Molecular Indicators of High-Risk Disease.

High-risk multiple myeloma (HRMM) is routinely defined by laboratory parameters alone or in combination in the Durie-Salmon and, more recently, the ISS staging systems. The Bartl grade, a cell morphology-based staging system, has seen limited use. The presence of abnormal cytogenetics, high BrdU labeling index, interphase FISH abnormalities, and flow cytometric measures have also been used. A molecular-based classification and risk stratification of MM may improve the definition of HRMM. Global gene expression profiling (GEP) with of CD138-selected plasma cells followed by unsupervised hierarchical cluster analysis revealed that MM comprises a spectrum of seven distinct reproducible subtypes. A validated molecular classification schema has been defined as follows: (MS = t(4;14); MF = t(14;16) or t(14;20); CD-1 = t(11;14) or t(6;14) and CD-2 = t(11;14) or t(6;14) with high CD20 and/or VPREB3), hyperdiploidy (HY = high DKK1, FRZB, NCAM1, TNFSF10), low bone disease (LB = NF-kB signature, high CCND2, CST6, and IL6R) and proliferation (PR = high MIK67, CCNB1, CCNB2, TOP2A, and TYMS). Correlating GEP with outcome in two independent cohorts permitted the identification of a high-risk signature (UAMS 17-gene model), present in approximately 13% of newly diagnosed disease. GEP and high-resolution comparative genomic hybridization in 92 cases confirmed that the altered expression of the 17 genes in the model is driven by 1q gains and 1p losses. This high-risk signature is evident in a subset of all 7 molecular subtypes and negatively influences outcome. For example, low-risk MS disease fares much better than high-risk MS disease. We recently reported that the addition of bortezomib to TT3 has significantly improved outcome in low-risk MS disease, thereby demonstrating the value of GEP in evaluating benefits of new treatments that might be otherwise masked. When subjected to multivariate analysis including the International Staging System (ISS) and a gene expression-based proliferation index (GEP PI), the UAMS 17-gene model remained a significant predictor of outcome. Mulligan and colleagues developed outcome classifiers for relapsed disease treated with single agent bortezomib or high dose dexamethasone improved upon the risk stratification provided by the ISS. These predictive models showed some specificity for bortezomib. Using U133A data from newly diagnosed disease treated with ASCT, the Mayo clinic group validated the UAMS 17-gene model, but also showed that the t(4;14) translocation remained a significant adverse variable. The IFM recently reported on a 15-gene model of high-risk (IFM 15-gene model) related to cell proliferation. Multivariate showed that the UAMS 17-gene model was significant in all datasets, while the IFM 15-model was significant in a limited number. This difference might be attributed to the dependence of the IFM model to cell proliferation. GEP on 71 paired diagnostic and relapse samples indicate that the UAMS 17-gene model score increases in 80% of the cases and a low-risk to high-risk conversion in 14 of 24 (58%) severely impacted post-relapse survival. Expression of TP53 is a surrogate for 17p13 deletion, and TP53 expression below a specific threshold (seen in approximately 10% of newly diagnosed disease) imparts a poor prognosis in low-risk – but not high-risk – MM, defined by the UAMS 17-gene model. In conclusion, while the majority of patients with MM can anticipate long-term disease control, approximately 25% of patients with molecularly defined HRMM do not benefit from current approaches.