Long-Term Clinical Outcomes of Allogeneic Stem Cell Transplantation in Multiple Myeloma

Allogeneic stem cell transplantation (allo-SCT), which provides a tumor-free graft, is an alternative approach to autologous transplantation for multiple myeloma (MM) that offers the possibility of cure through a graft vs myeloma effect (GVM). However, the inherent non relapse mortality (NRM) and the high post-transplant relapse rate are the major shortcomings of this strategy which continues to have a controversial role in MM treatment.

To highlight the long-term clinical outcomes of allo-SCT, we performed a retrospective analysis on 102 patients (pts) with MM who received at our Institution either a myeloablative (MA) (74 pts) or a non-myeloablative (NMA) (28 pts) conditioning regimen between 1990 and 2014. The MA regimen consisted in low dose TBI and cyclophosphamide (cyclo) ± melphalan (mel) or busulfan-cyclo. The NMA regimen was mel-fludarabine. Graft versus host disease (GVHD) prophylaxis consisted in cyclosporine + methotrexate (83%) or mycophenolate (17%), with the addition of thymoglobulin for unrelated or related female vs male recipient donors. The median age was 42 yrs (IQR 38-47), 60% pts were male. The hematopoietic cell donors were sibling in 77 pts and unrelated in 25, while the source of stem cells was peripheral blood in 65 and bone marrow in 37 pts. Fifty-six pts received allo-SCT as first-line therapy, 31 as second-line and 13 as third or fourth-line. Median time from diagnosis to allo-SCT was 22 months. Response status at the time of transplant was at least PR in 63% of the pts, including ≥ VGPR in 36% and CR in 12%.

Overall, the response rate after allo-SCT was as follows: CR 58%, VGPR 19%, PR 18%. Median CR duration was 10 years (44% at 15 years).

The incidence of acute grade II-IV and III-IV GVHD was 25% and 15%, respectively. The incidence of all grades chronic cGVHD was 48%, grade ≥2: 24%, with a median onset time of 178 days. By competitive risks analysis, the cumulative incidence of cGVHD at 1 and 3 years was 20% and 32%, respectively. On univariate analysis, the gender combination of female donor-male patient resulted in significantly higher incidence of cGVHD (sub hazard ratio, SHR, 2.3, P= 0.03). The cumulative incidence of NRM was 6.6% at 100 days, 11% at 1 year and 14.4% at 3 years, with lack of statistically significant relationship with the conditioning regimen. By univariate analysis, < PR prior to allo-SCT (SHR 2.8) and all grades cGVHD (SHR 4.6) were significantly associated with an increased NRM.

By competitive risks analysis, the cumulative incidence of relapse was 50% at 5 years, 58% at 10 years and 59% at 15 years.  On univariate analysis, sibling donor (SHR 0.54, P= 0.047), all grades cGVHD (SHR 0.5, P= 0.011) and ≥VGPR after allo-SCT (SHR 0.38, P= 0.001) were significantly associated with extended time to progression. All grades cGVHD (SHR 0.43, P= 0.005) and ≥VGPR after allo-SCT (SHR 0.36, P= 0.001) were independent predictors for a lower risk of progression on multivariate analysis.

With a median follow-up of 13 years, overall survival (OS) was 43% at 5 years and 34% at 10 years. In univariate analysis, low ISS stage at diagnosis, sibling donor, absence of female donor-male patient gender combination, ≥VGPR prior to allo-SCT and first-line allo-SCT were significantly related to OS. Multivariate analysis confirmed an OS benefit for having a sibling donor (HR 0.30, P< 0.001), absence of female donor-male patient gender combination (HR 0.51, P= 0.022) and being in ≥ VGPR prior to allo-SCT (HR 0.41, P= 0.003).

In conclusion, this retrospective analysis performed with an extended follow-up of 13 years shows that a fraction of MM pts can be long-term survivors after allo-SCT, one third of them being potentially cured. The best outcomes were obtained when allo-SCT was applied in pts in an early phase of their disease and with a small residual tumor size. The major challenge, both with MA and NMA allo-SCT, was the relatively high post-transplant relapse rate. The presence of cGVHD was protective for the risk of progression, supporting the role of GVM, but increased the risk of NRM. New approaches aimed at modulating cGVHD are warranted. In addition,  incorporation of novel agents before and after allo-SCT to increase the rate and duration of high-quality responses, as well as identification of those patients mostly benefiting from this procedure, will likely contribute to improve long-term outcomes.