High-Risk Cytogenetics in Newly Diagnosed Multiple Myeloma: Prognostic Relevance of Co-Segregations and Analysis of the Role of Double Versus Single Autotransplantation

INTRODUCTION: Cytogenetic analysis performed by FISH at diagnosis of Multiple Myeloma (MM) allows to stratify patients in groups at different risk of progression or death. We analyzed a large cohort of patients with newly diagnosed MM to evaluate the prognostic impact of multiple high-risk cytogenetic abnormalities (HRAs), detected either as a single lesion (e.g. isolated) or multiple lesions (e.g. co-segregated).

METHODS: Probes for FISH analysis of CD138+ bone marrow plasma cells were chosen to detect t(4;14), t(6;14), t(11,14), t(14;16), t(14;20), deletion of 1p32, 13q14, 17p13, gain of 1q21 and hyperdiploidy (HD). HRAs were defined by the presence of one or more of t(4;14), t(14;16), t(14;20), del(17p), gain(1q) and del(1p). The analysis was performed in a large cohort of patients who received the diagnosis of MM in Italian centers and were enrolled in the phase III EMNO2/HO95 study. By study design, patients were randomized to receive either standard-dose intensification treatment with bortezomib-melphalan-prednisone (VMP) or high-dose intensification treatment with melphalan at 200 mg/m² (HDM) plus autotransplantation (ASCT), either single (ASCT-1) or double(ASCT-2). A second randomization to receive or not receive consolidation therapy was planned after the intensification phase, followed by lenalidomide maintenance in both arms.

RESULTS: A total of 461 patients were evaluable for all HRAs listed above. 151 of them were randomly assigned (stratification by ISS stage) to VMP and 310 to ASCT-1 or ASCT-2. HRAs were detected in 233/461 patients (50.5%): 155 (33.6%) gain(1q), 51 (11.1%) del(1p), 46 (10.0%) del(17p), 58 (12.6%) t(4;14), 20 (4.3%) t(14;16) and 6 (1.3%) t(14;20). 141/461 patients (30.6%) showed an isolated HRA: 76 (53.9%) gain(1q), 27 (19.1%) a high-risk translocation (HRT), 21 (14.9%) del(17p) and 17 (12.1%) del(1p). 81/461 patients (17.6%) carried 2 HRAs, most frequently a HRT co-segregating with gain(1q) (detected in 48.2%), while 3 HRAs were found in the remaining 11 (2.4%). With a median follow-up of 37.8 months, 3-year PFS estimates for patients without HRA or with 1, 2 and 3 HRAs were 80.3%, 53.5%, 45% and 21.8%, respectively (P<0.0001). The corresponding OS estimates were 95.4%, 81.7%, 73.5% and 40.9%, respectively (P<0.0001). All the HRAs, when evaluated either as isolated or in co-segregation, retained their adverse impact on PFS and OS in comparison with the groups lacking either 1 or more HRAs. Among isolated HRAs, del(17p) was the single lesion with the highest adverse impact on OS (HR=3.11; 95%CI=1.20-8.07; P=0.019). Notably, PFS benefit with ASCT-2 as compared with ASCT-1 was observed among the 148 patients with at least 1 HRA (3-year estimate: 61.8 vs 45.2; P= 0.0387). The corresponding OS rates with ASCT-2 vs ASCT-1 were 83.1 vs 67.6, P=0.04. Randomization to ASCT-2 significantly prolonged the 3-year PFS in comparison with ASCT-1 in patients with del(17p) (61% vs 38.7%; HR=0.27; CI=0.076-0.986; P=0.047). In comparison with isolated gain(1q), co-segregation of gain(1q) with a HRT was associated with significantly shorter 3-year PFS (33.3% vs 57.4%; HR 1.76; CI=1.02-3.03; P=0.041) and shorter OS rates (56.5% vs 85.7% ; HR=3.33; CI=1.52-7.30; P=0.027). Gain(1q) plus HRT was associated with lower OS (68.5% vs 89.6%; HR=4.78; 95%CI=1.09-20.93; P=0.038) compared with gain(1q) and del(1p). As expected, HD was more frequent in patients without HRA (59.2%) than in those with 1 HRA and 2 or more HRAs (53.2% and 34.8%). Differently from patients with 1 or ≥2 HRAs, PFS for patients with del(17p) and HD was significantly longer compared with that seen for patients with del(17) not associated with HD (43.9 vs 19.4 months, P=0.032).

CONCLUSIONS: 50% of patients with NDMM presented with one or more HRA and 20% with two or more of them. In patients with at least 1 HRA, randomization to ASCT-2 was superior over ASCT-1 in terms of prolonged PFS and OS. The association of gain(1q) with a HRT was the most frequently observed co-segregation of 2 HRAs and was correlated with significantly shorter PFS and OS compared with isolated gain(1q) and gain(1q) plus del(1p). The association of HD with del(17p) mitigated the adverse PFS of del(17p).