Direct Second Autologous Stem Cell Transplantation Versus Re-Induction Followed by Transplantation in Relapsed...

Background: Whether patients with relapsed multiple myeloma (MM) should proceed directly to a second autologous stem cell transplantation (ASCT2) or receive re-induction chemotherapy beforehand remains uncertain. In real-world practice, treatment allocation is highly heterogeneous, and comparative data addressing the impact of pre-ASCT2 strategy on survival outcomes are limited, partly because ASCT2 is reserved for selected patients and is not routinely performed in all relapsed MM cases. We aimed to evaluate real-world outcomes of ASCT2 performed either directly at relapse or following re-induction chemotherapy in patients with relapsed MM, with a focus on progression-free survival after ASCT2 (PFS2) and overall survival (OS). Methods: We conducted a multicenter retrospective cohort study including 42 patients with relapsed MM who underwent ASCT2 between 2012 and 2024 across eight centers. Patients were grouped according to pre-ASCT2 management: those proceeding directly to ASCT2 without additional systemic therapy (direct-ASCT2, n = 21) and those receiving ≥ 1 line of salvage chemotherapy prior to ASCT2 (re-induction, n = 21). Survival outcomes were estimated using Kaplan-Meier methods and explored using multivariable Cox regression analyses. Results: The median PFS2 for the entire cohort was 19.7 months (95% CI, 7.8-31.6), with no statistically significant difference between the direct-ASCT2 and re-induction groups (21.3 vs. 13.8 months; p = 0.790). Median OS was 48.2 months (95% CI, 41.3-55.1). Although OS was significantly longer in the unadjusted analysis (51.7 vs. 39.6 months; p = 0.019), this difference was not maintained after multivariable adjustment. Post-ASCT2 response rates were comparable between groups, and day-100 transplant-related mortality was low at 2.4%. Conclusions: In this multicenter real-world cohort, pre-ASCT2 re-induction chemotherapy was not associated with a clear improvement in PFS2 compared with proceeding directly to ASCT2. Observed differences in unadjusted OS likely reflect confounding by indication and selection bias inherent to retrospective treatment allocation. These findings should therefore be interpreted as hypothesis-generating, but they add to the limited real-world evidence base on ASCT2, a salvage strategy that is typically applied to highly selected patients.