Gene Expression Profiling (GEP) As a Prognostic Indicator of Multiple Myeloma, Irrespective of Disease Stage, Prior Treatment and Prior Transplant - a Single Center Experience

Background

Multiple myeloma (MM) risk is classified using fluorescence in situ hybridization (FISH) routine cytogenetics, and gene expression profiling (GEP). GEP is a clinically validated tool for prognostication of MM in newly diagnosed and relapsed patients (pts), its role in very heavily treated stages is less clear. A study (Nishihori et al, ASH, 2016) concluded that GEP is useful regardless of disease stage. We aim to determine the prognostic significance of GEP heavily pre-treated MM cohort.

Methods

299 pts with MyPRS^TM GEP results at Hackensack University Medical Center (NJ) between 2012 and 2016 were included. High risk GEP was defined with the validated cutoff score of >45.2, high risk FISH was determined according to the mSMART classification as del17p, t(14,16), t(14,20). Chi-square tests were performed on categorical data and Wilcoxon rank sum tests on continuous data. Overall survival (OS) and progression free survival (PFS) were performed using Kaplan-Meier.

Results

Median age was 63 years (range, 25-91), 170 pts (56%) were men, 81 pts (28%) had International Staging System (ISS) 3 and 191 pts (62%) had Durie-Salmon Staging (DSS) III. 245 pts (81%) had at least one bone marrow transplant (BMT). Median time from myeloma diagnosis to GEP was 30.1 (range 0.1-178.4) months. Most pts had GEP on diagnosis but 24% were heavily pretreated (range, 1-6 lines of treatment), and 13% had a BMT prior to GEP testing. 77 pts (25%) had high-risk disease based on GEP with a median score of 36.5 (range, 9.03-84.05). Molecular subtypes were hyperdiploidy (HY) 31%, cyclin family (CD)-2 20%, low bone disease (LB) 13%, proliferation (PR) 11%, MAF associated (MF) 10%, CD-1 8%, and MMSET associated (MS) 6%. Median OS and PFS were significantly inferior in patients with high-risk by GEP, 25 vs 33 months (p<0.021), 13 vs 21 months (p<0.004) respectively. Molecular subtypes were not individual predictors of OS (p<0.2). When MF, MS and PR subtypes were combined, they had inferior OS compared to other subtypes but this was not statistically significant (p<0.3).

Conclusions

GEP can accurately predict risk status in heavily pre-treated and post-transplant MM pts. In conclusion, our data confirm that GEP is a good prognostic indicator in MM pts irrespective of disease stage and prior treatments. Further prospective studies with larger samples and correlating with baseline and intra-treatment GEP are needed.