Prospective target assessment and multimodal prediction of survival for personalized and risk-adapted treatment strategies in multiple myeloma in the GMMG-MM5 multicenter trial.

backgroundPersonalized and risk-adapted treatment strategies in multiple myeloma prerequisite feasibility of prospective assessment, reporting of targets, and prediction of survival probability in clinical routine. Our aim was first to set up and prospectively test our experimental and analysis strategy to perform advanced molecular diagnostics, i.e., interphase fluorescence in-situ hybridization (iFISH) in ≥ 90% and gene expression profiling (GEP) in ≥ 80% of patients within the first cycle of induction chemotherapy in a phase III trial, seen as prerequisite for target expression-based personalized treatment strategies. Secondly, whether the assessment of risk based on the integration of clinical, cytogenetic, and expression-based parameters ("metascoring") is possible in this setting and superior to the use of single prognostic factors.

methodsWe prospectively performed plasma cell purification, GEP using DNA-microarrays, and iFISH within our randomized multicenter GMMG-MM5-trial recruiting 604 patients between July 2010 and November 2013. Patient data were analyzed using our published gene expression report (GEP-R): after quality and identity control, integrated risk assessment (HM metascore) and targets were reported in clinical routine as pdf-document.

resultsBone marrow aspirates were obtained from 573/604 patients (95%) and could be CD138-purified in 559/573 (97.6%). Of these, iFISH-analysis was possible in 556 (99.5%), GEP in 458 (82%). Identity control using predictors for sex, light and heavy chain type allowed the exclusion of potential sample interchanges (none occurred). All samples passed quality control. As exemplary targets, IGF1R-expression was reported expressed in 33.1%, AURKA in 43.2% of patients. Risk stratification using an integrated approach, i.e., HM metascore, delineated 10/77/13% of patients as high/medium/low risk, transmitting into significantly different median progression-free survival (PFS) of 15 vs. 39 months vs. not reached (NR; P < 0.001) and median overall survival (OS) of 41 months vs. NR vs. NR (P < 0.001). Five-year PFS and OS-rates were 5/31/54% and 25/68/98%, respectively. Survival prediction by HM metascore (Brier score 0.132, P < 0.001) is superior compared with the current gold standard, i.e., revised ISS score (0.137, P = 0.005).

conclusionsProspective assessment and reporting of targets and risk by GEP-R in clinical routine are feasible in ≥ 80% of patients within the first cycle of induction chemotherapy, simultaneously allowing superior survival prediction.