WO2017083513A1 - Methods to classify multiple myeloma - Google Patents

Abstract

The present disclosure provides a classification diagnostic, termed TC10, that uses gene expression to assess the primary aberration for individuals with multiple myeloma.

Description

METHODS TO CLASSIFY MULTIPLE MYELOMA

CROSS REFERENCE TO RELATED APPLICATIONS

[0001 ] This application claims the benefit of U.S. Provisional Application number 62/253,537, filed November 10, 2015, the disclosure of which is hereby incorporated by reference in its entirety.

FIELD OF THE INVENTION

[0002] The present disclosure provides a classification diagnostic, termed TC10, that uses gene expression to assess the primary aberration for individuals with multiple myeloma.

BACKGROUND OF THE INVENTION

[0003] Analysis of myeloma plasma cells using cytogenetics and FISH formed the basis of the initial classification and genetic subgrouping of myeloma (MM). These traditional investigations defined a broad based classification system dependent upon either the presence of a balanced translocation into the immunoglobulin heavy chain (IGH) locus (40%) or the presence of hyperdiploidy (HRD), characterized by gain of the odd numbered chromosomes. Further analysis of the translocated cases identified the overexpression of a D group cyclin as a key molecular aberration. A number of genomic classification systems have, therefore, been proposed, each with limitations. However, there is a need in the art for an ideal classification system that would identify clinically relevant subgroups with distinct biology that can be recognized in the clinic using robust diagnostic techniques. Importantly, such a classification should offer clinically useful information about prognosis and response to specific therapies and should split the disease into meaningfully sized subgroups.

SUMMARY OF THE INVENTION

[0004] In an aspect, the disclosure provides a classification system for multiple myeloma, the classification system comprising the following 10 classes:

t(1 1 ; 14) CD20/PAX5+, t(1 1 ; 14) CD20/PAX5-, t(1 1 ; 14) D1 -HRD:RRAS2+, t(1 1 ; 14) D1 - l HRD:RRAS2- t(11;14) D2:RRAS2+, t(11;14) D2:RRAS2- t(4;14), t(14;16), t(14;20) and t(6;14).

[0005] In another aspect, the disclosure provides a method of classifying multiple myeloma in a subject diagnosed with multiple myeloma, the method

comprising: detecting an aberration associated with t(11;14) CD20/PAX5+, t(11;14) CD20/PAX5-, t(11;14) D1-HRD:RRAS2+, t(11;14) D1-HRD:RRAS2- t(11;14)

D2:RRAS2+, t(11;14) D2:RRAS2- t(4;14), t(14;16), t(14;20) and t(6;14) in a biological sample obtained from a subject; and classifying a subject into one of the 10 classes selected from the group consisting of t(11 ;14) CD20/PAX5+, t(11 ;14) CD20/PAX5- t(11;14) D1-HRD:RRAS2+, t(11;14) D1-HRD:RRAS2-, t(11;14) D2:RRAS2+, t(11;14) D2:RRAS2- t(4;14), t(14;16), t(14;20) and t(6;14) based on the aberration detected.

[0006] In still another aspect, the disclosure provides a method to prognose a subject diagnosed with multiple myeloma, the method comprising:

classifying the subject into one of the 10 classes selected from the group consisting of t(11;14) CD20/PAX5+, t(11;14) CD20/PAX5-, t(11;14) D1-HRD:RRAS2+, t(11;14) D1- HRD:RRAS2- t(11;14) D2:RRAS2+, t(11;14) D2:RRAS2- t(4;14), t(14;16), t(14;20) and t(6; 14); and using a risk stratifier to further prognose the subject.

[0007] In still yet another aspect, the disclosure provides a method to determine treatment of a subject diagnosed with multiple myeloma, the method comprising: classifying the subject into one of the 10 classes selected from the group consisting of t(11 ; 14) CD20/PAX5+, t(11 ; 14) CD20/PAX5-, t(11 ; 14) D1 -HRD: RRAS2+, t(11;14) D1-HRD:RRAS2- t(11;14) D2:RRAS2+, t(11;14) D2:RRAS2- t(4;14), t(14;16), t(14;20) and t(6; 14); and administering treatment based on the classification of the subject.

BRIEF DESCRIPTION OF THE FIGURES

[0008] The application file contains at least one drawing executed in color. Copies of this patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee. [0009] FIG. 1 depicts an ordered bar plot of gene expression across TC10 subgroups. Spiked expression accompanies each respective IGH translocation, and every subgroup has a unique relationship with the D-group cyclins. The t(4; 14) subgroup has one of the most distinct expression patterns with elevated levels of WHSC1, often accompanied by FGFR3, DSG2, and BACE2. The t(14; 16) and t(14;20) have elevated expression within their respective MAF-gene and the related genes of NUAK1, ARID5A, and ITGB7. The t(6; 14) cases, although rare, uniquely over express CCND3 and USP49 while lacking expression in CCND1 and CCND2. The t(1 1 ; 14) cases uniquely overexpress CCND1, SLC8A1, and BANK1, and can be subdivided according to CD20 expression (MS4A1). The D1 subgroup is most distinguished by intermediate expression of CCND1, low expression of CCND2, and elevated expression of hyperdiploid-associated genes such as ISL2, TNFSF10, and CCRL2. The D2 subgroup expresses many of the same genes as the D1 subgroup; however, these cases over express CCND2 and genes associated with 1 q: S100A4 and NES. Each row represents log2 expression of probes ranging from 250 to 50,000 in raw expression values for 1355 NDMM UAMS cases.

[0010] FIG. 2A, FIG. 2B, FIG. 2C, FIG. 2D, FIG. 2E and FIG. 2F depict overall survival and response rate across TC10 subgroups. (FIG. 2A, FIG. 2B) Kaplan- Meier estimates of overall survival (FIG. 2A) and competing risk plot of response rates (FIG. 2B) for the D1 -HRD and D2 subgroups shows that both D1 -HRD subgroups are slower to respond to treatment than D2 subgroups, and that the D2:RRAS2- cases have the least favorable prognosis with a five-year survival rate of 68.1 %. (FIG. 2C, FIG. 2D) Kaplan-Meier estimates of overall survival (FIG. 2C) and competing risk plot of response rates (FIG. 2D) for the t(1 1 ; 14) and t(6; 14) cases. Curves for the t(1 1 ; 14) and t(6; 14) cases reveal a highly differential response to therapy as the CD20/PAX5+ cases are the slowest to respond to therapy, while the t(6; 14) and t(1 1 ;14):CD20/PAX5- cases are among the fastest. (FIG. 2E, FIG. 2F) Kaplan-Meier estimates of overall survival (FIG. 2E) and competing risk plot of response rates (FIG. 2F) for the t(4; 14), t(14; 16) and t(14;20) cases. Curves for the t(4;14), t(14; 16) and t(14;20) cases reveal poorer overall outcomes than those seen in subgroups above. Also, t(14;20) cases are slower to respond than t(4; 14) and t(14;16) cases.

[001 1 ] FIG. 3A, FIG. 3B, FIG. 3C and FIG. 3D depict a boxplot of mutational load and bar plot of frequency of mutations within TC10 subgroups for F1 mutation panels. (FIG. 3A) Boxplot of mutational load reveals that t(14; 16) has the highest median mutational load and the D1 -HRD and D2 RRAS2+ subgroups have lower mutational loads than their RRAS2- counterparts (subgroup adjusted p-value < 0.001 ). (FIG. 3B, FIG. 3C, FIG. 3D) Investigation of mutation within key genes revealed significantly increased frequency of RAS/MAPK mutations within the RRAS2- subgroups compared to their respective RRAS2+ counterparts (subgroup-adjusted p- value < 1 e-13). This disparity is greatest for the D1 cases as 76.0% of all D1 :RRAS2- cases have a RAS/MAPK mutation compared to just 23.3% of D1 :RRAS2+ cases.

Other associations of specific mutations with TC10 subgroups includes an excess of FAM46C mutations in the D1 -HRD:RRAS2- subgroup and the t(4;14) subgroup, ATM mutations within the t(14; 16) subgroup, and ATR mutations within the t(14;20) subgroup. In addition, CCND1, FGFR3, and MAF had higher levels of mutation in the appropriate cytogenetic TC10 subgroup. Values represent percentage of cases within TC10 subgroups identified with mutation by gene according to F1 panel performed on 552 cases.

[0012] FIG. 4A, FIG. 4B, FIG. 4C, FIG. 4D, FIG. 4E and FIG. 4F depict

Kaplan-Meier overall survival curves for vFISH adverse lesions compared to GEP70. (FIG. 4A, FIG. 4B, FIG. 4C) Association of the combined number of vFISH and TC defined adverse lesions [among 1 p- 1 q+, 17p-, t(4; 14), t(14; 16), and t(14;20)] with outcome. The addition of adverse lesions directly impacts survival, and this trend is seen across all data sets: UAMS (FIG. 4A); HOVON-65/GMMG-HD4 (FIG. 4B): and MRC-IX (FIG. 4C). (FIG. 4D, FIG. 4E, FIG. 4F) GEP70 HR model applied and validated across UAMS (FIG. 4D), MRC-IX (FIG. 4E), and HOVON-65/GMMG-HD4 (FIG. 4F) data sets. Overall, GEP70 has a greater association with survival than vFISH model (A), and reveals the increased power of the GEP70 to identify HR than FISH alone. [0013] FIG. 5A and FIG. 5B depict dysregulation of the G1 S cell cycle checkpoint - The central role of the D group cyclins in myeloma pathogenesis is preserved in the TC10 classification. Cell cycle progression from G1 to S phase requires the accumulation of complexes of the D cyclins (CCND1 -3) with the cyclin dependent kinases (CDK4/6). These complexes result in the phosphorylation of retinoblastoma tumor suppressor protein (Rb) resulting in its inactivation and thus allowing transcription of genes important for S phase progression via the E2F family of transcription factors. One of these genes encodes cyclin E (CCNE2) which, in complex with CDK2, is required for transition from G1 to S phase. The INK4 proteins CDKN2A-D act as a negative regulator of CCND-CDK4/6 complexes. The Cip/Kip proteins

CDKN1 A-C act predominantly as a negative regulator of the CCNE2-CDK2 complex but have also been described to activate CCND-CDK4/6 complexes. (FIG. 5A) Association of TC10 subgroups (dark blue background, white text) with dysregulated expression of each D cyclin and how each subgroup therefore contributes to disordered cell cycle progression via the G1 S checkpoint. Other previously identified points of dysregulation within the checkpoint in myeloma are identified in black text e.g. DNA methylation of the genes encoding CDKN2A and CDKN2B and mutations of RB1 . (FIG. 5B) Expression plot demonstrates the relative expression of each of the genes involved in the G1 S checkpoint across the TC10 subgroups in the 1336 NDMM (JAMS patients in this study. Of note CDK6 and CCNE2 have lower expression in the CD20+ t(1 1 ; 14) cases and higher expression in the t(14; 16) and t(14;20) subgroups.

[0014] FIG. 6A, FIG. 6B, FIG. 6C, FIG. 6D, FIG. 6E and FIG. 6F depict paired GEP expression with iFISH translocations. Scatterplots reveal that iFISH positivity for t(4; 14) (FIG. 6A, FIG. 6B), t(1 1 ;14) (FIG. 6C, FIG. 6D), and t(14; 16) (FIG. 6E, FIG. 6F) corresponds overwhelmingly with GEP probes associated with these translocations for MRC-IX and HOVON data sets. This correspondence serves as a foundation for establishment of the TC10 as proxy for iFISH translocation analyses.

[0015] FIG. 7 depicts an ordered bar plot visualizing distribution of key genes across the unsupervised six-group clusters. Sparse k-means clustering revealed an optimal six-cluster model based on 90 probes (Table 4). These six-group designations reveal clusters with similar gene expression signatures, and these signatures reflect Bergsagel's TC subgroups with D group cyclin, hyperdiploidy, and Ig translocations primarily dividing all cases.

[0016] FIG. 8A, FIG. 8B and FIG. 8C depict CCND1 and CCND2 expression across main TC subgroups. Scatterplot of GEP expression for CCND1 (20871 1_s_at) and CCND2 (200953_s_at) probes for the UAMS (FIG. 8A), MRC-IX (FIG. 8B), and HOVON-65/GMMG-HD4 (FIG. 8C) data sets. Note the strong separation of the primary TC subgroups, as well as the lack of D1 +D2 or "none" subgroups included in the original TC classification.

[0017] FIG. 9A, FIG. 9B, FIG. 9C and FIG. 9D depict NFkB and

proliferation signatures across TC10 subgroups. Across TC10 subgroups, NFkB signatures are highest in t(14; 16) and RRAS2+ D1 -HRD and D2 subgroups, and lowest in RRAS2- D1 -HRD and D2. Proliferation indexes showed that (FIG. 9A) NFkB

Proliferation Index based on mean of 1 1 -genes as described by Annunziata et al 2007 (BIRC3, TNFAIP3, NFKB2, IL2RG, NFKBIE, RELB, NFKBIA, CD74, PLEK, MALT1, WNT10A). (FIG. 9B) NFkB Proliferation index based on mean of 4-probes {CD74, IL2RG, and 2x TNFAIP3) by Keats et al 2007. (FIG. 9C) Median of 12-probe (TYMS, TK1, CCNB1, MKI67, KIAA101, KIAA0186, CKS1B, TOP2A, UBE2C, ZWINT, TRIP13, KIF11) Proliferation Index of Bergsagel et al 2005, however here we have reported unsealed value. (FIG. 9D) Mean of 50-proliferative genes included in GPI of Hose et al 201 1 reported here as mean of 50-proliferative associated gene.

[0018] FIG. 10A, FIG. 10B, FIG. 10C, FIG. 10D and FIG. 10E depict MAF and MAFB expression across TC10 subgroups and CD20 and PAX5 associated outcome. Scatterplot of MAF and MAFB probes (FIG. 10A) paired with a boxplot of MAF expression (FIG. 10B) and MAFB expression (FIG. 10C) across TC10 subgroups. Note the distinct t(14; 16) and t(14;20) clusters defined solely through MAF and MAFB expression. MMSET t(4; 14) cases also tend to overexpress MAF, however not as extreme as t(14; 16) cases. Note, low outliers in MAF expression included in the t(14; 16) MAF subgroup are presumed MAFA cases that exhibit similar overall expression and risk profiles as t(14; 16) cases while lacking MAF expression. (FIG. 10D, FIG. 10E) High MAF and MAFB expressers of CD20 and PAX5 show improved overall performance compared to lower expressers.

[0019] FIG. 11A, FIG. 11 B, FIG. 11C, FIG. 11 D, FIG. 11 E, FIG. 11 F, FIG.

11G, FIG. 11 H and FIG. 111 depict expression of PAX5, CD20, and RRAS2 across TC10 subgroups. Association between these probes reveal that MMSET uniquely overexpresses RRAS2 with lower CD20 and PAX5 expression. There is also a unique cohort of MAF and MAFB cases with increased CD20 expression. Correlation between PAX5 and CD20 has 0.50 pearson correlation coefficient.

[0020] FIG. 12 depicts ordered bar plot visualizing distribution of key genes across the TC10 subgroups applied to the MRC-IX and HOVON-65/GMMG-HD4. Similar gene expression patterns that mirror the TC10 as applied to the UAMS data are seen here on combined external data set. Key probes such as RRAS2 expression in the D1 and D2 subgroups and CD20 (MS4A1) in the t(1 1 ; 14) subdivide these subgroups similar to TC10 on UAMS data set.

[0021 ] FIG. 13 depicts ordered bar plot visualizing distribution of key genes relevant to NFkB and B-Cell differentiation across the TC10. Expression plot reveals similar distribution of many many of these genes including XBP1, PRDM1, and BCL10. BIRC3 and TNFAIP3 are elevated for RRAS2+ cases.

[0022] FIG. 14 depicts ordered bar plot visualizing distribution of key genes relevant to apoptosis across the TC10. Expression plot reveals higher levels of NOXA across t(1 1 ; 14) and t(4; 14) subgroups and lower levels of expression of BCL-XL within the t(1 1 ; 14) subgroup. Other related genes such as BCL2, MCL1, BAK1, BAD, BAX, PUMA, BIM, and BID show little difference across TC10 subgroups. Ratios of BCL2/MCL1 and BCL2/BCL-XL reveal no difference across TC10 subgroups; however, NOXA/BCL-XL has higher expression in t(1 1 ; 14) and t(4; 14) subgroups.

[0023] FIG. 15 depicts ordered bar plot visualizing distribution of key cell surface markers across the TC10. Expression plot reveals consistent expression for many probes across TC10 subgroups including CD38, SLAMF7, and CD138; however, CD56 lacks expression in t(14; 16) and t(14;20) cases. Also, CD19 expression is correlated with expression of PAX5. [0024] FIG. 16 depicts ordered bar plot visualizing distribution of key genes relevant to IMiD response across the TC10. Expression plot reveals consistent low expression of many genes relating to Imid response such as IKZF1, IKZF2, IKFZ3, and MAX; while others are consistently over expressed: IRF4, RBX1, CRBN, and DDB1.

[0025] FIG. 17A, FIG. 17B, FIG 17C, FIG. 17D, FIG. 17E, FIG. 17F, FIG.

17G, FIG. 17H, FIG. 171, FIG. 17J, FIG. 17K, FIG. 17L, FIG. 17M, FIG. 17N, FIG. 170, FIG. 17P, FIG. 17Q, FIG. 17R, FIG. 17S, FIG. 17T, FIG. 17U, FIG. 17V, FIG. 17W, FIG. 17X, FIG. 17Y, FIG. 17Z, FIG. 17AA, FIG. 17AB, FIG. 17AC, FIG. 17AD, FIG. 17AE, FIG. 17AF, FIG. 17AG, FIG. 17AH, FIG. 17AI, FIG. 17AJ, FIG. 17AK, FIG. 17AL, FIG. 17AM, FIG. 17AN, FIG. 17AO, FIG. 17AP, FIG. 17AQ, FIG. 17AR and FIG. 17AS depict transition of expression patterns in plasma cells at different stages of disease. Scatterplots of key genes for TC10 subgroups across different disease states including healthy donor Normal Plasma Cells (NPC), Waldenstrom Macroglobulinemia (WM), monoclonal gammopathy of undetermined significance (MGUS), smoldering or asymptomatic multiple myeloma (SMM), and newly diagnosed multiple myeloma

(NDMM). Each row of figures shows the transition of plasma cells from normal to myeloma for key gene pairs. (FIG. 17A, FIG. 17B, FIG 17C, FIG. 17D, FIG. 17E)

CCND1 and CCND2 transition from normal levels to deregulated levels for all cases. Note that cases uniquely deregulate one D-group cyclin. (FIG. 17F, FIG. 17G, FIG. 17H, FIG. 171, FIG. 17J) MMSET and FGFR3 have distinct expression patterns in newly diagnosed cases; furthermore, this abnormality is less common in MGUS stages than in NDMM. (FIG. 17K, FIG. 17L, FIG. 17M, FIG. 17N, FIG. 170) MAF and MAFB

abnormalities are common from MGUS stages throughout disease progression and have unique expression patterns. (FIG. 17P, FIG. 17Q, FIG. 17R, FIG. 17S, FIG. 17T) Average RRAS2 expression decreases throughout progression of disease where high expression is the norm for NPCs and low expression for NDMM. This transition may relate to increased frequency of MAPK mutations as disease progresses. (FIG. 17U, FIG. 17V, FIG. 17W, FIG. 17X, FIG. 17Y) Hyperdiploidy marker ISL2 is often low in WM cases, higher in NPC, and distinctly bimodal in NDMM cases with the highest levels seen in D1 and lowest in t(4; 14). (FIG. 17Z, FIG. 17AA, FIG. 17AB, FIG. 17AC, FIG. 17AD) High CD20 (MS4A 1) expression is absent in NPC, but commonly seen in other disease stages; however, low CD20 expression is the increasingly the norm as MM disease progresses. Also, PTP4A3 often increases with progression of disease. (FIG. 17AE, FIG. 17AF, FIG. 17AG, FIG. 17AH, FIG. 17AI) Plasma cell markers XBP1 and IRF4 are high for all stages shown. (FIG. 17AJ, FIG. 17AK, FIG. 17AL, FIG. 17AM, FIG. 17AN) Similarly, PRDM1 is high across all stages; however, PAX5 expression is increasingly differential as disease progresses. (FIG. 17AO, FIG. 17AP, FIG. 17AQ, FIG. 17AR, FIG. 17AS) FOXP1 and IGHM are shown to illustrate distinct expression pattern for WM cases.

[0026] FIG. 18 depicts ordered bar plot visualizing distribution of key genes across the UAMS molecular subgroups. Original Zhan seven-group clusters shown across key genes. Note that expression of CCND1 and CCND2 is not uniformly distributed within subgroups: some CD-1 , CD-2, and HY samples express CCND2 or CCND3 and some cases of LB and PR express CCND1. This is the primary difference between the TC and molecular subgroup framework as CCND1, CCND2, and CCND3 are not primary divisors in the UAMS molecular subgroups.

[0027] FIG. 19 depicts Kaplan-Meier overall survival estimate of PR and non-PR cases by GEP70 HR. Kaplan-Meier curve reveals that there is no significant difference in overall survival between PR and non-PR cases after accounting for GEP70 HR. GEP70 low risk cases that are also PR underperform when compared to non-PR low risk; however, this difference is not significant.

DETAILED DESCRIPTION OF THE INVENTION

[0028] The inventors examined a group of 1 ,355 patients enrolled in the Total Therapy trials, characterized at multiple genetic levels, to develop a novel molecular classifier and risk stratification approach for newly diagnosed myeloma patients (NDMM). Gene expression profiling was used to determine GEP70 risk status, molecular subgroup by UAMS and TC classifications, and to devise a new

Translocation Cyclin D 10-group (TC10) classification. The classification was validated on samples from the UK MRC Myeloma IX and HOVON-65/GMMG-HD4 studies. The TC10 combines known etiologic subgroups with clinically relevant subdivisions to create 10 novel subgroups. Interphase FISH data for IGH translocations were compared to final TC10 subgroups, and 1 p- 1 q+, 13q-, and 17p- iFISH data were used to build GEP proxies that, along with proliferation and NFkB gene signatures, were compared across subgroups. Data from mutational analysis generated by the FoundationOne targeted sequencing panel were also incorporated to understand the distribution of mutations within specific TC10 subgroups. The TC10 is an approach that improves the molecular subclassification of MM by defining both etiological and clinically meaningful subgroups and is able to divide hyperdiploid MM into biologically distinct cohorts.

I. TC10 TO CLASSIFY MULTIPLE MYELOMA

[0029] One aspect of the present disclosure provides a classification system to classify multiple myeloma into 10 distinct classes. The classification system is referred here in as TC10 (Translocation Cyclin D 10) and comprises the following 10 classes: t(1 1 ; 14) CD20/PAX5+, t(1 1 ;14) CD20/PAX5-, t(1 1 ; 14) D1 -HRD:RRAS2+, t(1 1 ; 14) D1 -HRD:RRAS2- t(1 1 ; 14) D2:RRAS2+, t(1 1 ; 14) D2:RRAS2- t(4; 14), t(14; 16), t(14;20) and t(6; 14). The classifications relate to primary cytogenetic aberrations of immunoglobulin (Ig) regions as well as dysregulation of CCND1 and CCND2 and expression patterns for CD20 and RRAS2. This TC10 classification is driven by the primary genetic aberrations associated with myeloma while adding functionally relevant subtypes that relate to survival, response rate, chromosomal aberrations, and mutational profile. In addition, the TC10 classification is now able to subclassify hyperdiploid MM in a meaningful fashion. Simple hyperdiploidy is grouped in the D1 - HRD group, whereas the more complex hyperdiploidy and other karyotypes cluster in the D2 group. Subgrouping each of these groups by RRAS2 expression generates groups with distinct biology and clinical outcomes.

(a) biological sample

[0030] The presence of an aberration of TC10 may be detected in several different biological samples. Non-limiting examples of biological samples may include whole blood, peripheral blood, plasma, serum, bone marrow, urine, lymph, bile, pleural fluid, semen, saliva, sweat, and CSF. The biological sample may be used "as is", the cellular components may be isolated from the biological sample, or a protein faction may be isolated from the biological sample using standard techniques. In one

embodiment, the biological sample is selected from the group consisting of whole blood, peripheral blood, plasma, serum and bone marrow. In another embodiment, the biological sample is whole blood. In yet another embodiment, the biological sample is plasma. In still yet another embodiment, the biological sample is serum. In a different embodiment, the biological sample is peripheral blood. In other embodiments, the biological sample is bone marrow.

[0031 ] As will be appreciated by a skilled artisan, the method of collecting a biological sample from a subject can and will vary depending upon the nature of the biological sample. Any of a variety of methods generally known in the art may be utilized to collect a biological sample from a subject. Generally speaking, the method preferably maintains the integrity of the molecular signature such that it can be accurately quantified in the biological sample. Methods for collecting bone marrow are well known in the art. For example, see US Patent No. 6,846,314, which is hereby incorporated by reference in its entirety. Methods for collecting blood or fractions thereof are also well known in the art. For example, see US Patent No. 5,286,262, which is hereby

incorporated by reference in its entirety.

[0032] A biological sample may be used "as is", the cellular components may be isolated from the fluid, or a protein fraction may be isolated from the fluid using standard techniques. In an embodiment, plasma cells may be isolated from a biological sample. In a specific embodiment, CD138 may be used to isolate plasma cells from the biological sample.

[0033] A biological sample may be collected from any subject diagnosed with MM or used as a disease model for MM. As used herein, "subject" or "patient" is used interchangeably. Suitable subjects include, but are not limited to, a human, a livestock animal, a companion animal, a lab animal, and a zoological animal. In one embodiment, the subject may be a rodent, e.g. a mouse, a rat, a guinea pig, etc. In another embodiment, the subject may be a livestock animal. Non-limiting examples of suitable livestock animals may include pigs, cows, horses, goats, sheep, llamas and alpacas. In yet another embodiment, the subject may be a companion animal. Non- limiting examples of companion animals may include pets such as dogs, cats, rabbits, and birds. In yet another embodiment, the subject may be a zoological animal. As used herein, a "zoological animal" refers to an animal that may be found in a zoo. Such animals may include non-human primates, large cats, wolves, and bears. In specific embodiments, the animal is a laboratory animal. Non-limiting examples of a laboratory animal may include rodents, canines, felines, and non-human primates. In certain embodiments, the animal is a rodent. Non-limiting examples of rodents may include mice, rats, guinea pigs, etc. In a preferred embodiment, the subject is human.

[0034] In some embodiments, the subject has no clinical signs or symptoms of MM. In other embodiments, the subject has mild clinical signs or symptoms of MM, for instance, monoclonal gammopathy of undetermined significance (MGUS), micro-residual disease or smoldering MM. In yet other embodiments, the subject may be at risk for MM. In different embodiments, the subject may have clinical signs or symptoms of MM. In still other embodiments, the subject has been diagnosed with MM. In yet other embodiments, the subject has achieved a complete response (CR) or very good partial response (VGPR) following treatment of MM. "Multiple myeloma" includes symptomatic myeloma, asymptomatic myeloma (smoldering or indolent myeloma), and monoclonal gammopathy of undetermined significance

(MGUS), as defined in Kyle and Rajkumar Leukemia 23: 3-9 (2009, PubMedID

18971951 , the disclosure of which is incorporated by reference in its entirety), as well as the other stratifications and stages described in Kyle and Rajkumar 2009. In a specific embodiment, the subject is newly diagnosed with MM.

(b) detecting the aberration

[0035] In the classification system disclosed herein, one of the 10 aberrations of TC10 is detected to classify a subject. Gene expression may be used to assess the primary aberration for a subject with multiple myeloma. In particular, gene expression profiling may be used. Methods of performing gene expressing profiling are standard in the art and include the steps of RNA processing, target labeling, and hybridization to gene expression arrays. It is known in the art that there is a strong correlation between gene expression profiling (GEP) and iFISH (interphase fluorescent in situ hybridization) as FISH positivity for translocations are accompanied by spikes in gene expression. Accordingly, GEP may be used as a proxy for FISH determined translocations. As such, gene expression may be used to detect the 10 aberrations of TC10. In a specific embodiment, GEP data may be analyzed on the U133Plus 2.0 platform and MAS5 normalized. Gene expression of one or more of the following genes may be used to detect to the 10 aberrations of TC10: ARID5A, BACE2, CCND1, CCND2, CCND3, CDK6, DSG2, FGFR3, ISL2, ITGB7, LAMP5, MAF, MAFB, MS4A 1, NES, NUAK1, PAX5, PTP4A3, RRAS2, S100A4, SLC8A 1, SULF2, USP49, VREB3 and WHSC1. Accordingly, the TC10 classification system may comprise one or more probes used to detect expression of ARID5A, BACE2, CCND1, CCND2, CCND3, CDK6, DSG2, FGFR3, ISL2, ITGB7, LAMP5, MAF, MAFB, MS4A1, NES, NUAK1, PAX5, PTP4A3, RRAS2, S100A4, SLC8A 1, SULF2, USP49, VREB3 and WHSC1.

Additionally, the TC10 classification system comprises probes used to detect

expression of ARID5A, BACE2, CCND1, CCND2, CCND3, CDK6, DSG2, FGFR3, ISL2, ITGB7, LAMP5, MAF, MAFB, MS4A1, NES, NUAK1, PAX5, PTP4A3, RRAS2, S100A4, SLC8A1, SULF2, USP49, VREB3 and WHSC1. Specifically, the TC10 classification system comprises the probes listed in Table 5 to detect expression of ARID5A, BACE2, CCND1, CCND2, CCND3, CDK6, DSG2, FGFR3, ISL2, ITGB7, LAMP5, MAF, MAFB, MS4A1, NES, NUAK1, PAX5, PTP4A3, RRAS2, S100A4, SLC8A1, SULF2, USP49, VREB3 and WHSC1. Expression levels of the genes may be used to determine the TC10 aberration in a subject. FIG. 1 and Table 1 present the various expression levels of genes used to classify a subject into one of the 10 classes. Further, an algorithm may be used to classify a subject into one of the 10 TC10 classes based on expression levels. Specifically, an algorithm may be used to classify a subject into one of the 10 TC10 classes based on expression levels of the genes selected from the group consisting of ARID5A, BACE2, CCND1, CCND2, CCND3, CDK6, DSG2, FGFR3, ISL2, ITGB7, LAMP5, MAF, MAFB, MS4A1, NES, NUAK1, PAX5, PTP4A3, RRAS2, S100A4, SLC8A1, SULF2, USP49, VREB3 and WHSC1

[0036] Other methods for assessing an amount of gene, also referred to herein as "nucleic acid", expression in cells are well known in the art, and all suitable methods for assessing an amount of nucleic acid expression known to one of skill in the art are contemplated within the scope of the invention. The term "amount of nucleic acid expression" or "level of nucleic acid expression" as used herein refers to a measurable level of expression of the nucleic acids, such as, without limitation, the level of messenger RNA (mRNA) transcript expressed or a specific variant or other portion of the mRNA, the enzymatic or other activities of the nucleic acids, and the level of a specific metabolite. The term "nucleic acid" includes DNA and RNA and can be either double stranded or single stranded. Non-limiting examples of suitable methods to assess an amount of nucleic acid expression may include arrays, such as microarrays, PCR, such as RT-PCR (including quantitative RT-PCR), nuclease protection assays and Northern blot analyses. In a specific embodiment, determining the amount of expression of a target nucleic acid comprises, in part, measuring the level of target nucleic acid mRNA expression.

[0037] In one embodiment, the amount of nucleic acid expression may be determined by using an array, such as a microarray. Methods of using a nucleic acid microarray are well and widely known in the art. For example, a nucleic acid probe that is complementary or hybridizable to an expression product of a target gene may be used in the array. The term "hybridize" or "hybridizable" refers to the sequence specific non-covalent binding interaction with a complementary nucleic acid. In a preferred embodiment, the hybridization is under high stringency conditions. Appropriate stringency conditions which promote hybridization are known to those skilled in the art, or can be found in Current Protocols in Molecular Biology, John Wiley & Sons, N.Y. (1989), 6.3.1 6.3.6. The term "probe" as used herein refers to a nucleic acid sequence that will hybridize to a nucleic acid target sequence. In one example, the probe hybridizes to an RNA product of the nucleic acid or a nucleic acid sequence

complementary thereof. The length of probe depends on the hybridization conditions and the sequences of the probe and nucleic acid target sequence. In one embodiment, the probe is at least 8, 10, 15, 20, 25, 50, 75, 100, 150, 200, 250, 400, 500 or more nucleotides in length. In a specific embodiment, probes used in the TC10 classification system are listed in Table 5.

[0038] In another embodiment, the amount of nucleic acid expression may be determined using PCR. Methods of PCR are well and widely known in the art, and may include quantitative PCR, semi-quantitative PCR, multiplex PCR, or any

combination thereof. Specifically, the amount of nucleic acid expression may be determined using quantitative RT-PCR. Methods of performing quantitative RT-PCR are common in the art. In such an embodiment, the primers used for quantitative RT-PCR may comprise a forward and reverse primer for a target gene. The term "primer" as used herein refers to a nucleic acid sequence, whether occurring naturally as in a purified restriction digest or produced synthetically, which is capable of acting as a point of synthesis when placed under conditions in which synthesis of a primer extension product, which is complementary to a nucleic acid strand is induced (e.g. in the presence of nucleotides and an inducing agent such as DNA polymerase and at a suitable temperature and pH). The primer must be sufficiently long to prime the synthesis of the desired extension product in the presence of the inducing agent. The exact length of the primer will depend upon factors, including temperature, sequences of the primer and the methods used. A primer typically contains 15-25 or more nucleotides, although it can contain less or more. The factors involved in determining the appropriate length of primer are readily known to one of ordinary skill in the art.

[0039] The amount of nucleic acid expression may be measured by measuring an entire mRNA transcript for a nucleic acid sequence, or measuring a portion of the mRNA transcript for a nucleic acid sequence. For instance, if a nucleic acid array is utilized to measure the amount of mRNA expression, the array may comprise a probe for a portion of the mRNA of the nucleic acid sequence of interest, or the array may comprise a probe for the full mRNA of the nucleic acid sequence of interest. Similarly, in a PCR reaction, the primers may be designed to amplify the entire cDNA sequence of the nucleic acid sequence of interest, or a portion of the cDNA sequence. One of skill in the art will recognize that there is more than one set of primers that may be used to amplify either the entire cDNA or a portion of the cDNA for a nucleic acid sequence of interest. Methods of designing primers are known in the art. Methods of extracting RNA from a biological sample are known in the art.

[0040] The level of expression may or may not be normalized to the level of a control nucleic acid. This allows comparisons between assays that are performed on different occasions.

[0041 ] Nucleic acid expression may be increased or decreased relative to a control. In one embodiment, nucleic acid expression can be compared using the ratio of the level of expression of a nucleic acid in a subject as compared with the expression level of the nucleic acid in a control. For example, a nucleic acid is differentially expressed if the ratio of the level of expression of a nucleic acid in a subject as compared with the expression level of the nucleic acid in a control is greater than or less than 1 .0. For example, a ratio of greater than 1 , 1 .2, 1 .5, 1 .7, 2, 3, 3, 5, 10, 15, 20 or more, or a ratio less than 1 , 0.8, 0.6, 0.4, 0.2, 0.1 , 0.05, 0.001 or less. In another embodiment, the increase or decrease in expression is measured using p-value. For instance, when using p-value, a nucleic acid is identified as being differentially expressed between a nucleic acid in a subject and the nucleic acid in a control when the p-value is less than 0.1 , preferably less than 0.05, more preferably less than 0.01 , even more preferably less than 0.005, the most preferably less than 0.001 . A control may be the level of expression of a nucleic acid in a subject that is not known to have MM.

[0042] FISH (fluorescent in situ hybridization) may also be used to detect the primary cytogenetic aberrations of immunoglobulin (Ig) regions. FISH is a

cytogenetic technique used to visualize labeled DNA probes hybridized to a region of interest on a chromosome. FISH probes are designed to only bind those parts of the chromosome with which they show a high degree of sequence complementarity.

Generally, a probe is used to detect the IGH locus. The IGH locus includes variable (IGHV), diversity (IGHD), joining (IGHJ), and constant (IGHC) segments. A probe is used to detect translocations involving the IGH locus. Translocations involve many chromosomal partners (translocation partners). Accordingly, a probe may also detect translocation partners. The TC10 classification system may further comprise probes used to detect vFISH designations for 1 p-, 1 q+, 13q- 17p- and hyperdiploidy (HRD). Probes to detect HRD may be selected from the group consisting of ISL2, CCNE1, THG1L, ELOVL7, RPL35A, CCRL2, TNFSF10, C19orf12, LOC100506548 /// RPL37, ESRRA, ATP5L, IP07, RC3H2, SMIM7, SLC12A9, RPL13A /// SNORD32A ///

SNORD33 /// SNORD34 /// SNORD35A, RNF14, and GYG1. Probes to detect 1 p- may be selected from the group consisting of CSDE1, ATP5F1, AHCYL1, LAMTOR5, RSBN1, MAN1A2, CSDE1, CEPT1, TRIM33, BCAS2, HIAT1, LAMTOR5, SARS, CAPZA1, GNAI3, TMED5, AHCYL1, STXBP3, DRAM2, and MAN1A2. Probes to detect 1 q+ may be selected from the group consisting of UBE2Q1, TMEM183A /// TMEM183B, DESI2, PSMD4, PRCC, DCAF6, MRPL9, FBX028, SNAPIN, SLC19A2, VPS72, GPR89B, ETV3, GPR89A /// GPR89B /// GPR89C /// LOC101060247 ///

LOC101060636, MRPL24, LOC101060511 /// PRKAB2, TPM3, CDC42SE1, GNPAT, and TMEM189 /// TMEM189-UBE2V1 /// UBE2V1. Probes to detect 13q- may be selected from the group consisting of DIS3, PCID2, UTP14C, MZT1, STK24, SUGT1, SUCLA2, KPNA3, CDC16, NDFIP2, COMMD6, COG6, SUPT20H, TRIM13, MED4, KBTBD6, ANKRD10, CLN5, and NUDT15. Probes to detect 17p- may be selected form the group consisting of SAT2, TP53, ZBTB4, CYB5D1, FXR2, TMEM256, NDEL1, RANGRF, MAP2K4, MED11, MIS12, C17orf85, CYB5D1, FXR2, DERL2, KIAA0753, ELP5, VAMP2, and MED31. In a specific embodiment, probes used to detect vFISH designations for 1 p- 1 q+, 13q- 17p- and hyperdiploidy are listed in Table 3. Further, additional probes used to detect vFISH designations for 1 p-, 1 q+, 13q-, 17p- hyperdiploidy and TP53 mutation are listed in Table 8. As such, the TC10 classification system may further comprise one or more probes listed in Table 3. Additionally, the TC10 classification system may further comprise one or more probes listed in Table 8.

(c) classifying a subject

[0043] A subject may be classified into 1 of the 10 distinct classes of TC10 based on the detected aberration. Accordingly, the subject may be classified into one of the following 10 classes: t(1 1 ; 14) CD20/PAX5+, t(1 1 ; 14) CD20/PAX5-, t(1 1 ; 14) D1 - HRD:RRAS2+, t(1 1 ; 14) D1 -HRD:RRAS2-, t(1 1 ; 14) D2:RRAS2+, t(1 1 ; 14) D2:RRAS2-, t(4; 14), t(14;16), t(14;20) and t(6;14). For the expression profile used to classify a subject into one of the TC10 classes, see FIG. 1. Additionally, an algorithm derived from this expression profile may be used to classify a subject into one of the 10 TC10 classes.

[0044] A subject may be classified into the t(1 1 ; 14) class if the subject has increased expression of CCND1, SLC8A 1 and/or MS4A1 and decreased expression of SULF2, CDK6, and/or CCND2. In an aspect, a subject may be classified into the t(1 1 ; 14) class if the subject has increased expression of CCND1, SLC8A 1 and MS4A1 and decreased expression of SULF2, CDK6, and CCND2. The t(1 1 ;14) class may be further subdivided based on the expression of CD20 (MS4A1) and PAX5. If a subject has increased expression of CD20 and PAX5 the subject is classified as t(1 1 ; 14) CD20/PAX5+. The t(1 1 ; 14) CD20/PAX5+ classified subject may also have increased expression of VPREB3 and decreased expression of LAMP5. If the subject has decreased expression of CD20 and PAX5 the subject is classified as t(1 1 ; 14)

CD20/PAX5-.

[0045] Additionally, the t(1 1 ; 14) class may further subdivided based on D1-HRD. The subject is classified into D1-HRD if the subject has increased expression of genes on odd number chromosomes, non-limiting examples include ISL2 (chr15q) and CCRL2 (chr3p), and decreased expression of genes on 1 q, non-limiting examples include NES and S100A4. In an aspect, the subject is classified into D1-HRD if the subject has increased expression of ISL2 (chrl 5q) and CCRL2 (chr3p) and decreased expression of NES and S100A4. Additionally, a subject classified into D1-HRD may have increased SULF2 and/or FRZB expression and decreased CCND2 and/or SOCS3 expression. In an aspect, a subject classified into D1-HRD may also have increased SULF2 and FRZB expression and decreased CCND2 and SOCS3 expression. A subject classified into D1-HRD may be further classified based on expression of RRAS2. If a subject has increased expression of RRAS2 the subject is classified as t(1 1 ; 14) D1 -HRD:RRAS2+. The t(1 1 ; 14) D1 -HRD:RRAS2+ classified subject may also have decreased expression of PTP4A3. If the subject has decreased expression of RRAS2 the subject is classified as t(1 1 ; 14) D1 -HRD: RRAS2-. The t(1 1 ; 14) D1 - HRD:RRAS2- classified subject may also have increased expression of LAMP5.

[0046] Additionally, the t(1 1 ; 14) class may further subdivided based on D2. The subject is classified into D2 if the subject has increased expression of CCND2, SOCS3 and/or PTP4A3 and decreased expression of CCND1. In an aspect, the subject is classified into D2 if the subject has increased expression of CCND2, SOCS3 and PTP4A3 and decreased expression of CCND1. A subject classified into D2 may be further classified based on expression of RRAS2. If a subject has increased expression of RRAS2 the subject is classified as t(1 1 ; 14) D2:RRAS2+. If the subject has decreased expression of RRAS2 the subject is classified as t(1 1 ;14) D2:RRAS2-.

[0047] A subject is classified into the t(4; 14) class if the subject has increased expression of WHSC1, FGFR3, BACE2 and/or DSG2 and decreased expression of ISL2, CCND1 and/or FRZB. In an aspect, a subject is classified into the t(4; 14) class if the subject has increased expression of WHSC1, FGFR3, BACE2 and DSG2 and decreased expression of ISL2, CCND1 and FRZB. A subject classified into the t(4; 14) class may be further classified as MAF+. The subject is classified as MAF+ if the subject has increased expression of NUAK1, ARID5A, SMARCA1 and/or ITGB7 and decreased expression of DDK1 and/or SULF2. In an aspect, the subject is classified as MAF+ if the subject has increased expression of NUAK1, ARID5A, SMARCA 1 and ITGB7 and decreased expression of DDK1 and SULF2.

[0048] A subject is classified into the t(14; 16) class if the subject has increased expression of MAF, NUAK1, ARID5A, and/or SMARCA1. In an aspect, a subject is classified into the t(14; 16) class if the subject has increased expression of MAF, NUAK1, ARID5A, and SMARCA 1.

[0049] A subject is classified into the t(14;20) class if the subject has increased expression of MAFB, NUAK1, ARID5A and/or ITGB7. In an aspect, a subject is classified into the t(14;20) class if the subject has increased expression of MAFB, NUAK1, ARID5A and ITGB7. [0050] A subject is classified into the t(6; 14) class if the subject has increased expression of CCND3 and/or USP49. In an aspect, a subject is classified into the t(6; 14) class if the subject has increased expression of CCND3 and USP49.

[0051 ] FIG. 1 presents a full diagram of the gene expression algorithm used to classify a subject via the TC10 classification system. Other means of classifying a subject into a TC10 class are presented in Table 1 and may include hyperdiploidy, loss of 1 p, gain of 1 q, loss of 13q, and/or loss of 17p. Accordingly, the expression profile, or algorithm derived there from, used to classify TC10 classes may further comprise FISH designations. The FISH designations may be 1 p-, 1 q+, 13q-, 17p- and hyperdiploidy. Specifically, the probes depicted in Table 3 may be used in combination with the TC10 classification system. Additionally, the probes depicted in Table 8 may be used in combination with the TC10 classification system.

II. METHODS

[0052] The TC10 classification system disclosed herein may be used in a method to classify a subject diagnosed with multiple myeloma. The method may comprise detecting an aberration in a biological sample obtained from the subject, wherein the aberration is used to classify the subject in one of 10 classes selected from the group consisting of: t(1 1 ; 14) CD20/PAX5+, t(1 1 ; 14) CD20/PAX5-, t(1 1 ; 14) D1 - HRD:RRAS2+, t(1 1 ; 14) D1 -HRD:RRAS2-, t(1 1 ; 14) D2:RRAS2+, t(1 1 ; 14) D2:RRAS2-, t(4; 14), t(14;16), t(14;20) and t(6;14). The classification system may comprise the probes delineated in Table 5. Additionally, vFISH may be used to further classify a subject. In a specific embodiment, vFISH probes may be those delineated in Table 3. In another specific embodiment, vFISH probes may further comprise those delineated in Table 8. Classification of a subject may be useful in enrolling patients for clinical trials. Such classification may result in advancement in therapeutic targeting of myeloma subgroups.

[0053] In another aspect, the TC10 classification system disclosed herein may be used in combination with a risk stratifier to prognose a subject. In a specific example, the risk stratifier may be GEP70. GEP70 is a 70-gene classifier that identifies patients with high risk for short progression-free survival (PFS) and overall survival (OS). The T10 classification system may also provide some indication of prognosis. For example, a subject classified into the t(1 1 ; 14) may be better performing with a five-year OS rate of about 72.3%. Whereas, a subject classified into CD20/PAX5+ may be among the slowest to respond to treatment and a subject classified into CD20/PAX5- may be among the fastest. Additionally, a subject classified into D1-HRD cases may be better performing with a five-year OS rate of about 74.9% while being slower to respond to therapy. Further, a subject classified into D2 may have a faster response to therapy. Still further, a subject classified into t(14; 16) or t(14;20) may have poor prognosis with a five-year OS rate of about 50%. Additionally, a subject classified into t(6; 14) may have the best five-year OS rate of the TC10 subgroups and the fastest response to therapy. Table 1 provides additional information regarding prognosis for the TC10 subgroups.

[0054] In still another aspect, the TC10 classification system disclosed herein may be used to determine treatment of a subject diagnosed with multiple myeloma. Based on the classification of a subject, with or without the use of a risk stratifier, the subject may be treated more or less aggressively. A skilled artisan would be able to determine standard treatment versus aggressive treatment. The TC10 classification may be used to identify groups that are in need of treatment or not or in need of more aggressive treatment. The term "treatment" or "therapy" as used herein means any treatment suitable for the treatment of multiple myeloma. For example, multiple myeloma may be treated with chemotherapy, radiotherapy, immunotherapy, and bone marrow transplant. Non-limiting examples of chemotherapy include

proteosome inhibitors (e.g. bortezomib, carfilzomib), alkylating agents (e.g., melphalan, cyclophosphamide, cisplatin, carboplatin, oxaliplatin), anti-metabolites, taxanes

(paclitaxel, docetaxel), vinca alkaloids, (e.g. vincristine, vinblastine, vinorelbine, vindesine), topoisomerase inhibitors (etoposide, irinotecan, topotecan), cytotoxic antibiotics (doxorubicin, daunorubicin, epirubicin, bleomycin, mitomycin), histone deacetylase inhibitors, dexamethasone, thalidomide, and inhibitors of vascular endothelial growth factors. In some embodiments, the treatment is chemotherapy. In other embodiments, the treatment is radiotherapy. In still other embodiments, the treatment is immunotherapy. In yet other embodiments, the treatment is bone marrow transplant. In other embodiments, the treatment is a proteosome inhibitor.

EXAMPLES

[0055] The following examples are included to demonstrate preferred embodiments of the invention. It should be appreciated by those of skill in the art that the techniques disclosed in the examples that follow represent techniques discovered by the inventors to function well in the practice of the invention, and thus can be considered to constitute preferred modes for its practice. However, those of skill in the art should, in light of the present disclosure, appreciate that many changes can be made in the specific embodiments which are disclosed and still obtain a like or similar result without departing from the spirit and scope of the invention.

Introduction to the Examples.

[0056] Analysis of myeloma plasma cells using cytogenetics and FISH formed the basis of the initial classification and genetic subgrouping of myeloma

(MM).^{1 ,2} These traditional investigations defined a broad based classification system dependent upon either the presence of a balanced translocation into the

immunoglobulin heavy chain (IGH) locus (40%) or the presence of hyperdiploidy (HRD), characterized by gain of the odd numbered chromosomes. Further analysis of the translocated cases identified the overexpression of a D group cyclin as a key molecular aberration.^3,4 In this respect t(1 1 ; 14) directly deregulates CCND1, t(6; 14) CCND3, and t(4; 14) and t(14; 16) indirectly deregulate CCND2. The HRD group can also be sub- classified by the expression of CCND1 and CCND2 providing a pathologically important classifier, termed the Translocation Cyclin D classification (TC).⁴ The TC classification, based on the presence of an IGH translocation and overexpression of a D group cyclin, defines 8 molecular subgroups and recognized that the t(4; 14) and t(14; 16) groups have adverse clinical outcomes.

[0057] The UAMS molecular classification, was derived through a more unbiased analysis of gene expression data.⁵ This classification defined a set of 7 subgroups, each characterized by distinct clinical behavior, which did not correspond exactly to the TC groups. An important clinical feature identified by the UAMS group was the presence of a high-risk subset (15%) defined by a 70-gene signature

associated with a very poor clinical outcome: median PFS of 1 .75 years and OS of 2.87 years.⁶

[0058] In order to improve the ability to predict prognosis, gene mapping studies were utilized to identify a number of recurrent copy number abnormalities (CNA) associated with adverse outcomes including ampl q (10%), del17p (8%) and del1 p32 (30%).^{7 8} In initial FISH studies del(13) (40%) was also identified as an adverse feature, but this was subsequently shown not to be an independent prognostic factor.⁹ Adverse copy number changes in combination with adverse translocations, can be used to form a prognostic system based on summating the number of adverse lesions.¹⁰

[0059] A number of genomic classification systems have, therefore, been proposed, each with limitations. An ideal classification system would identify clinically relevant subgroups with distinct biology that can be recognized in the clinic using robust diagnostic techniques. Importantly such a classification should offer clinically useful information about prognosis and response to specific therapies and should split the disease into meaningfully sized subgroups. In this work, we have readdressed the molecular subclassification of MM and have devised a new classification, the TC10, which has important advantages over previous classifications. The TC10 is based on the results of an analysis of the largest case series across diverse data sets, and defines subgroups using newer bioinformatic techniques including machine learning classification and sparse clustering. Another key feature of this analysis is its external validation using datasets generated from patient cohorts undergoing different treatment approaches in the UK^11"14 and Germany¹⁵, showing that the information can be generalized.

Example 1. Development of the TC10.

[0060] The development of a cytogenetic proxy through gene expression is only feasible due to the strong correlation between GEP and iFISH.³⁰ The t(4;14), t(6; 14), t(1 1 ;14), t(14; 16), and t(14;20) IGH translocation events identified by iFISH strongly correlate with gene expression patterns as FISH positivity for translocations are accompanied by spikes in gene expression, FIG. 6. These distinct patterns of gene expression related to translocations involving 14q32 result in visible clusters that were previously identified based on fixed cutoffs of gene expression. In our further analyses, we generated a 10 group classifier based on the self-identification of IGH translocation and cyclin-D related subgroups through sparse clustering allowing a diverse subset of genes to contribute simultaneously to cluster designations.

[0061 ] An initial unsupervised sparse clustering of the most variable and highest expressed 2,500 probes across the UAMS GEP data generated six subgroups, which we denoted CCND1, D1-HRD, D2, MMSET, MAF/D2, and CCND3 based on the predominant feature of each cluster, FIG. 7. Without any prior knowledge of the importance of specific genes or IGH translocations, the feature selection within sparse clustering identified 90 probes which included CCND1, CCND2, FGFR3, WHSC1, and others genes related to known IGH translocations and cyclin D dysregulation, Table 4. As we searched for the optimal number of partitions in this initial approach, we note that hyperdiploidy, CCND1, and CCND2 expression were the primary divisors of all cases. The divisions due to CCND1 and CCND2 are visually apparent across all data sets, FIG. 8

[0062] After generating these six groups through an unbiased approach, we similarly applied sparse clustering within each subgroup to determine the primary differentiators in each cluster. The CCND1 cluster divided into two subgroups according to paired CD20 and PAX5 expression. The D1-HRD cluster divided into two subgroups according to RRAS2 expression. The MMSET and CCND3 clusters were too

homogeneous to successfully divide further. The D2 and MAF/D2 subgroups were combined together and then re-clustered to isolate a MAF subgroup, identified by high expression of NUAK1 and ARID5A. The remaining non-MAF D2 cases clustered according to RRAS2 expression. The MAF subgroup was then clustered again, and MAF and MAFB subgroups emerged. In the end, sparse clustering across genes selected in an unsupervised fashion, chosen due to high expression and variability, generated 10 distinct subgroups: t(11;14):CD20/PAX5 +/-, D1 -HRD.RRAS2 +/-, D2:RRAS2 +/-, t(4;14) MMSET, t(14;16) MAF, t(14;20) MAFB, and t(6;14) CCND3. These 10-cluster designations across the UAMS data were then trained with a SVM classifier using 25 essential probes and then applied to external data sets, Table 4.

[0063] The TC10 designations show overwhelming agreement with iFISH identification of t(4; 14), t(1 1 ; 14) and t(14; 16) on the external MRC-IX and HOVON- 65/GMMG-HD4 data sets. For the MRC-IX data set, the TC10 has 98.8% agreement with t(4; 14) from iFISH, 99.6% agreement with t(14;16), and 96.5% agreement with t(1 1 ; 14) designations. For the HOVON-65/GMMG-HD4 data set, the TC10 has 97.0% agreement with t(4; 14), 100.0% agreement with t(14; 16), and 93.0% agreement with t(1 1 ; 14) designations.

[0064] The TC10 novel and distinct subgroups have a unique relationship to outcome, response to therapy, and adverse risk related events (FIG. 1 and Table 1 ). Overlap with the UAMS molecular subgroups and Kaplan-Meier curves for overall survival and response to therapy are shown in Table 2 and FIG. 2. The distribution of NFkB and Proliferation Indexes across the TC10 is found in FIG. 9.

Example 2. Summary of TC10 subgroups.

[0065] t(11;14): This subgroup, comprising nearly 20% of all newly diagnosed cases, is best distinguished by elevated expression of CCND1. These cases also overexpress SLC8A1 and BANK1 while under-expressing CDK6, SULF2, and CCND2. Cases with t(1 1 ; 14) are among the better performing, with a five-year OS rate of 72.3%. These cases are rarely hyperdiploid (vFISH for HRD positive 5.3%) and less than a quarter have gain of 1 q. Nearly all of these cases are classified as CD-1 or CD-2 by the UAMS molecular subgroup (32% CD-1 and 61 % CD-2). They can be further subdivided based on the expression of CD20 (MS4A 1) and PAX5, which are positively correlated with VPREB3 expression and negatively with LAMP5. Over 60% of t(1 1 ; 14) cases are classified as CD20/PAX5+ and are distinct both clinically and biologically from their CD20/PAX5- counterparts. The CD20/PAX5+ cases are among the slowest to respond to treatment while CD20/PAX5- cases are among the fastest. The CD20/PAX5- are also more prone to chromosomal losses of 1 p, 13q, and 17p, are more frequently GEP70 HR, and are more proliferative than their CD20/PAX5+ counterparts. CD20/PAX5- cases have poorer outcomes than CD20/PAX5+ cases, however the difference is not statistically significant (five-year OS rates of 74.7 vs. 68.2, log rank p- value of 0.284).

[0066] D1-HRD: This is the most common TC subgroup with nearly a third of all newly diagnosed cases classifying into this hyperdiploid subgroup. D1-HRD cases are distinguishable by their elevated expression of genes on odd numbered

chromosomes such as ISL2 (chr15q), TNFSF10 (chr3q), and CCRL2 (chr3p), while under-expressing genes on 1 q (NES and S100A4). The D1-HRD cases lack CCND2 expression, while expressing CCND1 at an intermediate level: less than t(1 1 ; 14) cases but higher than most CCND2 elevated cases. D1-HRD cases are among the best performers with a five-year OS rate of 74.9% while being among the slowest to respond to therapy. Nearly all of the UAMS HY subgroup classify within this D1-HRD TC10 subgroup (88% of HY cases are D1-HRD). This is consistent with the cytogenetic finding of hyperdiploidy in this group and is also seen in our vFISH HRD proxy (98.4% HRD by vFISH). While the mechanisms of CCND1 overexpression in this group are unknown, it is possible that the intermediate expression of CCND1 is due to trisomy 1 1 . The D1-HRD subtype clusters into RRAS2+ and RRAS2- subtypes according to RRAS2 expression, which is associated with high levels of SNX9, BIRC3, LAPTM5, and

TNFAIP3 expression and low PTP4A3 expression. Less than 40% of D1-HRD cases are RRAS2+ and they are distinct from their RRAS2- counterparts having lower frequency of 1 q+, slower response to therapy, and increased NFkB activation.

[0067] D2: This group comprises over a quarter of all newly diagnosed cases and tends to have increased frequency of adverse features compared to D1- HRD. The D2 group is diverse at a cytogenetic level lacking an IGH translocation and with frequent but not consistent hyperdiploidy. In contrast to the D1-HRD subgroup, D2 cases have a greater frequency of both 1 q+ and GEP70 HR, along with a faster response to therapy. In addition to overexpressing CCND2, D2 cases are among the highest expressers of SOCS3 and PTP4A3. This subgroup comprises two of the original UAMS subgroups that have distinct clinical outcomes: the LB and PR subgroups. In the TC10 the D2 cases are further divided based on RRAS2 expression. The RRAS2+ cases are less common than their RRAS2- counterparts (39% vs. 61 %) and tend to overexpress SNX9, LAPTM5, BIRC3, and TNFAIP3. The RRAS2+ cases have increased NFkB activation and better clinical outcomes than the RRAS2- cases (five- year OS rates of 78.8% vs. 68.1 %, log-rank p-value < 0.01 ). The RRAS2- cases are more frequently HRD and GEP70 HR, which are common features of PR cases (nearly half of all PR cases classify as D2.RRAS2-). Note that D2.RRAS2+ is prone to higher expression of IG related probes than other subgroups, which is presumed to indicate infiltration of normal plasma cells; furthermore, healthy donor samples classify as D2:RRAS2+.

[0068] t(4;14): This group comprises 13% of newly diagnosed cases, is one of the most distinct subgroups, and is identified by overexpression of WHSC1 alone or accompanied with FGFR3. This group overexpresses CCND2 and a majority of cases have 1 q+ (60.0%) and 13q- (78.3%). These cases also tend to overexpress BACE2 and DSG2 while under-expressing ISL2 and FRZB. This group has a

background, low level, RRAS2 expression possibly as a consequence of the impact of MMSET on its promoter region.³¹ The t(4; 14) subgroup has a five-year OS rate of 62.1 % across all protocols, however, all cases treated from TT3 onward, with therapy that included bortezomib, had a significantly improved outcome with a five-year OS rate of 70.4% (TT2 vs. TT3 onward, log-rank p-value < 0.01 ).³²

[0069] t(14;16) & t(14;20): These two subgroups comprise just 6.5% of all cases, and are associated with poor prognosis, having a five-year OS rate of 51 .0%. These cases are most readily identified by their spiked expression of MAF or MAFb for t(14; 16) and t(14;20), respectively FIG. 10A-FIG. 10C. We note that t(4;14) cases have elevated MAF expression, however, t(14;16) cases have levels of expression which are significantly higher than all other subgroups. We also note that there are cases within t(14;20) subgroup with lower MAFB expression that clustered within this subgroup due to high expression of associated probes such as NUAK1, ARID5A, ITGB7, and

SMARCA 1. The t(14;20) cases are rarer than their t(14;16) counterparts and slower to respond to treatment. The t(14;16) and t(14;20) groups contain a subset of cases with elevated expression of CD20 and PAX5 which have significantly better outcomes than their negative counterparts. The highest expressers (Q4) of PAX5 within t(14;16) and t(14;20) have a five-year OS rate of 81.8% compared to 39.2% in the low expressers (log rank p-value < 0.001 ), FIG. 10D-FIG. 10E. A similar pattern is also observed with CD20 expression.

[0070] t(6;14): This is the rarest subgroup of MM comprising only 2.2% of all newly diagnosed cases. It is characterized by high expression of CCND3

accompanied by over expression of USP49 and under expression of CDK6. These cases have the best five-year OS rate of the TC10 subgroups and the fastest response to therapy. In the UAMS molecular classification, cases with t(6; 14) were spread across a variety of subgroups with half of these cases clustering as CD-2.

Example 3. CD20, PAX5, and RRAS2 as distinct dividers.

[0071 ] Beneath each of the molecular subgroups defined initially by deregulation of a D group cyclin and subsequently by the molecular cytogenetic abnormalities there is a further level of organization driven by the expression of PAX5 and the surface membrane expression of CD20. This pattern of expression is

predominantly seen in the t(1 1 ; 14) subgroup, however elevated CD20 expression is seen in 19% of cases across all subgroups with a varying proportion of CD20+ cases across subgroups, FIG. 11. Interestingly, elevated CD20 expression is far more common among high expressers of RRAS2 and within the MAF and CCND3 subgroups. CD20 expression correlates strongly with PAX5 expression (Pearson correlation of 0.50), which should, in the context of plasma cell differentiation, be switched off by PRDM1. It is clear that deregulation of PAX5 is a critical pathogenic feature of myeloma and elucidating how PAX5 is deregulated and its downstream consequences would further our understanding of the disease.

Example 4. Distribution of Mutations.

[0072] Importantly, any new MM classification based on gene expression should reflect other components of the disease biology and genetics. We show that key mutations significantly associated with MM are differentially distributed across the TC10 subgroups. The t(14;16) cases have the highest median mutational load of all of the TC10 subgroups and the RRAS2+ subgroups have lower mutational loads than their RRAS2- counterparts (subgroup adjusted p-value < 0.001 ), FIG. 3A. Furthermore, RRAS2+ cases within the D1-HRD and D2 subgroups have a significantly lower mutational load within the RAS/MAPK pathway (subgroup adjusted p-value < 1 e-13), FIG. 3B-FIG. 3D. This observation is consistent with the RAS/MAPK pathway being activated in this subgroup and the acquisition of additional mutations being redundant. Other associations of specific mutations with TC10 subgroups includes an excess of FAM46C mutations in the D1-HRD:RRAS2- subgroup and the t(4;14) subgroup, ATM mutations within the t(14;16) subgroup, and ATR mutations within the t(14;20) subgroup. In addition, CCND1, FGFR3, and MAF had higher frequency of mutation in the appropriate cytogenetic TC10 subgroup.

[0073] We hypothesized that there might be a distinct mutational pattern in the TC10 subgroups associated with abnormal transcriptional pattern associated with PAX5 (FIG. 3 and FIG. 12). We addressed whether there is an excess of mutations within key activated pathways possibly associated with this transcriptional pattern, e.g. plasma cell differentiation transcription factor cascade, the B-cell receptor pathway, such as is seen in ABC lymphoma, or the NFkB pathway. However, we could not define a mutational signal within our analyses to support any of these hypotheses. We did not look at the patterns of recombination at the translocated Ig loci or whether there was an excess of either abnormal VDJ or class switch recombination events.³³

Example 5. Validation of the TC10 on external datasets.

[0074] The TC10 classifier was applied to the MRC-IX and HOVON- 65/GMMG-HD4 data sets with identical results observed compared to those from the initial UAMS test set. The combined set of MRC-IX and HOVON-65/GMMG-HD4 data had similar proportions of cases within each TC10 subgroup ranging from the largest, D1 -HRD, to the smallest, t(14;20) and t(6; 14) subgroups (chi-square test on subgroup counts: p-value > 0.10), Table 6 and FIG. 12. The distribution of adverse risk markers was also similar in the test and validation datasets: 1 q+ and GEP70 HR were more frequently seen in subgroups with high CCND2 expression; hyperdiploidy was seen in 95.8% of all D1 -HRD cases; and 13q- was common in the t(4; 14) subgroup. Overall, the TC10 had similar proportions and characteristics in both the UAMS test and validation set of MRC-IX and HOVON-65/GMMG-HD4. Furthermore, the translocation related subgroups of the TC10 corresponded well with observed iFISH translocation data as previously described.

Example 6. Simplified Application of TC10 to additional external data sets.

[0075] A simplified TC10 classifier was created, independent of data transformation, so that TC10 classes can be easily defined for additional external data sets. So long as a given external GEP data set is analyzed on the U133Plus 2.0 platform, MAS5 normalized, composed of newly diagnosed patients, and of sufficiently large sample size to self-normalize, the simplified TC10 classifier will define subgroups nearly identical to methods described above. This simplified method uses median centered and median absolute deviation scaled values for the 25 probes in the TC10 model, and is publically available. When applying this simplified method to the external MRC-IX and HOVON-65/GMMG-HD4 data sets, simplified TC10 definitions had 96% agreement with previously described definitions above.

Example 7. Clinical Risk Stratification.

[0076] The GEP70 consistently identifies 15% of patients, with significantly worse outcomes in survival analyses across UAMS and other data sets. In an analysis of risk assessment methods, we compared GEP detection of adverse lesions (t(4; 14), t(14; 16), t(14;20), 17p- 1 p-, and 1 q+) with the GEP70 and showed that samples identified as GEP70 HR have lower OS rates than cases with two or more of these adverse translocation or chromosomal aberrations. The results obtained validated in both the MRC-IX and HOVON65/GMMG-4 data sets, FIG. 4A-FIG. 4C. As previously described, GEP70 HR is uniquely associated with 1 p- and 1 q+ as cases with at least one of these adverse lesions are 7.1 times as likely to be GEP70 HR as compared to cases with neither. Thus, while adverse chromosomal aberrations contribute to our understanding of risk and survival, no combination of these factors either alone or in combination provides as much power as the GEP70 to define adverse survival. With a log rank test statistic p-value of 2.019e-42 and five-year OS of 35.6% for HR and 76.5% for LR, the GEP70 remains the most powerful tool for predicting outcome. Furthermore, the GEP70 validates well on the external data sets used in this analyses irrespective of therapy used, FIG. 4D-FIG. 4F, and its clinical value can, therefore, be generalized.

Example 8. Expression patterns across TC10 subgroups.

[0077] In order to further characterize the nature of each TC10 subgroup, we examined the patterns of expression of a range of gene families important both in the pathogenesis and targeted therapy of myeloma. Previous studies have suggested that the balance between members of the BCL2 family, specifically the BCL2/MCL1 ratio, may predict response to BCL2 inhibitors in myeloma patients.^34,35 We demonstrate a uniform level of BCL2/MCL1 expression across all subgroups; however, we note high levels of NOXA in the t(11;14) and t(4;14) subgroups with correspondingly low levels of BCL-XL leading to a high NOXA/BCD(L ratio in these subgroups, FIG. 14.

[0078] Numerous immune-therapies are in development targeting cell surface proteins on myeloma plasma cells. Studying the expression of these molecules at the mRNA level across patient subgroups we see uniformly high expression of BAFF, SLAMF7, BCMA and CD38, targets for which antibodies are already in clinical trials, FIG. 15. As expected there is high expression of FGFR3 in the t(4;14) subgroup and CD56 (NCAM1) has lower expression in the t(11;14) subgroup and is not expressed in the MAF subgroups.

[0079] We looked at the patterns of expression of IMiD response genes, but did not identify any specific pattern of note, FIG. 16. We identified distinct differences in proliferation index between TC10 subgroups with some cases falling outside of the normal range, NFkB activation was also not distributed uniformly between TC10 subgroups, FIG. 9.

Discussion for the Examples.

[0080] We present a novel approach to the molecular subtyping of myeloma that builds upon the clinical foundations of the UAMS and TC classification systems. This TC10 classification is driven by the primary genetic aberrations associated with myeloma while adding functionally relevant subtypes that relate to survival, response rate, chromosomal aberrations, and mutational profile. In addition, this classification is now able to subclassify hyperdiploid MM in a meaningful fashion. Simple hyperdiploidy is grouped in the D1 -HRD group, whereas the more complex hyperdiploidy and other karyotypes cluster in the D2 group. Subgrouping each of these groups by RRAS2 expression generates groups with distinct biology and clinical outcomes.

[0081 ] The TC10 recognizes the cytogenetic mechanisms underlying myeloma etiology by retaining groups with primary IGH chromosomal translocation associated with deregulation of CCND1, CCND2 or CCND3. All groups recognized by the TC10 exemplify the convergent evolution characteristic of myeloma development where, irrespective of the molecular mechanism underlying it, the end result is deregulation of cyclin D. The deregulation of the G1 S cell cycle checkpoint is of central importance in MM, FIG. 5A. In addition to the characteristic patterns of D group cyclin expression, we identify abnormal distribution of CDK6 expression across the TC10 subgroups with low expression being seen in CD20+/PAX5+ t(11;14) and t(6;14) cases, moderate levels of expression in the other subgroups, and high levels in the MAF subgroups. In addition Cyclin E, required for the transition from G1 to S phase, and CDKN1C (p57/Kip2), an inhibitor of several G1 cyclin/CDK complexes, are more highly expressed in the MAF subgroups, FIG. 5B. These observations fit well with models which consider it essential for a long-lived plasma cell to exit the cell cycle and be maintained in a quiescent state in the bone marrow niche.^36"38 Deregulation of this pathway increases the likelihood that a further event will result in entry into cell cycle and clonal expansion.

[0082] Initial cell biology experiments identified the importance of signaling via the RAS/MAPK,³⁹ JAK/STAT,⁴⁰'⁴¹ PI3K/AKT⁴² and NFkB pathways⁴³ as being central to MM pathogenesis. Mutational analyses have confirmed the relevance of these observations, and, in addition, identified recurrent mutations of the RAS/MAPK pathway present in nearly 50% of NDMM.⁴⁴ Using previous classifications, no specific pattern of mutations in the RAS/MAPK pathway was seen; however, with the use of the TC10, we now see a concentration of mutations in the RRAS2- subgroups. RRAS2, encodes the TC21 protein, a known transforming oncogene, which activates the RAS/MAPK pathway and in normal B cell biology provides constitutive survival signals maintaining the survival of long lived plasma cells.⁴⁵ Analogous to the role of deregulation of the G1 S transition, RAS/MAPK deregulation via RRAS2 or mutation could be of central importance to myeloma development and be a further example of convergent evolution.

[0083] We show that RRAS2 is expressed in normal plasma cells and is down-regulated in malignant plasma cells. Its expression is high in normal plasma cells and cases of MGUS but subsequently its expression level decreases in SMM, reaching its lowest in NDMM, FIG. 17P-FIG. 17T. Using unsupervised clustering, RRAS2 expression can divide both D1 -HRD and D2 cases. In addition, we also observe that the CD20/PAX5+ t(11;14) cases overexpress RRAS2. The frequency of mutations in the MAPK pathway was significantly decreased in RRAS2+ cases suggesting that acquired mutation in this pathway is redundant, in terms of survival, when the pathway is already activated. In an analysis of coregulated genes associated with RRAS2 expression we show that it is associated with expression of PAX5, BIRC3, TNFAIP3, and negatively correlated with PTP4A3 expression.

[0084] Key features of the differentiation of germinal center B-cells to become normal plasma cells include the down-regulation of BCL6 and PAX5 during the exit from the germinal center, the up-regulation of PRDM1 then XBP1, and the activation of a transcriptional program typical of an immunoglobulin secreting plasma cell.^38,46 During this process the expression of CD20, typical of a B-cell until this stage of differentiation, is shut down as is signaling via the B-cell receptor. We show that in both normal and malignant plasma cells there is high expression of IRF4 and PRDM1, typical of a mature plasma cell FIG. 17AE-FIG. 17. Al and FIG. 17AJ-FIG. 17AN. We see expression of CD20 in 19% of myeloma cases, more common in the t(1 1 ; 14) cases (60%) where it subtypes cases into CD20/PAX5+ and CD20/PAX5- groups. In addition, cases with elevated CD20 and PAX5 expression are seen in the t(14; 16) and t(14;20) groups where intriguingly both CD20+ and PAX5+ cases have better clinical outcomes (log rank p-values < 0.05). CD20 expression on the cell surface is frequently seen in flow cytometry analyses, an observation consistent with the expression levels we see at the RNA level. The cluster analysis is driven by the expression of MS4A1, the gene encoding CD20, but interestingly the B cell transcription factor PAX5 is consistently upregulated in this group of cases. We analyzed the relationship of CD20 expression, PAX5 expression, and RRAS2 expression finding that CD20 and PAX5 correlate well, and that CD20 is more commonly overexpressed in RRAS2+ subgroups. The

transcriptional program associated with CD20 seems to be the same in all of the molecular subgroups where it is associated with PAX5 and VPREB3 expression. A pathologically important question in respect to this is why PAX5 remains overexpressed in the context of PRDM1/BL1 MP1 expression.

[0085] The TC10 classification system maintains the important clinical information derived from the UAMS classification by integrating the different biological and clinical information into the TC10 in an etiologically relevant fashion. In the UAMS molecular subgroups, a difference in time to complete response rates and OS is seen between the CD-1 and CD-2 subgroups, Table 7, consistent with them being

pathologically and clinically distinct. The TC10 captures this difference by splitting the t(11;14) cases (previously defined as either CD-1 or CD-2) into CD20/PAX5 positive and negative subsets. In the UAMS approach, the CD-2 subgroup included half of all t(6; 14) cases, which, although rare, are pathologically different and constitute a distinct etiologic subgroup in the TC10.

[0086] In the UAMS classification, the MF and PR subgroups have the poorest prognosis of the seven molecular subgroups and their prognosis failed to improve with novel treatment approaches. The UAMS MF subgroup has been maintained in the TC10 as the t(14;16) and t(14;20) subgroups. In contrast, the UAMS PR subgroup has been divided across numerous subgroups due to its etiologic diversity. The UAMS PR group is best characterized by overexpression of PTP4A3, RRM2, and the cancer testis antigens GAGE and MAGE the pathogenesis of which remains obscure, FIG. 18. Importantly while this pattern of expression may define a clinically relevant group, the underlying etiological aberrations contained within it are fundamentally different, and in the TC10 have been treated as such. Approximately 18% of all PR cases have an IGH translocation and hence these cases are assigned to their respective TC10 subgroups. Of the remaining cases, over 80% classify into an RRAS2- subgroup, either D1-HRD or D2. A further important feature we noted was that the adverse prognosis of the PR subgroup is dependent upon its association with GEP70 defined risk status. When PR status and GEP70 HR designation are included in the same Cox regression model, PR designation does not retain significance.

Additionally, LR PR cases have equivalent outcomes to non-PR LR cases, and HR PR cases have equivalent outcomes to non-PR HR cases (both log rank p-values > 0.05) FIG. 19. Therefore, since the clinical impact of PR designation is dependent upon GEP70 risk features, PR designation only identifies cases with greater likelihood of adverse molecular features such as 1 p-, 1 q+, and 17p- rather than identifying cases with a distinct biology, Table 6.

[0087] A MM classification was proposed by the HOVON group,⁴⁷ and many of their findings are reflected in the TC10 subgroups They provided data supporting the existence of the seven UAMS subgroups, and also developed an additional NFkB related group, a cancer testis antigen (CTA) group, and a group with high expression of PRL3 (PTP4A3). Their NFkB subgroup overexpresses TNFAIP3, which is a key feature for both the RRAS2+ subgroups, whilst over expression of PTP4A3 is a common feature of the RRAS2- D2 subgroup.

[0088] The TC10 subgroups have distinct pathology and clinical outcomes, but it is important not to consider any of the groups to be inherently high or low risk. Rather, patients within each subgroup have a different likelihood of being high risk, lowest in the t(11;14):CD20+, D1-HRD, and D2.RRAS2+ subgroups and highest in the t(14;16) subgroup. We explored the possibility of generating molecular subgroup specific prognostic factors and found it feasible, but initial investigations did not provide additional prognostic information beyond that contained in the GEP70. The genetic markers of chromosome 1 p, 1 q and 17p are important drivers of HR clinical behavior, but the GEP70 improves upon the use of these markers alone. When used for clinical purposes, combining the TC10 with a risk stratifier, such as the GEP70, is the optimal solution for gaining both biological and prognostic information.

[0089] The TC10 provides a framework within which to introduce targeted therapy for MM into the clinic. Each of the groups has a distinct biology and clinical outcome with mutations being distributed in a distinct fashion between subgroups. There are also differences in transcriptional programs active in each group, manifest as distinct patterns of expression, suggesting that each group may have a different response to agents targeting these programs. Notable in this respect are the distinct patterns of expression of genes in the apoptotic pathway consistent with the idea that differential responses may be expected for agents targeting this pathway³⁵. Differential responses to agents targeting the NFkB pathway may be expected based on the distribution of the activation index and of specific mutations. Clear targets for therapy include MAF and MMSET, which are uniquely deregulated and have a distinct impact on downstream transcriptional programs such as RAS/MAPK for the MAF deregulated cases. It is important, therefore, in clinical trials of novel agents to incorporate molecular subtyping either prospectively in the design or in a post-hoc fashion to determine what features may govern response.

[0090] In conclusion, the TC10 serves as a new standard for

understanding both the nature of multiple myeloma pathogenesis and the expected clinical outcome for individual cases. It also provides a basis for addressing

pathologically important questions across distinct groups of myeloma and how they could be targeted therapeutically. It identifies individual biological subgroups, which have different response and outcome patterns, and when paired with a risk stratifier, such as the GEP70, the survival and clinical outcome within each subgroup can be predicted and clinical management adjusted appropriately. The authors suggest that this classification should be the standard approach for classifying Myeloma going forward based on the 2015 Orlando consensus. Methods for the Examples.

[0091] Patient samples. The Myeloma Institute, UAMS dataset comprised newly diagnosed (NDMM) patients accrued to the Total Therapy trials over a 15 year period with a median follow up of 7.75 years.^16"19 In total 1 ,355 patients with baseline gene expression profiling (GEP) paired with corresponding iFISH data for 1 p- 1 q+, 13q- , and 17p- available in part for 1018 samples were used for this analysis. Patients gave written informed consent for bone marrow sampling and the research was approved by the institutional review board of UAMS.

[0092] The UK MRC-IX dataset comprised 273 NDMM patients with corresponding baseline GEP and iFISH.^{11 ,14} iFISH was available for 84.6% of all patients including probes for 1 q+, 17p- and hyperdiploidy.

[0093] The HOVON-65/GMMG-HD4 dataset comprised 390 NDMM patients with corresponding GEP and iFISH.¹⁵ iFISH was available for 51 .0% of all patients including 1 q+, 17p- and hyperdiploidy.

[0094] In all datasets plasma cells were CD138-purified from bone marrow aspirates and processed on Affymetrix U133Plus 2.0 microarrays (Santa Clara, CA) as previously described.²⁰ Raw signals were MAS5 normalized using Affymetrix Microarray GCOS 1 .1 software and the data were analyzed further with R statistical software.

[0095] 552 cases in the UAMS dataset also underwent targeted

sequencing to a median depth of 470x (range: 5-3781 ) using the FoundationOne Heme panel (Foundation Medicine, Cambridge MA). This panel includes sequencing of 405 cancer related genes and selected intronic regions from 31 genes. Targeted RNA-seq was performed for gene rearrangement for 265 genes. Sequences were analyzed for base substitutions, insertions, deletions, copy number alterations (focal amplifications and homozygous deletions) and gene fusions. Germline variants from the 1000

Genomes Project (dbSNP135) were removed. Known somatic alterations deposited in the Catalog of Somatic Mutations in Cancer (COSMIC v62) were highlighted as being biologically significant. All inactivating events (i.e. truncations and deletions) in known tumor suppressor genes were also called as significant. To maximize mutation- detection accuracy (sensitivity and specificity), the test is optimized and validated to detect base substitutions at a > 5% mutant allele frequency (MAF) and indels with a > 10% MAF to > 99% accuracy.

[0096] Statistical overview, data transformation, and cluster analyses. In order to analyze GEP data across each individual trial, data were transformed and combined into one large, unified set. The UAMS data was chosen as the reference set as it is the largest cohort and the training set for the UAMS molecular subgroups and GEP70. The MRC-IX and HOVON-65/GMMG-HD4 were transformed to be of similar distribution to the UAMS GEP data by spline fitting percentile-matched residuals between reference and external data sets across all probes. This method was chosen rather than ComBat,^{21 ,22} as it is more flexible for the analysis of non-normally distributed data typical of myeloma gene expression, which is often bimodal due to distinct populations or skewed by extreme outliers due to translocation spikes.

[0097] In general, sparse / -means clustering²³ was applied to filtered gene expression data to generate the TC10 subgroups. The gene expression data was filtered to remove probes associated with sex and contamination by normal plasma cells. Then the most variable and highest expressed of the remaining probes were included in sparse clustering, which simultaneously clusters while performing feature selection. Further statistical analyses were performed using R 3.1 .3 with frequent use of the survival²⁴, limma²⁵, and e1071²⁶ packages.

[0098] GEP proxies (vFISH) for 1 p- 1 q+, 13q- 17p- and hyperdiploidy were generated by pairing iFISH data with GEP, so that all GEP samples would have complete vFISH estimates for chromosomal aberrations. The hyperdiploidy model was constructed using a tuned support vector machine (SVM) based on MRC-IX data and validated at 90.5% accuracy on the HOVON-65/GMMG-HD4 data. The remaining proxies for chromosomal aberrations were constructed by modeling the percent loss or gain of cells using beta regression tuned with 10-fold cross validation on the UAMS iFISH data. The 1 q+ vFISH model was 96% accurate on UAMS training data and 90% and 86% accurate across the external MRC-IX and HOVON-65/GMMG-HD4 data sets, respectively. The 17p- vFISH model was 99% accurate on UAMS training data and 94% and 91 % accurate across the external MRC-IX and HOVON-65/GMMG-HD4 data sets, respectively. The 1 p- and 13q- models were 97% and 94% accurate on UAMS training data, respectively (external annotations were lacking for these aberrations). The top probes used within all five of these vFISH models are included in Table 3.

[0099] Previously described NFkB and Proliferation indexes were used to further examine distinctions between TC10 subgroups. The 1 1 -probe and 4-probe NFkB signatures were used as previously described.^27,28 An unsealed version of the median 12-gene Proliferation Index³ and the mean of the 50-genes identified in the GPI proliferation score are also reported here.²⁹

Table 1. Descriptive table of TC10 subgroups.

Analysis reveals key distinctions between TC10 subgroups. Primary distinctions include increased frequency of 1q+ within high expressers of CCND2 [D2, t(4;14), t(14; 16), and t(14;20)]. Hyperdiploidy is most common within the D1-HRD subgroup and also seen in the majority of D2:RRAS2- cases. Response rates are highly differential across TC subgroups as the t(6; 14), t(1 1 ; 14):CD20/PAX5-, t(14; 16), and D2 are among the fastest to respond to therapy while the t(1 1 ; 14):CD20/PAX5+, D1-HRD:RRAS2+, and t(14;20) are among the slowest.

Table features counts; list of key up and down- regulated genes; frequencies of hyperdiploidy, loss of 1 p, gain of 1q, loss of 13q, and loss of 17p according to vFISH proxies; GEP70 HR; five- year overall survival rates; and median time to complete response (NA when median not reached). The up and down-regulated genes listed on table are the top 4 most differentially expressed genes for each subgroup compared to all other subgroups with a minimum

moderated t-statistic p-value less than 1e-20, performed using limma.

GEP70 Median

Loss Gain Loss Loss Five-

Count Up- Down- Hyper- High Time to

Groups Count 1 p 1 q 13q 17p year

(%) regulated regulated diploid (%) Risk CR

(%) (%) (%) (%) OS%

(%) (months)

CCND1 SULF2

t(11 ;14) 262 19.3 SLC8A1 CDK6 5.3 7.3 21 .0 35.9 9.5 8.4 72.3 23.4

MS4A1 CCND2

CCND1 CDK6

VPREB3 SULF2

CD20(+) 167 12.3 6.6 1 .2 23.4 29.9 4.8 2.4 74.7 NA

MS4A1 LAMP5

PAX5 CCND2

SULF2

CD20(-) 95 7.0 CCND1 3.2 17.9 16.8 46.6 17.9 18.9 68.2 6.4

CCND2

ISL2 CCND2

CCRL2 S100A4

D1 428 31 .6 98.4 20.6 15.0 22.4 7.2 7.7 74.9 39.5

SULF2 NES

FRZB SOCS3

RRAS2 PTP4A3

ISL2 S100A4

RRAS2(+) 161 11 .9 98.8 25.5 7.5 24.2 9.3 6.2 75.4 NA

SULF2 CCND2

FRZB NES

ISL2

CCND2

SULF2

RRAS2(-) 267 19.7 RRAS2 98.1 17.6 19.5 21 .3 6.0 8.6 74.5 23.1

LAMP5

S100A4

FRZB

CCND2

D2 367 27.1 SOCS3 CCND1 63.5 16.3 49.0 53.1 7.1 13.6 72.2 12.5

PTP4A3 RRAS2

RRAS2(+) 143 10.6 40.6 20.3 40.6 47.6 7.7 7.0 78.8 11 .7

CCND2

CCND2

RRAS2

RRAS2(-) 224 16.5 PTP4A3 78.1 13.8 54.5 56.7 6.7 17.9 68.1 13.2

CCND1

SOCS3

WHSC1

ISL2

FGFR3

t(4;14) 180 13.3 CCND1 8.9 25.6 60.0 78.3 8.9 28.9 62.1 15.9

BACE2

FRZB DSG2

NUAK1

DDK1

ARID5A

MAF 88 6.5 SULF2 10.2 28.4 68.2 71 .6 10.2 44.3 51.0 21 .4

SMARCA1

FRZB ITGB7

MAF NUAK1

t(14;16) 54 4.0 7.4 27.8 74.1 77.8 9.3 50.0 48.0 11 .2

ARID5A

SMARCA1

MAFB NUAK1

t(14;20) 34 2.5 14.7 29.4 58.8 61 .8 1 1.8 35.3 56.0 NA

ARID5A

ITGB7

CCND3

t(6;14) 30 2.2 33.3 10.0 16.7 70.0 6.7 10.0 79.0 6.7

USP49

209658 at CDC16 chr13q34

224801 at NDFIP2 chr13q31.1

225312 at C0MMD6 chr13q22

225769 at C0G6 chr13q14.1 1

209659 s at CDC16 chr13q34

226330 s at SUPT20H chr13q13.3

230192 at TRIM13 chr13q14

217843 s at MED4 chr13q14.2

226479 at KBTBD6 chr13q14.1 1

218093 s at ANKRD10 chr13q34

204085 s at CLN5 Chr13q21.1-q32

219347 at NUDT15 chr13q14.2

Model EventID Symbol Chromosome

17p- 225272 at SAT2 chr17p13.1

201746 at TP53 chr17p13.1

225629 s at ZBTB4 chr17p13.1

226833 at CYB5D1 chr17p13.1

35265 at FXR2 chr17p13.1

227063 at TMEM256 chr17p13.1

208093 s at NDEL1 chr17p13.1

218526 s at RANGRF chr17p13.1

203266 s at MAP2K4 chr17p12

226958 s at MED11 chr17p13.2

221559 s at MIS12 chr17p13.2

218896 s at C17orf85 chr17p13.2

155271 1 a at CYB5D1 chr17p13.1

203172 at FXR2 chr17p13.1

231002 s at ... ...

218333 at DERL2 chr17p13.2

204711 at KIAA0753 chr17p13.1

223286 at ELP5 chr17p13.1

201556 s at VAMP2 chr17p13.1

222867 s at MED31 chr17p13.1

219377 at GAREM chr18q12.1

212724 at RND3 chr2q23.3

217865 at RNF130 chr5q35.3

203186 s at S100A4 chr1q21

235518 at SLC8A1 Chr2p23-p22

238546 at SLC8A1 Chr2p23-p22

223028 s at SNX9 Chr6q25.1-q26

227697 at S0CS3 chr17q25.3

201417 at S0X4 chr6p22.3

224724 at SULF2 Chr20q12-q13.2

233555 s at SULF2 Chr20q12-q13.2

217147 s at TRAT1 chr3q13

201387 s at UCHL1 chr4p14

209053 s at WHSC1 chr4p16.3

222777 s at WHSC1 chr4p16.3

209052 s at WHSC1 chr4p16.3

222778 s at WHSC1 chr4p16.3

209054 s at WHSC1 chr4p16.3

223472 at WHSC1 chr4p16.3

208606 s at WNT4 Chr1 p36.23-p35.1

Table 5. Probes for TC10 model. |

List of 25 probes used to perform T CIO designations with a | support vector machine multiclass classifier. Key genes j relevant to all 10 subgroups are present in this list including |

Table 7. Descriptive table for UAMS molecular subgroups.

Table reveals details of UAMS molecular subgroups identically to methods used to detail the

TC10 subgroups. We see that HY is the most common subgroup with nearly all cases identified as HRD by vFISH. CD-2 subgroup is the slowest to respond to therapy and also has the fewest cases identified asl p- and GEP70 HR. The PR and MF subgroups have the poorest five-year

OS rates of 54.2% and 51.6%, respectively. These two subgroups also have the greatest frequency o F GEP70 HR, 1q +, and 1 p-.

Median

Loss Gain Loss Loss GEP70 Five-

Count Up- Down- Hyperdiploid Time to

Groups Count 1 p 17p High year

gulated regulated 1 q 13q

(%) re (%) CR

(%) (%) (%) (%) Risk (%) OS%

(months)

SULF2

CD-1 93 6.9 CCND1 1 .1 14.0 14.0 49.5 14.0 18.3 72.7 6.3

CCND2

VPREB3

CDK6

MS4A1

CD-2 214 15.8 LAMP5 8.9 1 .9 17.8 33.6 6.5 0.9 74.8 NA

CCND1

SULF2

PAX5

ISL2

S100A4

CCRL2

HY 420 31 .0 CCND2 99.5 17.6 9.8 19.3 6.4 3.8 78.5 29.1

SULF2

NES

FRZB

WHSC1

ISL2

FGFR3

MS 177 13.1 CCND1 9.0 22.0 57.6 79.1 7.3 23.2 65.3 15.2

BACE2

FRZB DSG2

NUAK1 DKK1

ARID5A FRZB

MF 83 6.1 8.4 30.1 67.5 69.9 10.8 43.4 51 .6 21 .4

SMARCA1 SULF2

ITGB7 CCND1

CST6

LB 189 13.9 CCND1 60.8 14.3 57.1 55.0 5.8 2.6 75.4 14.4

CCND2

PR 179 13.2 PTP4A3 70.9 33.0 63.7 60.9 12.3 45.8 54.2 1 1 .7

202300 at 223531 x at 225769 at 218896 s at 208745 at 202284 s at

200802 at 227200 at 209659 s at 155271 1 a at 200992 at 235534 at

208374 s at 225463 x at 226330 s at 203172 at 200994 at 227578 at

201 180 s at 218270 at 230192 at 231002 s at 220201 at 217542 at

202195 s at 225278 at 217843 s at 218333 at 224717 s at 42361_g_at

200849 s at 222976 s at 226479 at 20471 1 at 223995 at 235571 at

203310 at 229120 s at 218093 s at 223286 at 200715 x at 244616 x at

225228 at 201956 s at 204085 s at 201556 s at 201823 s at 200770 s at

217921 at 201003 x at 219347 at 222867 s at 21 1275 s at 213036 x at

207855 s at 224634 at 223412 at 219435 at 244841 at 222022 at

200848 at 224513 s at 202304 at 230274 s at 205356 at 205903 s at

227278 at 221486 at 235339 at 224900 at 200993 at 233413 at

238653 at 209435 s at 209009 at 201530 x at 208289 s at 227584 at

202194 at 204478 s at 214252 s at 226703 at 214329 x at 208065 at

202776 at 220642 x at 224799 at 204690 at 213203 at 1560570 a at

225226 at 201786 s at 229298 at 211787 s at 213041 s at 209877 at

1555543 a at 222443 s at 202259 s at 220606 s at 219081 at 207534 at

223331 s at 209382 at 229078 s at 214552 s at 218038 at 234051 at

228026 at 21 1594 s at 218203 at 214805 at 224814 at 217876 at

220306 at 202244 at 218352 at 209295 at 212129 at 230777 s at

217920 at 201377 at 229970 at 223081 at 226189 at 231477 at

202647 s at 219696 at 202548 s at 235857 at 238590 x at 1560177 at

231894 at 210386 s at 218420 s at 203344 s at 229838 at 240599 x at

207157 s at 223470 at 236446 at 228597 at 204949 at 215515 at

223671 x at 221497 x at 223306 at 209208 at 213737 x at

22681 1 at 202243 s at 231871 at 201557 at 227056 at

227981 at 235196 at 222239 s at 200855 at 226902 at

201421 s at 201612 at 215096 s at 209141 at 224495 at

243501 at 224314 s at 240574 at 228183 s at 213348 at

212360 at 1552617 a at 233647 s at 209350 s at 234339 s at

202362 at 223046 at 228937 at 229551 x at 205672 at

224985 at 213027 at 221899 at 235728 at 202759 s at

228661 s at 228238 at 214028 x at 212601 at 209329 x at

212629 s at 209825 s at 234993 at 22451 1 s at 213223 at

201274 at 218728 s at 217732 s at 205055 at 219362 at

218040 at 201425 at 206235 at 202632 at 212716 s at

207654 x at 206468 s at 221503 s at 226787 at 223175 s at

238524 at 2261 15 at 236067 at 236241 at 237817 at

1554154 at 230917 at 236122 at 228482 at 227394 at

212893 at 228852 at 15571 18 a at 204170 s at 209533 s at 218462 at 205788 s at 239314 at 204162 at 235027 at

203346 s at 225786 at 229718 at 230012 at 218943 s at

235294 at 225244 at 213375 s at 202900 s at 227146 at

215285 s at 220059 at 213357 at 200854 at 200089 s at

215333 x at 202337 at 204831 at 2151 14 at 441 1 1 at

209204 at 41329 at 37512 at 203871 at 227385 at

238087 at 205661 s at 201672 s at 202282 at 244066 at

204646 at 202220 at 208753 s at 3281 1 at 220202 s at

207237 at 212839 s at 219562 at 229833 at 202387 at

36499 at 205945 at 226050 at 238733 at 228561 at

212704 at 238056 at 204226 at 224512 s at 227099 s at

213689 x at 203035 s at 223404 s at 201899 s at 212843 at

212291 at 225401 at 229751 s at 226262 at 38269 at

205087 at 223555 at 225570 at 209440 at 227518 at

202234 s at 223322 at 201551 s at 226996 at 213844 at

201420 s at 2141 13 s at 221561 at 215667 x at 244406 at

205371 s at 217894 at 213882 at 202580 x at 1554273 a at

202377 at 212530 at 204064 at 229175 at 229720 at

244481 at 228032 s at 205632 s at 204146 at 207480 s at

223584 s at 228191 at 15531 13 s at 202534 x at 204950 at

203362 s at 222987 s at 202413 s at 228361 at 219348 at

223849 s at 228385 at 222868 s at 2021 10 at 202760 s at

228297 at 229828 at 209934 s at 219148 at 213183 s at

212678 at 1558027 s at 221550 at 202284 s at 205363 at

227068 at 205607 s at 1556821 x at 212247 at 216396 s at

239596 at 221222 s at 238623 at 218308 at 44696 at

1552738 a at 206332 s at 223339 at 209190 s at 200083 at

220840 s at 15531 18 at 226224 at 208890 s at 219343 at

220948 s at 210589 s at 220768 s at 219123 at 44669 at

212676 at 226482 s at 212585 at 1557128 at 224655 at

227680 at 219373 at 204459 at 219417 s at 230277 at

223275 at 236280 at 214906 x at 205873 at 40020 at

207168 s at 216593 s at 225795 at 204817 at 1555575 a at

202304 at 214719 at 227465 at 203336 s at 231930 at

226568 at 217797 at 201746 at 202154 x at 207452 s at

204032 at 212408 at 1557066 at 201092 at 224948 at

218567 x at 1554057 at 235215 at 227085 at 202649 x at

218979 at 222605 at 202268 s at 44563 at 21 1563 s at

222734 at 229544 at 222077 s at 242153 at 212648 at

226488 at 228386 s at 216682 s at 228040 at 201573 s at 202236 s at 213261 at 225424 at 1552733 at 205774 at

218355 at 226384 at 214946 x at 228427 at 1558622 a at

202690 s at 224709 s at 223026 s at 229021 at 221706 s at

214526 x at 200797 s at 203745 at 22971 1 s at 224151 s at

202954 at 231907 at 205804 s at 200737 at 1554456 a at

224523 s at 200652 at 203358 s at 227578 at 1569022 a at

213088 s at 241262 at 213186 at 243191 at 46270 at

213647 at 210417 s at 201230 s at 237429 at 229693 at

225926 at 226233 at 58308 at 212597 s at 21 1988 at

203746 s at 1557166 at 1558953 s at 203159 at 226915 s at

2351 13 at 226447 at 228035 at 202842 s at 235035 at

223016 x at 235603 at 213261 at 214430 at 205578 at

241853 at 210978 s at 233750 s at 205191 at 231 188 at

219588 s at 1554575 a at 230465 at 210314 x at 236227 at

201614 s at 225763 at 221940 at 225017 at 209285 s at

226718 at 224855 at 235303 at 1557129 a at 202604 x at

222916 s at 1554049 s at 238919 at 204325 s at 227972 at

205369 x at 218518 at 214703 s at 213334 x at 1554455 at

201477 s at 204742 s at 201602 s at 1555874 x at 1554188 at

241922 at 1555226 s at 200699 at 224751 at 200758 s at

225928 at 218873 at 211069 s at 222396 at 205452 at

203283 s at 220992 s at 222869 s at 201 191 at 214895 s at

230766 at 234995 at 201247 at 203262 s at 200810 s at

226384 at 224669 at 224689 at 201 1 15 at 235521 at

222396 at 201912 s at 209628 at 200600 at 20081 1 at

218499 at 239825 at 235463 s at 216961 s at 212233 at

201832 s at 201273 s at 218373 at 217542 at 221483 s at

222468 at 225793 at 202225 at 201312 s at 2141 16 at

218882 s at 224791 at 213098 at 1561554 at 209282 at

233294 at 223624 at 203193 at 218494 s at 228592 at

65086 at 204923 at 227767 at 1557192 at 207102 at

203391 at 215438 x at 243539 at 224185 at 201260 s at

227267 at 235747 at 229067 at 1560391 at 213256 at

219562 at 221824 s at 219000 s at 207038 at

233917 s at 209546 s at 200815 s at 201310 s at

205191 at 226527 at 218399 s at 222868 s at

232066 x at 214615 at 224578 at 206662 at

201954 at 229670 at 209946 at 205208 at

201465 s at 218068 s at 230072 at 21 1012 s at

236122 at 222667 s at 221 176 x at 225425 s at 210707 x at 219875 s at 232706 s at 201259 s at

239038 at 243780 at 204197 s at 231997 at

2261 19 at 21 1080 s at 237643 at 1568997 at

219473 at 241433 at 206835 at 209640 at

1554878 a at 214228 x at 223480 s at 229063 s at

200738 s at 218401 s at 1564337 at 221490 at

202915 s at 1553856 s at 203832 at 226732 at

203832 at 219313 at 205240 at 208867 s at

203745 at 228493_at 221620 s at 210755_at

209537 at 1559776 at 231383 at 220182 at

203022 at 239536 at 218363 at

1553709 a at 212800 at 1555872 a at

235507 at 218407 x at 229519 at

202235 at 1553314 a at 202824 s at

50965 at 223402 at 201558 at

241720 at 219922 s at 2011 11_at

223294 at 243570 at 213647 at

1558305 at 226615 at 235299 at

239982 at 229447 x at 235507 at

203640 at 229074 at 219843 at

219531 at 212594 at 243634 at

235780 at 201959 s at 218472 s at

202706 s at 212483 at 201068 s at

200600 at 203523 at 244519 at

225792 at 233248 at 209149 s at

225153 at 220193 at 218966 at

209308 s at 235292 at 218380 at

211759 x at 208759 at 230840 at

239409 at 220338 at 205486 at

242374 at 238005 s at 1558304 s at

231909 x at 1558508 a at 217329 x at

203764 at 213658 at 224177 s at

224413 s at 226962 at 238247 at

215380 s at 205076 s at 1560339 s at

223568 s at 219056 at 49452 at

209189 at 229145 at 216548 x at

223024 at 223625 at 214 80 at

2391 18 at 212409 s at 213302 at

213340 s at 206919 at 220800 s at

2021 10 at 228619 x at 225441 x at 1554520 at 243843 at 219540 at

201590 x at 206978 at 236655 at

204342 at 229303 at 229131 at

235612 at 223265 at 230633 at

21231 1 at 213279 at 218669 at

211071 s at 201804 x at 200980 s at

227414 at 2381 14 at 239082 at

212522 at 223548 at 203396 at

229595 at 203143 s at 224896 s at

234295 at 212604 at 223041 at

1561884 at 56197 at 1553434 at

217886 at 213902 at 227846 at

207556 s at 201930 at 238105 x at

225201 s at 239843 at 240615 at

214757 at 216250 s at 234454 at

1558957 s at 225012 at 229436 x at

229000 at 217967 s at 225533 at

200597 at 201061 s at 210378 s at

1552973 at 227776 at 205662 at

203077 s at 53076 at 1556457 s at

244383 at 209868 s at 238021 s at

202852 s at 227414 at 205671 s at

223037 at 206674 at 1568656 at

222976 s at 230605 at 204727 at

202209 at 238574 at 224461 s at

209946 at 227031 at 243004 at

223434 at 213155 at 225741 at

229693 at 213988 s at 217399 s at

230350 at 207350 s at 1557838 at

215313 x at 204603 at 1566597 at

1561762 s at 200978 at 2081 17 s at

229025 s at 209536 s at 204054 at

219037 at 231832 at 1553373 at

218558 s at 214056 at 208407 s at

208943 s at 218788 s at 222487 s at

217047 s at 218723 s at 1553258 at

213044 at 224885 s at 237392 at

223542 at 1552892 at 242723 at

238758 at 231 103 at 243269 s at

225302 at 228745 at 221474 at 235343 at 222026 at 1566631 at

211069 s at 227962 at 231 183 s at

210034 s at 214450 at 212096 s at

238615 at 225722 at 204588 s at

241396 at 216885 s at 201798 s at

222889 at 213180 s at 217356 s at

202225 at 200980 s at 223522 at

243671 at 241321 at 217795 s at

243527 at 212418 at

218592 s at 220214 at

204759 at 226760 at

208632 at 238781 at

201 141 at 201555 at

213912 at 238070 at

208370 s at 235039 x at

231962 at 212642 s at

205270 s at 200658 s at

203197 s at 203641 s at

209289 at 50965 at

212797 at 221704 s at

227134 at 226893 at

49452 at 219248 at

202201 at 217513 at

202580 x at 218994 s at

222418 s at 235064 s at

227928 at 2441 10 at

209300 s at 221234 s at

226255 at 240310 at

229928 at 1568597 at

226896 at

210605 s at

242560 at

219648 at

200980 s at

229969 at

238881 at

227552 at

202934 at

223804 s at

217356 s at 244593 at

222622 at

226915 s at

225344 at

222843 at

230362 at

225702 at

213951 s at

224628 at

1553587 a at

217929 s at

References for the Examples.

1 . Fonseca R, Barlogie B, Bataille R, et al. Genetics and cytogenetics of multiple myeloma: a workshop report. Cancer research. Feb 15 2004;64(4): 1546-1558.

2. Cremer FW, Bila J, Buck I, et al. Delineation of distinct subgroups of multiple myeloma and a model for clonal evolution based on interphase cytogenetics. Genes, chromosomes & cancer Oct 2005;44(2): 194-203.

3. Bergsagel PL, Kuehl WM, Zhan F, Sawyer J, Barlogie B, Shaughnessy J. Cyclin D dysregulation: an early and unifying pathogenic event in multiple myeloma. Blood. Jul 2005; 106(1 ):296-303.

4. Bergsagel PL, Kuehl WM. Molecular pathogenesis and a consequent classification of multiple myeloma. J Clin Oncol. Sep 2005;23(26):6333-6338.

5. Zhan F, Huang Y, Colla S, et al. The molecular classification of multiple myeloma. Blood. Sep 2006; 108(6):2020-2028.

6. Shaughnessy JD, Zhan F, Burington BE, et al. A validated gene expression model of high-risk multiple myeloma is defined by deregulated expression of genes mapping to chromosome 1 . Blood. Mar 2007; 109(6):2276-2284.

7. Walker BA, Leone PE, Jenner MW, et al. Integration of global SNP-based mapping and expression arrays reveals key regions, mechanisms, and genes important in the pathogenesis of multiple myeloma. Blood. Sep 2006; 108(5): 1733-1743. Walker BA, Leone PE, Chiecchio L, et al. A compendium of myeloma-associated chromosomal copy number abnormalities and their prognostic value. Blood. Oct 14 2010; 1 16(15):e56-65.

Shaughnessy J, Tian E, Sawyer J, et al. High incidence of chromosome 13 deletion in multiple myeloma detected by multiprobe interphase FISH. Blood. Aug 2000;96(4): 1505-151 1 .

Boyd KD, Ross FM, Chiecchio L, et al. A novel prognostic model in myeloma based on co-segregating adverse FISH lesions and the ISS: analysis of patients treated in the MRC Myeloma IX trial. Leukemia. Feb 2012;26(2):349-355.

Morgan GJ, Davies FE, Gregory WM, et al. Cyclophosphamide, thalidomide, and dexamethasone as induction therapy for newly diagnosed multiple myeloma patients destined for autologous stem-cell transplantation: MRC Myeloma IX randomized trial results. Haematologica. Mar 2012;97(3):442-450.

Morgan GJ, Davies FE, Gregory WM, et al. Effects of induction and maintenance plus long-term bisphosphonates on bone disease in patients with multiple myeloma: the Medical Research Council Myeloma IX Trial. Blood. Jun 2012; 1 19(23):5374-5383.

Morgan GJ, Davies FE, Gregory WM, et al. Long-Term Follow-Up of MRC Myeloma IX Trial: Survival Outcomes with Bisphosphonate and Thalidomide Treatment. Clinical cancer research : an official journal of the American Association for Cancer Research. Aug 30 2013.

Morgan GJ, Davies FE, Gregory WM, et al. Cyclophosphamide, thalidomide, and dexamethasone (CTD) as initial therapy for patients with multiple myeloma unsuitable for autologous transplantation. Blood. Aug 4 201 1 ; 1 18(5): 1231 -1238. Sonneveld P, Schmidt-Wolf IG, van der Holt B, et al. Bortezomib induction and maintenance treatment in patients with newly diagnosed multiple myeloma: results of the randomized phase III HOVON-65/ GMMG-HD4 trial. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. Aug 20 2012;30(24):2946-2955. Barlogie B, Tricot G, Anaissie E, et al. Thalidomide and hematopoietic-cell transplantation for multiple myeloma. N Engl J Med. Mar 2006;354(10): 1021 - 1030.

Barlogie B, Pineda-Roman M, van Rhee F, et al. Thalidomide arm of Total Therapy 2 improves complete remission duration and survival in myeloma patients with metaphase cytogenetic abnormalities. Blood. Oct 2008; 1 12(8):31 15-3121 .

Barlogie B, Anaissie E, van Rhee F, et al. Incorporating bortezomib into upfront treatment for multiple myeloma: early results of total therapy 3. British journal of haematology. Jul 2007; 138(2): 176-185.

Barlogie B, Mitchell A, van Rhee F, Epstein J, Morgan GJ, Crowley J. Curing myeloma at last: defining criteria and providing the evidence. Blood. Nov 2014; 124(20):3043-3051 .

Zhan F, Hardin J, Kordsmeier B, et al. Global gene expression profiling of multiple myeloma, monoclonal gammopathy of undetermined significance, and normal bone marrow plasma cells. Blood. Mar 1 2002;99(5): 1745-1757.

Chen C, Grennan K, Badner J, et al. Removing batch effects in analysis of expression microarray data: an evaluation of six batch adjustment methods. PLoS One. 201 1 ;6(2):e17238.

Stein CK, Qu P, Epstein J, et al. Removing batch effects from purified plasma cell gene expression microarrays with modified Com Bat. BMC Bioinformatics. 2015; 16:63.

Witten DM, Tibshirani R. A framework for feature selection in clustering. J Am Stat Assoc. Jun 2010;105(490):713-726.

Therneau TM, Grambsch PM. Modeling Survival Data: Extending the Cox Model. New York: Springer; 2000.

Ritchie ME, Phipson B, Wu D, et al. limma powers differential expression analyses for R/V/A-sequencing and microarray studies. Nucleic Acids Research. 2015;43:doi: 10.1093/nar/gkv1007. Meyer D, Dimitriadou E, Hornik K, Weingessel A, Leisch F. e1071: Misc Functions of the Department of Statistics (e1071), TU Wien. 2014.

Annunziata CM, Davis RE, Demchenko Y, et al. Frequent engagement of the classical and alternative NF-kappaB pathways by diverse genetic abnormalities in multiple myeloma. Cancer Cell. Aug 2007; 12(2): 1 15-130.

Keats JJ, Fonseca R, Chesi M, et al. Promiscuous mutations activate the noncanonical NF-kappaB pathway in multiple myeloma. Cancer cell. Aug 2007; 12(2): 131 -144.

Hose D, Reme T, Hielscher T, et al. Proliferation is a central independent prognostic factor and target for personalized and risk-adapted treatment in multiple myeloma. Haematologica. Jan 201 1 ;96(1 ):87-95.

Tian E, Sawyer JR, Heuck CJ, et al. In multiple myeloma, 14q32 translocations are nonrandom chromosomal fusions driving high expression levels of the respective partner genes. Genes, chromosomes & cancer. Jul 2014;53(7):549- 557.

Kuo AJ, Cheung P, Chen K, et al. NSD2 links dimethylation of histone H3 at lysine 36 to oncogenic programming. Molecular cell. Nov 18 201 1 ;44(4):609-620. Pineda-Roman M, Zangari M, Haessler J, et al. Sustained complete remissions in multiple myeloma linked to bortezomib in total therapy 3: comparison with total therapy 2. Br J Haematol. Mar 2008; 140(6):625-634.

Walker BA, Wardell CP, Johnson DC, et al. Characterization of IGH locus breakpoints in multiple myeloma indicates a subset of translocations appear to occur in pregerminal center B cells. Blood. Apr 2013; 121 (17):3413-3419.

Bodet L, Gomez-Bougie P, Touzeau C, et al. ABT-737 is highly effective against molecular subgroups of multiple myeloma. Blood. Oct 201 1 ; 1 18(14):3901 -3910. Touzeau C, Dousset C, Le Gouill S, et al. The Bcl-2 specific BH3 mimetic ABT- 199: a promising targeted therapy for t(1 1 ; 14) multiple myeloma. Leukemia. Jan 2014;28(1 ):210-212.

Chen-Kiang S. Cell-cycle control of plasma cell differentiation and tumorigenesis. Immunological reviews. Aug 2003; 194:39-47. Amanna IJ, Slifka MK. Mechanisms that determine plasma cell lifespan and the duration of humoral immunity. Immunological reviews. Jul 2010;236:125-138. Tooze RM. A replicative self-renewal model for long-lived plasma cells: questioning irreversible cell cycle exit. Frontiers in immunology. 2013;4:460.

Chang-Yew Leow C, Gerondakis S, Spencer A. MEK inhibitors as a chemotherapeutic intervention in multiple myeloma. Blood Cancer J. 2013;3:e105.

Ward AC, Touw I, Yoshimura A. The Jak-Stat pathway in normal and perturbed hematopoiesis. Blood. Jan 1 2000;95(1 ): 19-29.

Catlett-Falcone R, Landowski TH, Oshiro MM, et al. Constitutive activation of Stat3 signaling confers resistance to apoptosis in human U266 myeloma cells. Immunity. Jan 1999; 10(1 ): 105-1 15.

Ramakrishnan V, Kimlinger T, Haug J, et al. Anti-myeloma activity of Akt inhibition is linked to the activation status of PI3K/Akt and MEK/ERK pathway. PLoS One. 2012;7(1 1 ):e50005.

Demchenko YN, Kuehl WM. A critical role for the NFkB pathway in multiple myeloma. Oncotarget. May 2010; 1 (1 ):59-68.

Walker BA, Boyle EM, Wardell CP, et al. Mutational Spectrum, Copy Number Changes, and Outcome: Results of a Sequencing Study of Patients With Newly Diagnosed Myeloma. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. Aug 17 2015.

Movilla N, Crespo P, Bustelo XR. Signal transduction elements of TC21 , an oncogenic member of the R-Ras subfamily of GTP-binding proteins. Oncogene. Oct 1999; 18(43):5860-5869.

Matthias P, Rolink AG. Transcriptional networks in developing and mature B cells. Nat Rev Immunol. Jun 2005;5(6):497-508.

BroyI A, Hose D, Lokhorst H, et al. Gene expression profiling for molecular classification of multiple myeloma in newly diagnosed patients. Blood. Oct 7 2010; 1 16(14):2543-2553.

Claims

CLAIMS What is claimed is:

1. A classification system for multiple myeloma, the classification system

comprising the following 10 classes: t(11;14) CD20/PAX5+, t(11;14) CD20/PAX5- , t(11;14) D1-HRD:RRAS2+, t(11;14) D1-HRD:RRAS2-, t(11;14) D2:RRAS2+, t(11;14) D2:RRAS2-, t(4;14), t(14;16), t(14;20) and t(6;14).

2. The classification system of claim 1 , wherein the classification system comprises one or more probes used to detect expression of ARID5A, BACE2, CCND1, CCND2, CCND3, CDK6, DSG2, FGFR3, ISL2, ITGB7, LAMP5, MAF, MAFB, MS4A1, NES, NUAK1, PAX5, PTP4A3, RRAS2, S100A4, SLC8A1, SULF2, USP49, VREB3 and WHSC1.

3. The classification system of claim 1 , wherein the classification system comprises one or more of the probes listed in Table 5:

4. The classification system of claim 1 , wherein the classification system further comprises probes used to detect vFISH designations for 1 p-, 1 q+, 13q-, 17p- and hyperdiploidy (HRD).

5. The classification system of claim 4, wherein probes to detect HRD are selected from the group consisting of ISL2, CCNE1, THG1L, ELOVL7, RPL35A, CCRL2, TNFSF10, C19orf12, LOC100506548 /// RPL37, ESRRA, ATP5L, IP07, RC3H2, SMIM7, SLC12A9, RPL13A /// SNORD32A /// SNORD33 /// SNORD34 /// SNORD35A, RNF14, and GYG1.

6. The classification system of claim 4, wherein probes to detect 1 p- are selected from the group consisting of CSDE1, ATP5F1, AHCYL1, LAMTOR5, RSBN1, MAN1A2, CSDE1, CEPT1, TRIM33, BCAS2, HIAT1, LAMTOR5, SARS,

CAPZA1, GNAI3, TMED5, AHCYL1, STXBP3, DRAM2, and MAN1A2.

7. The classification system of claim 4, wherein probes to detect 1 q+ are selected from the group consisting of UBE2Q1, TMEM183A /// TMEM183B, DESI2, PSMD4, PRCC, DCAF6, MRPL9, FBX028, SNAPIN, SLC19A2, VPS72,

GPR89B, ETV3, GPR89A /// GPR89B /// GPR89C /// LOC101060247 ///

LOC101060636, MRPL24, LOC101060511 /// PRKAB2, TPM3, CDC42SE1, GNPAT, and TMEM189 /// TMEM189-UBE2V1 /// UBE2V1.

8. The classification system of claim 4, wherein probes to detect 13q- are selected from the group consisting of DIS3, PCID2, UTP14C, MZT1, STK24, SUGT1, SUCLA2, KPNA3, CDC16, NDFIP2, COMMD6, COG6, SUPT20H, TRIM13, MED4, KBTBD6, ANKRD10, CLN5, and NUDT15.

9. The classification system of claim 4, wherein probes to detect 17p- are selected form the group consisting of SAT2, TP53, ZBTB4, CYB5D1, FXR2, TMEM256, NDEL1, RANGRF, MAP2K4, MED11, MIS12, C17orf85, CYB5D1, FXR2, DERL2, KIAA0753, ELP5, VAMP2, and MED31.

10. The classification system of claim 4, wherein the classification system further comprises one or more of the probes listed in Table 3:

226330 s at SUPT20H chr13q13.3

230192 at TRIM13 chr13q14

217843 s at MED4 chr13q14.2

226479 at KBTBD6 chr13q14.1 1

218093 s at ANKRD10 chr13q34

204085 s at CLN5 Chr13q21.1-q32

219347 at NUDT15 chr13q14.2

Model EventID Symbol Chromosome

17p- 225272 at SAT2 chr17p13.1

201746 at TP53 chr17p13.1

225629 s at ZBTB4 chr17p13.1

226833 at CYB5D1 chr17p13.1

35265 at FXR2 chr17p13.1

227063 at TMEM256 chr17p13.1

208093 s at NDEL1 chr17p13.1

218526 s at RANGRF chr17p13.1

203266 s at MAP2K4 chr17p12

226958 s at MED11 chr17p13.2

221559 s at MIS12 chr17p13.2

218896 s at C17orf85 chr17p13.2

155271 1 a at CYB5D1 chr17p13.1

203172 at FXR2 chr17p13.1

231002 s at — —

218333 at DERL2 chr17p13.2

204711 at KIAA0753 chr17p13.1

223286 at ELP5 chr17p13.1

201556 s at VAMP2 chr17p13.1

222867 s at MED31 chr17p13.1

1 1 . The classification system of claim 4, wherein the classification system further comprises one or more of the probes listed in Table 8:

208374 s at 225463 x at 226330 s at 203172 at 200994 at 227578 at

201 180 s at 218270 at 230192 at 231002 s at 220201 at 217542 at

202195 s at 225278 at 217843 s at 218333 at 224717 s at 42361_g_at

200849 s at 222976 s at 226479 at 20471 1 at 223995 at 235571 at

203310 at 229120 s at 218093 s at 223286 at 200715 x at 244616 x at

225228 at 201956 s at 204085 s at 201556 s at 201823 s at 200770 s at

217921 at 201003 x at 219347 at 222867 s at 21 1275 s at 213036 x at

207855 s at 224634 at 223412 at 219435 at 244841 at 222022 at

200848 at 224513 s at 202304 at 230274 s at 205356 at 205903 s at

227278 at 221486 at 235339 at 224900 at 200993 at 233413 at

238653 at 209435 s at 209009 at 201530 x at 208289 s at 227584 at

202194 at 204478 s at 214252 s at 226703 at 214329 x at 208065 at

202776 at 220642 x at 224799 at 204690 at 213203 at 1560570 a at

225226 at 201786 s at 229298 at 211787 s at 213041 s at 209877 at

1555543 a at 222443 s at 202259 s at 220606 s at 219081 at 207534 at

223331 s at 209382 at 229078 s at 214552 s at 218038 at 234051 at

228026 at 21 1594 s at 218203 at 214805 at 224814 at 217876 at

220306 at 202244 at 218352 at 209295 at 212129 at 230777 s at

217920 at 201377 at 229970 at 223081 at 226189 at 231477 at

202647 s at 219696 at 202548 s at 235857 at 238590 x at 1560177 at

231894 at 210386 s at 218420 s at 203344 s at 229838 at 240599 x at

207157 s at 223470 at 236446 at 228597 at 204949 at 215515 at

223671 x at 221497 x at 223306 at 209208 at 213737 x at

22681 1 at 202243 s at 231871 at 201557 at 227056 at

227981 at 235196 at 222239 s at 200855 at 226902 at

201421 s at 201612 at 215096 s at 209141 at 224495 at

243501 at 224314 s at 240574 at 228183 s at 213348 at

212360 at 1552617 a at 233647 s at 209350 s at 234339 s at

202362 at 223046 at 228937 at 229551 x at 205672 at

224985 at 213027 at 221899 at 235728 at 202759 s at

228661 s at 228238 at 214028 x at 212601 at 209329 x at

212629 s at 209825 s at 234993 at 22451 1 s at 213223 at

201274 at 218728 s at 217732 s at 205055 at 219362 at

218040 at 201425 at 206235 at 202632 at 212716 s at

207654 x at 206468 s at 221503 s at 226787 at 223175 s at

238524 at 2261 15 at 236067 at 236241 at 237817 at

1554154 at 230917 at 236122 at 228482 at 227394 at

212893 at 228852 at 15571 18 a at 204170 s at 209533 s at

218462 at 205788 s at 239314 at 204162 at 235027 at

203346 s at 225786 at 229718 at 230012 at 218943 s at

235294 at 225244 at 213375 s at 202900 s at 227146 at

215285 s at 220059 at 213357 at 200854 at 200089 s at

215333 x at 202337 at 204831 at 2151 14 at 441 1 1 at

209204 at 41329 at 37512 at 203871 at 227385 at

238087 at 205661 s at 201672 s at 202282 at 244066 at

204646 at 202220 at 208753 s at 3281 1 at 220202 s at

207237 at 212839 s at 219562 at 229833 at 202387 at

36499 at 205945 at 226050 at 238733 at 228561 at

212704 at 238056 at 204226 at 224512 s at 227099 s at

213689 x at 203035 s at 223404 s at 201899 s at 212843 at

212291 at 225401 at 229751 s at 226262 at 38269 at 205087 at 223555 at 225570 at 209440 at 227518 at

202234 s at 223322 at 201551 s at 226996 at 213844 at

201420 s at 2141 13 s at 221561 at 215667 x at 244406 at

205371 s at 217894 at 213882 at 202580 x at 1554273 a at

202377 at 212530 at 204064 at 229175 at 229720 at

244481 at 228032 s at 205632 s at 204146 at 207480 s at

223584 s at 228191 at 15531 13 s at 202534 x at 204950 at

203362 s at 222987 s at 202413 s at 228361 at 219348 at

223849 s at 228385 at 222868 s at 2021 10 at 202760 s at

228297 at 229828 at 209934 s at 219148 at 213183 s at

212678 at 1558027 s at 221550 at 202284 s at 205363 at

227068 at 205607 s at 1556821 x at 212247 at 216396 s at

239596 at 221222 s at 238623 at 218308 at 44696 at

1552738 a at 206332 s at 223339 at 209190 s at 200083 at

220840 s at 15531 18 at 226224 at 208890 s at 219343 at

220948 s at 210589 s at 220768 s at 219123 at 44669 at

212676 at 226482 s at 212585 at 1557128 at 224655 at

227680 at 219373 at 204459 at 219417 s at 230277 at

223275 at 236280 at 214906 x at 205873 at 40020 at

207168 s at 216593 s at 225795 at 204817 at 1555575 a at

202304 at 214719 at 227465 at 203336 s at 231930 at

226568 at 217797 at 201746 at 202154 x at 207452 s at

204032 at 212408 at 1557066 at 201092 at 224948 at

218567 x at 1554057 at 235215 at 227085 at 202649 x at

218979 at 222605 at 202268 s at 44563 at 21 1563 s at

222734 at 229544 at 222077 s at 242153 at 212648 at

226488 at 228386 s at 216682 s at 228040 at 201573 s at

202236 s at 213261 at 225424 at 1552733 at 205774 at

218355 at 226384 at 214946 x at 228427 at 1558622 a at

202690 s at 224709 s at 223026 s at 229021 at 221706 s at

214526 x at 200797 s at 203745 at 22971 1 s at 224151 s at

202954 at 231907 at 205804 s at 200737 at 1554456 a at

224523 s at 200652 at 203358 s at 227578 at 1569022 a at

213088 s at 241262 at 213186 at 243191 at 46270 at

213647 at 210417 s at 201230 s at 237429 at 229693 at

225926 at 226233 at 58308 at 212597 s at 21 1988 at

203746 s at 1557166 at 1558953 s at 203159 at 226915 s at

2351 13 at 226447 at 228035 at 202842 s at 235035 at

223016 x at 235603 at 213261 at 214430 at 205578 at

241853 at 210978 s at 233750 s at 205191 at 231 188 at

219588 s at 1554575 a at 230465 at 210314 x at 236227 at

201614 s at 225763 at 221940 at 225017 at 209285 s at

226718 at 224855 at 235303 at 1557129 a at 202604 x at

222916 s at 1554049 s at 238919 at 204325 s at 227972 at

205369 x at 218518 at 214703 s at 213334 x at 1554455 at

201477 s at 204742 s at 201602 s at 1555874 x at 1554188 at

241922 at 1555226 s at 200699 at 224751 at 200758 s at

225928 at 218873 at 211069 s at 222396 at 205452 at

203283 s at 220992 s at 222869 s at 201 191 at 214895 s at

230766 at 234995 at 201247 at 203262 s at 200810 s at

226384 at 224669 at 224689 at 201 1 15 at 235521 at 222396 at 201912 s at 209628 at 200600 at 20081 1 at

218499 at 239825 at 235463 s at 216961 s at 212233 at

201832 s at 201273 s at 218373 at 217542 at 221483 s at

222468 at 225793 at 202225 at 201312 s at 2141 16 at

218882 s at 224791 at 213098 at 1561554 at 209282 at

233294 at 223624 at 203193 at 218494 s at 228592 at

65086 at 204923 at 227767 at 1557192 at 207102 at

203391 at 215438 x at 243539 at 224185 at 201260 s at

227267 at 235747 at 229067 at 1560391 at 213256 at

219562 at 221824 s at 219000 s at 207038 at

233917 s at 209546 s at 200815 s at 201310 s at

205191 at 226527 at 218399 s at 222868 s at

232066 x at 214615 at 224578 at 206662 at

201954 at 229670 at 209946 at 205208 at

201465 s at 218068 s at 230072 at 21 1012 s at

236122 at 222667 s at 221 176 x at 225425 s at

210707 x at 219875 s at 232706 s at 201259 s at

239038 at 243780 at 204197 s at 231997 at

2261 19 at 21 1080 s at 237643 at 1568997 at

219473 at 241433 at 206835 at 209640 at

1554878 a at 214228 x at 223480 s at 229063 s at

200738 s at 218401 s at 1564337 at 221490 at

202915 s at 1553856 s at 203832 at 226732 at

203832 at 219313 at 205240 at 208867 s at

203745 at 228493 at 221620 s at 210755 at

209537 at 1559776 at 231383 at 220182 at

203022 at 239536 at 218363 at

1553709 a at 212800 at 1555872 a at

235507 at 218407 x at 229519 at

202235 at 1553314 a at 202824 s at

50965 at 223402 at 201558 at

241720 at 219922 s at 201 1 1 1 at

223294 at 243570 at 213647 at

1558305 at 226615 at 235299 at

239982 at 229447 x at 235507 at

203640 at 229074 at 219843 at

219531 at 212594 at 243634 at

235780 at 201959 s at 218472 s at

202706 s at 212483 at 201068 s at

200600 at 203523 at 244519 at

225792 at 233248 at 209149 s at

225153 at 220193 at 218966 at

209308 s at 235292 at 218380 at

211759 x at 208759 at 230840 at

239409 at 220338 at 205486 at

242374 at 238005 s at 1558304 s at

231909 x at 1558508 a at 217329 x at

203764 at 213658 at 224177 s at

224413 s at 226962 at 238247 at

215380 s at 205076 s at 1560339 s at

223568 s at 219056 at 49452 at 209189 at 229145 at 216548 x at

223024 at 223625 at 214180 at

2391 18 at 212409 s at 213302 at

213340 s at 206919 at 220800 s at

2021 10 at 228619 x at 225441 x at

1554520 at 243843 at 219540 at

201590 x at 206978 at 236655 at

204342 at 229303 at 229131 at

235612 at 223265 at 230633 at

21231 1 at 213279 at 218669 at

211071 s at 201804 x at 200980 s at

227414 at 2381 14 at 239082 at

212522 at 223548 at 203396 at

229595 at 203143 s at 224896 s at

234295 at 212604 at 223041 at

1561884 at 56197 at 1553434 at

217886 at 213902 at 227846 at

207556 s at 201930 at 238105 x at

225201 s at 239843 at 240615 at

214757 at 216250 s at 234454 at

1558957 s at 225012 at 229436 x at

229000 at 217967 s at 225533 at

200597 at 201061 s at 210378 s at

1552973 at 227776 at 205662 at

203077 s at 53076 at 1556457 s at

244383 at 209868 s at 238021 s at

202852 s at 227414 at 205671 s at

223037 at 206674 at 1568656 at

222976 s at 230605 at 204727 at

202209 at 238574 at 224461 s at

209946 at 227031 at 243004 at

223434 at 213155 at 225741 at

229693 at 213988 s at 217399 s at

230350 at 207350 s at 1557838 at

215313 x at 204603 at 1566597 at

1561762 s at 200978 at 2081 17 s at

229025 s at 209536 s at 204054 at

219037 at 231832 at 1553373 at

218558 s at 214056 at 208407 s at

208943 s at 218788 s at 222487 s at

217047 s at 218723 s at 1553258 at

213044 at 224885 s at 237392 at

223542 at 1552892 at 242723 at

238758 at 231 103 at 243269 s at

225302 at 228745 at 221474 at

235343 at 222026 at 1566631 at

211069 s at 227962 at 231 183 s at

210034 s at 214450 at 212096 s at

238615 at 225722 at 204588 s at

241396 at 216885 s at 201798 s at

222889 at 213180 s at 217356 s at 202225 at 200980 s at 223522 at

243671 at 241321 at 217795 s at

243527 at 212418 at

218592 s at 220214 at

204759 at 226760 at

208632 at 238781 at

201 141 at 201555 at

213912 at 238070 at

208370 s at 235039 x at

231962 at 212642 s at

205270 s at 200658 s at

203197 s at 203641 s at

209289 at 50965 at

212797 at 221704 s at

227134 at 226893 at

49452 at 219248 at

202201 at 217513 at

202580 x at 218994 s at

222418 s at 235064 s at

227928 at 2441 10 at

209300 s at 221234 s at

226255 at 240310 at

229928 at 1568597 at

226896 at

210605 s at

242560 at

219648 at

200980 s at

229969 at

238881 at

227552 at

202934 at

223804 s at

217356 s at

244593 at

222622 at

226915 s at

225344 at

222843 at

230362 at

225702 at

213951 s at

224628 at

1553587 a at

217929 s at

12. A method of classifying multiple myeloma in a subject diagnosed with multiple myeloma, the method comprising: a. detecting an aberration associated with t(11 ; 14) CD20/PAX5+, t(11 ; 14) CD20/PAX5- t(11;14) D1-HRD:RRAS2+, t(11;14) D1-HRD:RRAS2- t(11;14) D2:RRAS2+, t(11;14) D2:RRAS2- t(4;14), t(14;16), t(14;20) and t(6;14) in a biological sample obtained from a subject; and

b. classifying a subject into one of the 10 classes selected from the group consisting of t(11 ; 14) CD20/PAX5+, t(11 ; 14) CD20/PAX5-, t(11 ; 14) D1 - HRD:RRAS2+, t(11;14) D1-HRD:RRAS2-, t(11;14) D2:RRAS2+, t(11;14) D2:RRAS2- t(4;14), t(14;16), t(14;20) and t(6;14) based on the aberration detected.

13. The method of claim 12, wherein an aberration is detected via gene expression.

14. The method of claim 13, wherein gene expression of the following genes are detected: ARID5A, BACE2, CCND1, CCND2, CCND3, CDK6, DSG2, FGFR3, ISL2, ITGB7, LAMP5, MAF, MAFB, MS4A1, NES, NUAK1, PAX5, PTP4A3, RRAS2, S100A4, SLC8A1, SULF2, USP49, VREB3 and WHSC1.

15. The method of claim 14, wherein gene expression is detected using one or more of the probes listed in Table 5:

16. The method of claim 13, wherein an aberration is also detected via vFISH

designations for 1 p- 1 q+, 13q- 17p- and hyperdiploidy (HRD).

17. The method of claim 16, wherein genes to detect HRD are selected from the group consisting of ISL2, CCNE1, THG1L, ELOVL7, RPL35A, CCRL2,

TNFSF10, C19orf12, LOC100506548 /// RPL37, ESRRA, ATP5L, IP07, RC3H2, SMIM7, SLC12A9, RPL13A /// SNORD32A /// SNORD33 /// SNORD34 /// SNORD35A, RNF14, and GYG1.

18. The method of claim 16, wherein genes to detect 1 p- are selected from the group consisting of CSDE1, ATP5F1, AHCYL1, LAMTOR5, RSBN1, MAN1A2, CSDE1, CEPT1, TRIM33, BCAS2, HIAT1, LAMTOR5, SARS, CAPZA 1, GNAI3, TMED5, AHCYL1, STXBP3, DRAM2, and MAN1A2.

19. The method of claim 16, wherein genes to detect 1 q+ are selected from the

group consisting of UBE2Q1, TMEM183A /// TMEM183B, DESI2, PSMD4, PRCC, DCAF6, MRPL9, FBX028, SNAPIN, SLC19A2, VPS72, GPR89B, ETV3, GPR89A /// GPR89B /// GPR89C /// LOC101060247 /// LOC101060636,

MRPL24, LOC101060511 /// PRKAB2, TPM3, CDC42SE1, GNPAT, and

TMEM189 /// TMEM189-UBE2 V1 /// UBE2V1.

20. The method of claim 16, wherein genes to detect 13q- are selected from the

group consisting of DIS3, PCID2, UTP14C, MZT1, STK24, SUGT1, SUCLA2, KPN A3, CDC16, NDFIP2, COMMD6, COG6, SUPT20H, TRIM 13, MED4, KBTBD6, ANKRD10, CLN5, and NUDT15.

21 . The method of claim 16, wherein genes to detect 17p- are selected form the group consisting of SAT2, TP53, ZBTB4, CYB5D1, FXR2, TMEM256, NDEL1, RANGRF, MAP2K4, MED11, MIS12, C17orf85, CYB5D1, FXR2, DERL2, KIAA0753, ELP5, VAMP2, and MED31.

22. The method of claim 16, wherein vFISH designations for 1 p- 1 q+, 13q- 17p- and hyperdiploidy (HRD) is detected using one or more of the probes listed in Table 3:

225769 at C0G6 chr13q14.1 1

209659 s at CDC16 chr13q34

226330 s at SUPT20H chr13q13.3

230192 at TRIM13 chr13q14

217843 s at MED4 chr13q14.2

226479 at KBTBD6 chr13q14.1 1

218093 s at ANKRD10 chr13q34

204085 s at CLN5 Chr13q21.1-q32

219347 at NUDT15 chr13q14.2

Model EventID Symbol Chromosome

17p- 225272 at SAT2 chr17p13.1

201746 at TP53 chr17p13.1

225629 s at ZBTB4 chr17p13.1

226833 at CYB5D1 chr17p13.1

35265 at FXR2 chr17p13.1

227063 at TMEM256 chr17p13.1

208093 s at NDEL1 chr17p13.1

218526 s at RANGRF chr17p13.1

203266 s at MAP2K4 chr17p12

226958 s at MED11 chr17p13.2

221559 s at MIS12 chr17p13.2

218896 s at C17orf85 chr17p13.2

155271 1 a at CYB5D1 chr17p13.1

203172 at FXR2 chr17p13.1

231002 s at — —

218333 at DERL2 chr17p13.2

204711 at KIAA0753 chr17p13.1

223286 at ELP5 chr17p13.1

201556 s at VAMP2 chr17p13.1

222867 s at MED31 chr17p13.1

23. The method of claim 13, wherein an aberration is also detected via vFISH

designations for 1 p- 1 q+, 13q- 17p-, hyperdiploidy (HRD) and TP53 mutation using one or more of the probes listed in Table 8:

203053 at 209681 at 224801 at 226958 s at 224767 at 209295 at

225222 at 202261 at 225312 at 221559 s at 1487 at 217640 x at

202300 at 223531 x at 225769 at 218896 s at 208745 at 202284 s at

200802 at 227200 at 209659 s at 155271 1 a at 200992 at 235534 at

208374 s at 225463 x at 226330 s at 203172 at 200994 at 227578 at

201 180 s at 218270 at 230192 at 231002 s at 220201 at 217542 at

202195 s at 225278 at 217843 s at 218333 at 224717 s at 42361_g_at

200849 s at 222976 s at 226479 at 20471 1 at 223995 at 235571 at

203310 at 229120 s at 218093 s at 223286 at 200715 x at 244616 x at

225228 at 201956 s at 204085 s at 201556 s at 201823 s at 200770 s at

217921 at 201003 x at 219347 at 222867 s at 21 1275 s at 213036 x at

207855 s at 224634 at 223412 at 219435 at 244841 at 222022 at

200848 at 224513 s at 202304 at 230274 s at 205356 at 205903 s at

227278 at 221486 at 235339 at 224900 at 200993 at 233413 at

238653 at 209435 s at 209009 at 201530 x at 208289 s at 227584 at

202194 at 204478 s at 214252 s at 226703 at 214329 x at 208065 at

202776 at 220642 x at 224799 at 204690 at 213203 at 1560570 a at

225226 at 201786 s at 229298 at 211787 s at 213041 s at 209877 at

1555543 a at 222443 s at 202259 s at 220606 s at 219081 at 207534 at

223331 s at 209382 at 229078 s at 214552 s at 218038 at 234051 at

228026 at 21 1594 s at 218203 at 214805 at 224814 at 217876 at

220306 at 202244 at 218352 at 209295 at 212129 at 230777 s at

217920 at 201377 at 229970 at 223081 at 226189 at 231477 at

202647 s at 219696 at 202548 s at 235857 at 238590 x at 1560177 at

231894 at 210386 s at 218420 s at 203344 s at 229838 at 240599 x at

207157 s at 223470 at 236446 at 228597 at 204949 at 215515 at

223671 x at 221497 x at 223306 at 209208 at 213737 x at

22681 1 at 202243 s at 231871 at 201557 at 227056 at

227981 at 235196 at 222239 s at 200855 at 226902 at

201421 s at 201612 at 215096 s at 209141 at 224495 at

243501 at 224314 s at 240574 at 228183 s at 213348 at

212360 at 1552617 a at 233647 s at 209350 s at 234339 s at

202362 at 223046 at 228937 at 229551 x at 205672 at

224985 at 213027 at 221899 at 235728 at 202759 s at

228661 s at 228238 at 214028 x at 212601 at 209329 x at

212629 s at 209825 s at 234993 at 22451 1 s at 213223 at

201274 at 218728 s at 217732 s at 205055 at 219362 at

218040 at 201425 at 206235 at 202632 at 212716 s at

207654 x at 206468 s at 221503 s at 226787 at 223175 s at

238524 at 2261 15 at 236067 at 236241 at 237817 at

1554154 at 230917 at 236122 at 228482 at 227394 at

212893 at 228852 at 15571 18 a at 204170 s at 209533 s at

218462 at 205788 s at 239314 at 204162 at 235027 at

203346 s at 225786 at 229718 at 230012 at 218943 s at

235294 at 225244 at 213375 s at 202900 s at 227146 at

215285 s at 220059 at 213357 at 200854 at 200089 s at

215333 x at 202337 at 204831 at 2151 14 at 441 1 1 at

209204 at 41329 at 37512 at 203871 at 227385 at

238087 at 205661 s at 201672 s at 202282 at 244066 at

204646 at 202220 at 208753 s at 3281 1 at 220202 s at

207237 at 212839 s at 219562 at 229833 at 202387 at 36499 at 205945 at 226050 at 238733 at 228561 at

212704 at 238056 at 204226 at 224512 s at 227099 s at

213689 x at 203035 s at 223404 s at 201899 s at 212843 at

212291 at 225401 at 229751 s at 226262 at 38269 at

205087 at 223555 at 225570 at 209440 at 227518 at

202234 s at 223322 at 201551 s at 226996 at 213844 at

201420 s at 2141 13 s at 221561 at 215667 x at 244406 at

205371 s at 217894 at 213882 at 202580 x at 1554273 a at

202377 at 212530 at 204064 at 229175 at 229720 at

244481 at 228032 s at 205632 s at 204146 at 207480 s at

223584 s at 228191 at 15531 13 s at 202534 x at 204950 at

203362 s at 222987 s at 202413 s at 228361 at 219348 at

223849 s at 228385 at 222868 s at 2021 10 at 202760 s at

228297 at 229828 at 209934 s at 219148 at 213183 s at

212678 at 1558027 s at 221550 at 202284 s at 205363 at

227068 at 205607 s at 1556821 x at 212247 at 216396 s at

239596 at 221222 s at 238623 at 218308 at 44696 at

1552738 a at 206332 s at 223339 at 209190 s at 200083 at

220840 s at 15531 18 at 226224 at 208890 s at 219343 at

220948 s at 210589 s at 220768 s at 219123 at 44669 at

212676 at 226482 s at 212585 at 1557128 at 224655 at

227680 at 219373 at 204459 at 219417 s at 230277 at

223275 at 236280 at 214906 x at 205873 at 40020 at

207168 s at 216593 s at 225795 at 204817 at 1555575 a at

202304 at 214719 at 227465 at 203336 s at 231930 at

226568 at 217797 at 201746 at 202154 x at 207452 s at

204032 at 212408 at 1557066 at 201092 at 224948 at

218567 x at 1554057 at 235215 at 227085 at 202649 x at

218979 at 222605 at 202268 s at 44563 at 21 1563 s at

222734 at 229544 at 222077 s at 242153 at 212648 at

226488 at 228386 s at 216682 s at 228040 at 201573 s at

202236 s at 213261 at 225424 at 1552733 at 205774 at

218355 at 226384 at 214946 x at 228427 at 1558622 a at

202690 s at 224709 s at 223026 s at 229021 at 221706 s at

214526 x at 200797 s at 203745 at 22971 1 s at 224151 s at

202954 at 231907 at 205804 s at 200737 at 1554456 a at

224523 s at 200652 at 203358 s at 227578 at 1569022 a at

213088 s at 241262 at 213186 at 243191 at 46270 at

213647 at 210417 s at 201230 s at 237429 at 229693 at

225926 at 226233 at 58308 at 212597 s at 21 1988 at

203746 s at 1557166 at 1558953 s at 203159 at 226915 s at

2351 13 at 226447 at 228035 at 202842 s at 235035 at

223016 x at 235603 at 213261 at 214430 at 205578 at

241853 at 210978 s at 233750 s at 205191 at 231 188 at

219588 s at 1554575 a at 230465 at 210314 x at 236227 at

201614 s at 225763 at 221940 at 225017 at 209285 s at

226718 at 224855 at 235303 at 1557129 a at 202604 x at

222916 s at 1554049 s at 238919 at 204325 s at 227972 at

205369 x at 218518 at 214703 s at 213334 x at 1554455 at

201477 s at 204742 s at 201602 s at 1555874 x at 1554188 at

241922 at 1555226 s at 200699 at 224751 at 200758 s at 225928 at 218873 at 211069 s at 222396 at 205452 at

203283 s at 220992 s at 222869 s at 201 191 at 214895 s at

230766 at 234995 at 201247 at 203262 s at 200810 s at

226384 at 224669 at 224689 at 201 1 15 at 235521 at

222396 at 201912 s at 209628 at 200600 at 20081 1 at

218499 at 239825 at 235463 s at 216961 s at 212233 at

201832 s at 201273 s at 218373 at 217542 at 221483 s at

222468 at 225793 at 202225 at 201312 s at 2141 16 at

218882 s at 224791 at 213098 at 1561554 at 209282 at

233294 at 223624 at 203193 at 218494 s at 228592 at

65086 at 204923 at 227767 at 1557192 at 207102 at

203391 at 215438 x at 243539 at 224185 at 201260 s at

227267 at 235747 at 229067 at 1560391 at 213256 at

219562 at 221824 s at 219000 s at 207038 at

233917 s at 209546 s at 200815 s at 201310 s at

205191 at 226527 at 218399 s at 222868 s at

232066 x at 214615 at 224578 at 206662 at

201954 at 229670 at 209946 at 205208 at

201465 s at 218068 s at 230072 at 21 1012 s at

236122 at 222667 s at 221 176 x at 225425 s at

210707 x at 219875 s at 232706 s at 201259 s at

239038 at 243780 at 204197 s at 231997 at

2261 19 at 21 1080 s at 237643 at 1568997 at

219473 at 241433 at 206835 at 209640 at

1554878 a at 214228 x at 223480 s at 229063 s at

200738 s at 218401 s at 1564337 at 221490 at

202915 s at 1553856 s at 203832 at 226732 at

203832 at 219313 at 205240 at 208867 s at

203745 at 228493 at 221620 s at 210755 at

209537 at 1559776 at 231383 at 220182 at

203022 at 239536 at 218363 at

1553709 a at 212800 at 1555872 a at

235507 at 218407 x at 229519 at

202235 at 1553314 a at 202824 s at

50965 at 223402 at 201558 at

241720 at 219922 s at 201 1 1 1 at

223294 at 243570 at 213647 at

1558305 at 226615 at 235299 at

239982 at 229447 x at 235507 at

203640 at 229074 at 219843 at

219531 at 212594 at 243634 at

235780 at 201959 s at 218472 s at

202706 s at 212483 at 201068 s at

200600 at 203523 at 244519 at

225792 at 233248 at 209149 s at

225153 at 220193 at 218966 at

209308 s at 235292 at 218380 at

211759 x at 208759 at 230840 at

239409 at 220338 at 205486 at

242374 at 238005 s at 1558304 s at

231909 x at 1558508 a at 217329 x at 203764 at 213658 at 224177 s at

224413 s at 226962 at 238247 at

215380 s at 205076 s at 1560339 s at

223568 s at 219056 at 49452 at

209189 at 229145 at 216548 x at

223024 at 223625 at 214180 at

2391 18 at 212409 s at 213302 at

213340 s at 206919 at 220800 s at

2021 10 at 228619 x at 225441 x at

1554520 at 243843 at 219540 at

201590 x at 206978 at 236655 at

204342 at 229303 at 229131 at

235612 at 223265 at 230633 at

21231 1 at 213279 at 218669 at

211071 s at 201804 x at 200980 s at

227414 at 2381 14 at 239082 at

212522 at 223548 at 203396 at

229595 at 203143 s at 224896 s at

234295 at 212604 at 223041 at

1561884 at 56197 at 1553434 at

217886 at 213902 at 227846 at

207556 s at 201930 at 238105 x at

225201 s at 239843 at 240615 at

214757 at 216250 s at 234454 at

1558957 s at 225012 at 229436 x at

229000 at 217967 s at 225533 at

200597 at 201061 s at 210378 s at

1552973 at 227776 at 205662 at

203077 s at 53076 at 1556457 s at

244383 at 209868 s at 238021 s at

202852 s at 227414 at 205671 s at

223037 at 206674 at 1568656 at

222976 s at 230605 at 204727 at

202209 at 238574 at 224461 s at

209946 at 227031 at 243004 at

223434 at 213155 at 225741 at

229693 at 213988 s at 217399 s at

230350 at 207350 s at 1557838 at

215313 x at 204603 at 1566597 at

1561762 s at 200978 at 2081 17 s at

229025 s at 209536 s at 204054 at

219037 at 231832 at 1553373 at

218558 s at 214056 at 208407 s at

208943 s at 218788 s at 222487 s at

217047 s at 218723 s at 1553258 at

213044 at 224885 s at 237392 at

223542 at 1552892 at 242723 at

238758 at 231 103 at 243269 s at

225302 at 228745 at 221474 at

235343 at 222026 at 1566631 at

211069 s at 227962 at 231 183 s at 210034 s at 214450 at 212096 s at

238615 at 225722 at 204588 s at

241396 at 216885 s at 201798 s at

222889 at 213180 s at 217356 s at

202225 at 200980 s at 223522 at

243671 at 241321 at 217795 s at

243527 at 212418 at

218592 s at 220214 at

204759 at 226760 at

208632 at 238781 at

201 141 at 201555 at

213912 at 238070 at

208370 s at 235039 x at

231962 at 212642 s at

205270 s at 200658 s at

203197 s at 203641 s at

209289 at 50965 at

212797 at 221704 s at

227134 at 226893 at

49452 at 219248 at

202201 at 217513 at

202580 x at 218994 s at

222418 s at 235064 s at

227928 at 2441 10 at

209300 s at 221234 s at

226255 at 240310 at

229928 at 1568597 at

226896 at

210605 s at

242560 at

219648 at

200980 s at

229969 at

238881 at

227552 at

202934 at

223804 s at

217356 s at

244593 at

222622 at

226915 s at

225344 at

222843 at

230362 at

225702 at

213951 s at

224628 at

1553587 a at

217929 s at

24. The method of claim 14, wherein the subject is classified into the t(1 1 ; 14) CD20/PAX5+ class if

i. increased expression of CCND1, SLC8A 1 and/or MS4A1 and decreased expression of CDK6, SULF2 and/or CCND2 is detected; and

ii. increased expression of CD20 and PAX5 is detected.

25. The method of claim 24, wherein increased expression of VPREB3 and/or

decreased expression of LAMP5 is detected.

26. The method of claim 14, wherein the subject is classified into the t(1 1 ; 14)

CD20/PAX5- class if

i. increased expression of CCND1, SLC8A 1 and/or MS4A1 and decreased expression of CDK6, SULF2 and/or CCND2 is detected; and

ii. decreased expression of CD20 and PAX5 is detected.

27. The method of claim 14, wherein the subject is classified into the t(1 1 ; 14) D1 - HRD:RRAS2+ class if

i. increased expression of genes on odd number chromosomes

including ISL2 (chr15q) and/or CCRL2 (chr3p) and decreased expression of genes on 1 q including NES and/or S100A4 is detected; and

ii. increased expression of RRAS2 is detected.

28. The method of claim 27, wherein increased expression of SULF2 and/or FRZB and/or decreased expression of CCND2 and/or SOCS3 is detected.

29. The method of claim 14, wherein the subject is classified into the t(1 1 ; 14) D1 - HRD:RRAS2- class if

i. increased expression of genes on odd number chromosomes

including ISL2 (chr15q) and/or CCRL2 (chr3p) and decreased expression of genes on 1 q including NES and/or S100A4 is detected; and

ii. decreased expression of RRAS2 is detected.

30. The method of claim 29, wherein increased expression of LAMP5 is detected.

31 . The method of claim 14, wherein the subject is classified into t(1 1 ; 14)

D2:RRAS2+ class if

i. increased expression of CCND2, SOCS3 and/or PTP4A3 is detected;

ii. decreased expression of CCND1 is detected; and

iii. increased expression of RRAS2 is detected.

32. The method of claim 14, wherein the subject is classified into t(1 1 ; 14)

D2:RRAS2- class if

i. increased expression of CCND2, SOCS3 and/or PTP4A3 is

detected;

ii. decreased expression of CCND1 is detected; and

iii. decreased expression of RRAS2 is detected.

33. The method of claim 14, wherein the subject is classified into t(4; 14) class if i. increased expression of WHSC1, FGFR3, BACE2 and/or DSG2 and decreased expression of ISL2, CCND1 and/or FRZB is detected.

34. The method of claim 33, wherein increased expression of NUAK1, ARID5A, SMARCA 1 and/or ITGB7 and decreased expression of DDK1 and/or SULF2 is detected.

35. The method of claim 14, wherein the subject is classified into t(14; 16) class if i. increased expression of MAF, NUAK1, ARID5A and/or SMARCA1 is detected.

36. The method of claim 14, wherein the subject is classified into t(14;20) class if i. increased expression of MAFB, NUAK1, ARID5A and/or ITGB7 is detected.

37. The method of claim 14, wherein the subject is classified into t(6;14) class if i. increased expression of CCND3 and/or USP49 is detected.

38. A method to prognose a subject diagnosed with multiple myeloma, the method comprising:

a. classifying the subject into one of the 10 classes selected from the group consisting of t(11 ; 14) CD20/PAX5+, t(11 ; 14) CD20/PAX5-, t(11 ; 14) D1 - HRD:RRAS2+, t(11;14) D1-HRD:RRAS2-, t(11;14) D2:RRAS2+, t(11;14) D2:RRAS2- t(4;14), t(14;16), t(14;20) and t(6;14); and

b. using a risk stratifier to further prognose the subject.

39. The method of claim 38, wherein the risk stratifier is GEP70.

40. A method to determine treatment of a subject diagnosed with multiple myeloma, the method comprising:

a. classifying the subject into one of the 10 classes selected from the group consisting of t(11 ; 14) CD20/PAX5+, t(11 ; 14) CD20/PAX5-, t(11 ; 14) D1 - HRD:RRAS2+, t(11;14) D1-HRD:RRAS2-, t(11;14) D2:RRAS2+, t(11;14) D2:RRAS2- t(4;14), t(14;16), t(14;20) and t(6;14); and

b. administering treatment based on the classification of the subject.