Treatment Patterns, Treatment Response, and Survival Outcomes in Patients with Multiple Myeloma: A Single-Center Real-World Study

Background/Objectives: Real-world data on multiple myeloma (MM) treatment patterns under evolving drug access are limited. We describe first-line treatment over two decades and factors associated with survival. Methods: We retrospectively analyzed 254 patients diagnosed between 2002 and 2024. Survival was estimated with the Kaplan–Meier method; multivariable Cox models included prespecified covariates and autologous stem cell transplantation (ASCT) as a time-dependent covariate, non-proportional effects were estimated as interval-specific hazard ratios (HRs), and first-line response was assessed in 6-month landmark analyses. Results: Median age was 68 years. First-line induction shifted from VAD (64.9% until 2014) to VCD (44.0% in 2015–2019) and VRD (45.1% in 2020–2024). Transplantation was performed in 79 patients (31.1%), 58 as first-line consolidation. With a median follow-up of 101.2 months, median overall survival (OS) was 54.6 months and time to next line of therapy or death (TNLT) was 11.4 months. ISS stage III (HR 1.98) and elevated LDH (HR 1.55) were associated with shorter OS. The associations of ASCT and induction regimen with OS changed over follow-up: ASCT was associated with lower mortality during the first 5 years after diagnosis (HR 0.18 for 0–24 months and 0.20 for 24–60 months) but not thereafter (HR 0.99), and VD induction, compared with VCD, with higher mortality during the same period (HR 4.16 and 3.90). Stable or progressive disease after induction was associated with shorter OS (HR 2.93) and shorter TNLT (HR 3.55). Laboratory ratios were not prognostic. Conclusions: ISS stage, LDH, induction regimen, transplantation and early response were associated with survival in this real-world cohort. Time-dependent modeling of ASCT removes immortal-time bias but not selection by age, fitness, organ function and response, and treatment eras changed; these associations should therefore not be interpreted as treatment effects.