

3. The prognosis method according to any of claims 1 or 2, wherein the combined score calculated in b) is proportional to the expression levels of QKI-C, SLC2A1-MB, BRCA1-N, SIRT2-N and is inversely proportional to the expression levels of the QKI-N, STC1-N and CDC6-N, wherein the higher the score, the worst the prognosis.

4. The prognosis method according to any of claims 1 to 3, wherein the term prognosis refers to disease free survival (DFS) and/or overall survival (OS).

5. The prognosis method according to any of claims 1 to 4, wherein said method comprises determining in step a) the protein expression levels of :

- QKI-N, SLC2A1-MB and BRCA1-N; or

- QKI-N, SLC2A1-MB and QKI-C.

6. The prognosis method according to any of claims 1 to 5, wherein said method comprises determining in step a) the protein expression levels of:

- QKI-N, SLC2A1-MB, BRCA1-N and QKI-C; or

- QKI-N, SLC2A1-MB, BRCA1-N, QKI-C and STC1-N; or

- QKI-N, SLC2A1-MB, BRCA1-N, STC1-N, CDC6-N and SIRT2-N.

7. The prognosis method according to any of claims 1 to 6, wherein said method comprises determining in step a) the protein expression levels of:

QKI-N, SLC2A1-MB, BRCA1-N and QKI-C.

8. The prognosis method according to any of claims 1 to 7, wherein said biological sample isolated from the subject is a tumor biopsy sample, preferably obtained from a resected tumor.

9. The prognosis method according to any of claims 1 to 8, wherein in step a) the protein expression levels are determined by a method comprising:

a. obtaining from the biological sample protein extracts of the subcellular location of the target marker; and

b. determining the target protein expression levels in said protein extracts.

10. The prognosis method according to claim 9, comprising:

a. incubating nuclear protein extracts with an affinity reagent for QKI,

b. incubating membrane protein extracts with an affinity reagent for SLC2A1 ; and c. optionally, incubating target protein extracts with an affinity reagent for each of the at least one, two, three, or the four additional proteins defined in step a).

11. The prognosis method according to any of claims 1 to 10, wherein in step a) protein expression levels are determined by immunohistochemistry.

12. The prognosis method according to claims 1 to 8 or 11, wherein determining the protein expression levels in step a) comprises incubating the biological sample isolated from said subject with:

a) an affinity reagent for QKI,

b) an affinity reagent for SLC2A1 , and

c) optionally, an affinity reagent for each of the at least one, two, three, or the four additional proteins defined in step a).

13. The prognosis method according to any of claims 11 or 12, wherein said sample is a fixed biopsy sample, preferably a formalin fixed-paraffin embedded sample.

14. The prognosis method according to any of claims 11 to 13, wherein one or more fixed cancer cells, preferably formalin- fixed paraffin embedded cancer cells, are used as low and/or high expression controls.

15. The prognosis method according to any of claims 13 or 14, wherein said sample, and optionally said control cells, is submitted to an antigen retrieval process prior to incubation with the affinity reagent.

16. The method according to claim 15, wherein the antigen retrieval process comprises the treatment with a citrate buffer when the antigen to be retrieved is selected from the group consisting of BRCA1, SIRT2, SLC2A1 and STC1; and comprises the treatment with a EDTA buffer when the antigen to be retrieved is selected from the group consisting of CDC6 and QKI.

17. The method according to any of claims 15 or 16, wherein the antigen retrieval process is conducted at a temperature of 90°C to 100°C during 10 to 30 minutes, preferably at 95°C during 20 minutes.

18. The prognosis method according to any of claims 11 to 17, wherein the protein expression levels are expressed as H-Score values, and wherein H-Score values are determined by adding the products of the percentage of cells stained with a given intensity (0-100) by the staining intensity using the following formula: ∑ intensity grade x % stained cells.

19. The prognosis method according to claim 18, wherein the staining intensity has a value between 1 and 3, wherein l=weak, 2=moderate and 3=strong staining intensity, and H-Score values are determined by using the following formula:

H-Score = 1 x (% of cells with weak staining) + 2 x (% of cells with moderate staining) + 3 x (% of cells with strong staining).

20. The prognosis method according to any of claims 18 or 19, wherein the combined score is calculated as the sum of the products of the standardized beta coefficients for each marker obtained in a regression analysis and the H-Score value.

21. The prognosis method according to claim 20, wherein said beta coefficients are positive for markers selected from the group consisting of QKI-C, SLC2A1-MB, BRCA1-N, SIRT2-N and negative for markers selected from the group consisting of QKI-N, STC1-N and CDC6- N.

22. The prognosis method according to any of claims 18 to 21, wherein the combined score is obtained by using the following formula:

- 0.004 x H-Score_QKi-N + 0.005 x H-Score_SLc2Ai-MB + 0.006 x H-Score_BRCAi-N + 0.006 x H-Score_QKi-c.

23. The prognosis method according to any of claims 1 to 22, wherein further to step a) said method comprises:

Al) determining in said sample the stage of lung adenocarcinoma according to the TNM classification of lung tumors; and

b) calculating a combined score from the protein expression levels of the markers determined in the biological sample as defined in step a) and the stage of TNM classification as defined in step Al); and

c) classifying the subject as having good prognosis or poor prognosis based on the combined score obtained in b).

24. The prognosis method according to any of claims 1 to 23, wherein said method is a computer-implemented method.

25. A data-processing apparatus comprising means for carrying out the steps of a method of claim 24.

26. A computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the steps of the method of claim 24.

27. A computer-readable storage medium having stored thereon a computer program according to claim 26.

28. In vitro method for selecting those subjects having lung adenocarcinoma who are expected to benefit from an adjuvant treatment, wherein said method comprises:

a) classifying said subject according to the prognosis method as defined in any of claims 1 to 24;

b) selecting for administration of an adjuvant treatment a subject classified in step a) as having poor prognosis.

29. The method according to claim 28, wherein said adjuvant treatment comprises the administration of a platinum anticancer agent.

30. In vitro method for predicting the efficacy of an adjuvant chemotherapy in a subject having lung adenocarcinoma, wherein said method comprises:

a) classifying said subject according to the prognosis method as defined in any of claims 1 to 24; and

b) predicting the efficacy of said adjuvant chemotherapy according to the prognosis classification;

wherein classification of the subject in step a) as having poor prognosis is indicative of increased efficacy of the adjuvant chemotherapy.

31. The method according to claim 30, wherein said adjuvant chemotherapy comprises the administration of a platinum anticancer agent, and wherein a poor prognosis is indicative of increased efficacy of the platinum anticancer agent.

32. In vitro method for selecting an adjuvant treatment for a subject having lung adenocarcinoma, wherein said method comprises:

a) classifying said subject according to the prognosis method as defined in any of claims 1 to 24; and

b) selecting an adjuvant treatment according to the prognosis classification;

wherein when the subject is classified in step a) as having poor prognosis then adjuvant chemotherapy is selected as adjuvant treatment.

33. The method according to claim 32, wherein when the subject is classified as having poor prognosis then an adjuvant treatment comprising the administration of a platinum anticancer agent is selected.

34. The method according to any of claims 29, 31 or 33, wherein the platinum anticancer agent is cisplatin and/or carboplatin.

35. The method according to any of claims 1 to 24 and 28 to 34, wherein said adenocarcinoma is resectable lung adenocarcinoma.

36. The method according to any of claims 1 to 24 and 28 to 35, wherein said lung adenocarcinoma is classified as stage I or II lung adenocarcinoma according to the TNM classification of lung tumors, preferably is classified as stage I adenocarcinoma.

37. The method according to any of claims 1 to 24 and 28 to 36, wherein said subject has been submitted to tumor resection surgery and has not received any neoadjuvant treatment.

38. The method according to any of claims 1 to 24 and 28 to 37, wherein said subject is a human subject.

39. A kit suitable for determining the levels of QKI, SLC2A1 and optionally, at least one of the additional protein markers defined in step a) in a method for the prognosis of a subject having lung adenocarcinoma according to any of claims 1 to 24, wherein said kit comprises:

- an affinity reagent for QKI; and

- an affinity reagent for SLC2A1; - optionally, further comprising an affinity reagent for each of the at least one, two, three or the four additional proteins of the additional protein markers defined in step a) of the method of any of claims 1 to 7;

- optionally, further comprising cancer cells to be used as low and/or high expression controls;

- optionally, further comprising instructions for the use of said reagents in determining said protein expression levels in a biological sample isolated from a subject.

40. Use of a kit in a method for the prognosis of a subject having lung adenocarcinoma according to any of claims 1 to 24, wherein said kit comprises:

- a reagent for determining the protein expression levels of QKI; and

- a reagent for determining the protein expression levels of SLC2A1 ;

- optionally, further comprising a reagent for each of the at least one, two, three, or the four additional proteins of the additional protein markers defined in step a) of the method of any of claims 1 to 7 for determining the protein expression levels thereof;

- optionally, further comprising cancer cells to be used as low and/or high expression controls;

- optionally, further comprising instructions for the use of said reagents in determining said proteins expression levels in a biological sample isolated from a subject.

41. Use of a kit according to claim 40, wherein said kit is as defined in claim 39.

42. Kit according to claim 39 or use of a kit according to any of claims 40 or 41, wherein said affinity reagent is an antibody.

43. An in vitro method for determining the prognosis of a subject having lung adenocarcinoma, wherein said method comprises: a. determining in a biological sample isolated from said subject the protein expression levels of nuclear QKI (QKI-N) and membrane SLC2A1 (SLC2A1-MB), optionally, further determining in said biological sample the protein expression levels of at least one, two, three, four or five additional markers selected from the group consisting of:

-nuclear BRCA1 (BRCA1-N),

-cytoplasmic QKI (QKI-C), -nuclear STC1 (STC1-N),

-nuclear CDC6 (CDC6-N), and

-nuclear SIRT2 (SIRT2-N); b. calculating, using a computer, a combined score from the protein expression levels of the markers determined in the biological sample as defined in step a);

wherein the protein expression levels are calculated as H-Score values, and wherein the H-Score values are determined by adding the products of the percentage of cells stained with a given intensity (0-100) by the staining intensity using the following formula: H- Score =∑ intensity grade x % stained cells; and c. classifying the subject as having good prognosis or poor prognosis based on the combined score.

44. A plurality of biomarkers for predicting disease progression or survival of a subject with lung adenocarcinoma (ADC), the plurality of biomarkers comprising nuclear QKI (QKI-N), membrane SLC2A1 (SLC2A1-MB), nuclear BRCA1 (BRCA1-N) and cytoplasmic QKI (QKI-C).