PS1374 PROGNOSIS IN ELDERLY MULTIPLE MYELOMA PATIENTS IN THE HOVON‐87/NMSG‐18 STUDY BASED ON REVISED ISS AND SKY92‐ISS

Background: The incidence of multiple myeloma (MM) increases with age with a median at diagnosis of 69 years. Patients older than 65 years old are often considered to be ineligible for stem‐cell transplantation. Treatment for older patients is lenalidomide and dexamethasone, or melphalan/prednisone with thalidomide or bortezomib. The HOVON‐87/NMSG‐18 trial showed that replacing thalidomide with lenalidomide, followed by maintenance therapy until disease progression, i.e. MPR‐R vs MPT‐T, did not result in improved progression free and overall survival (PFS and OS) of MPR‐R treated patients compared to MPT‐T treated patients (Zweegman et al. Blood 2016;127(9):1109–1116). MPR‐R treated patients did demonstrate significantly lower grade 3/4 neuropathy. Aims: The aim of this analysis was to evaluate the SKY92 gene expression classifier in comparison to revised‐ISS (R‐ISS) in elderly, non‐transplant eligible patients included in the HOVON‐87/NMSG‐18 trial using updated survival data. Methods: For 190 patients, CD138 purified plasma cells were available to determine the SKY92 risk score (median age 72 years, inter‐quartile range: 69 – 76). SKY92 scores were determined using the MMprofiler™ CE IVD assay. In addition, treatment arm, FISH and R‐ISS were analyzed by survival analysis for both PFS and OS. Hazard ratios (HR) are given with 95% confidence intervals in brackets. Likelihood ratio (LR) tests were used to evaluate the significance of each prognostic model. Results: At the time of analysis the median follow up was 6 yrs. The SKY92 classifier identified 14% of patients as high‐risk. The median PFS and OS of the high‐risk patients was 12 months and 19 months, respectively, compared to 23 (PFS) and 61 months (OS) for standard risk patients (OS, HR = 2.6 [1.6–4.1]; LR p = 7x10 –5 ; PFS, HR = 2.4 [1.6–3.7]; LR p = 7x 10 –5 ). Based on ISS (n ISS I/ II/ III = 48/ 92/ 46) and SKY92, 186 patients were classed into four risk groups: SKY92 high‐risk combined with any ISS stage (13%), SKY92 standard‐risk and ISS III (21%), SKY92 standard‐risk and ISS II (45%) and SKY92 standard risk and ISS I (21%; Kuiper et al., 2015; Blood, 126: 1996–2004). The median PFS of these respective groups was 11, 21, 22 and 25 months and the median OS was 18, 49, 56 and 88 mo (PFS: LR p‐value = 5x10 –3 ; OS: LR p‐value = 2x10 –4 ). Classifying in R‐ISS stages (n R‐ISS I/II/III = 12/129/28) demonstrated a median PFS of 13, 20 and 30 months (LR p‐value = 5x10 –3 ) and a median OS of 25, 54 and 78 months (LR p‐value = 1x10 –3 ). Factors independently associated with OS in the multivariate analysis were SKY92‐ISS, R‐ISS and del17p, whereas only SKY92‐ISS and R‐ISS remained independently associated with PFS. Eleven SKY92 high‐risk patients were treated with lenalidomide and demonstrated a median OS of 55 months compared to 17 months for thalidomide treated high‐risk patients (n = 15). The median OS in standard‐risk patients was 59 months (lenalidomide) vs 61 months (thalidomide). Using an interaction term in the Cox regression model, a significant difference in OS (p = 0.04) was found between the treatment arms conditional on SKY92 risk status. Summary/Conclusion: Also in non‐transplant eligible MM patients, the SKY92 classifier is a robust marker to identify high‐risk patients. The SKY92‐ISS has prognostic value independent of the revised ISS. In addition, SKY92 high‐risk patients appear to have a survival benefit of lenalidomide treatment over thalidomide treatment, which is not found for SKY92 standard risk patients.