Allogeneic SCT in Multiple Myeloma: Limited Predictive Value of Chimerism Analysis for Heralding Relapse/Progression Due to Extramedullary Disease

Background Allogeneic stem cell transplantion (allo SCT) for multiple myeloma (MM) takes advantage of a tumor cell-free graft along with potential immunosurveillance of residual malignant plasma cells. However, multiple factors such as treatment-related mortality (TRM) and morbidity and post-allo disease progression question the more widespread use of allo SCT. The induction of an early complete donor chimerism (DC) is thought to be associated with a low incidence of relapse and improved survival. Interestingly, data on chimerism analysis are relatively scarce for allo SCT in MM. Therefore, we retrospectively evaluated longitudinal DC results by short tandem repeat (STR)-based techniques from a single lab. Data were correlated with a set of baseline and outcome variables. In addition, disease- and treatment related factors like extramedullary disease (EMD), acute and chronic graft-versus-host disease (GvHD) and post-transplant consolidation were analyzed for impact on outcome.

Patients and methods We identified 118 MM patients (pts) receiving non-myeloablative allo SCT at two German centers between 01/2006 and 12/2014. 90% of pts underwent allo SCT at relapse. Median follow-up was 42 months (mos), median overall survival (OS) 25 mos (range, 1-106) and TRM was 15%. Two-thirds of pts had received treosulfan/fludarabine as conditioning regimen. Median interval between MM diagnosis and allo SCT was 40.5 mos (range, 7-175) and median number of prior treatment lines was 3 (range, 1-10). 24 pts received allografts from matched related donors, 61 from matched-unrelated donors, and 33 pts from mismatched donors. Chimerism analysis was performed by STR-PCR on whole blood and T cell subsets using a validated multiplex STR-assay with a documented sensitivity of 1%. Response assessment was performed according the international myeloma working group guidelines.

Results 1398 chimerism samples (mean: 12/patient, 119 marrow, 1279 blood)) were analyzed. Median interval from allo SCT to first assessment was 15 (range, 3-62) days. Median duration of chimerism sampling was 24 mos. No graft rejections occurred. Full DC (> 99%) was achieved in 117/118 pts. Pts who lost full DC had a significantly shorter progression free survival (PFS) than those with sustained DC (HR 0.2; p<.0001).  OS, however, was not different. In patients who developed disease progression, loss of DC versus sustained DC was not predictive for outcome. In spite of relapse, a full DC was found in peripheral blood in 82% and in 45% in bone marrow samples of pts, respectively. With respect to clinical predictors of outcome, the occurrence of acute GvHD had no impact on PFS nor OS. In contrast, induction of chronic GvHD was associated with prolonged PFS (HR 0.50; p=.0088) and OS (HR 0.39; p<.001). Despite the induction of full DC in the blood, 13/118 (11%) of pts not in complete response received preemptive donor lymphocyte infusions (DLI). DLI without any concomitant signs of graft rejection or disease progression was associated with favourable OS (HR 0.43; p=.0067). When analysing relapse patterns after allo SCT, a 35% rate of EMD was observed. Interestingly, OS was not different when compared to intramedullary relapses. However, when allo SCT was performed as part of a consolidation therapy for prior EMD or plasma cell leukemia, OS was significantly shorter than compared to patients without EMD history (median OS 6 vs. 41 mos, p=.0035) Nevertheless, 25% of these EMD patients showed long-lasting remissions beyond 4 years. High risk cytogenetics (del17p, t(4;14) and +1q21), CMV reactivation or use of ATG had no impact on outcome.

Conclusions To the best of our knowledge, this is the largest study of serial DC analysis in a cohort of clinically well characterized MM pts. The robust engraftment in almost all patients as well as the lack of graft rejections indicate the severe immunosuppression in MM due to pre-treatment and underlying disease. On the other hand, the most important predictors for PFS and OS was the occurrence of cGvHD. Preemptive DLI may further improve tumour control. The frequent spread of myeloma to extramedullary sites can be interpreted as an escape to more immuno-privileged regions as compared to the bone marrow. This pattern clearly limits the potential of prolonged serial DC analysis as the only tool for relapse detection. Our observations underscore the need to prospectively evaluate complementary strategies in this setting.