Adverse Cytogenetic Abnormalities At Diagnosis Are The Most Important Predictor Of Relapse Post T Cell Depleted (TCD) Allogeneic Hematopoietic Stem Cell Transplantation In Patients With Advanced Myelodysplastic Syndrome (MDS)

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is the only curative treatment for advanced MDS. High relapse rate of up to 40% is a major obstacle limiting its success. TCD allo-HSCT has the advantage of reducing graft vs. host disease (GvHD), however, with concerns that depletion of donor T cells could lead to loss of graft vs. leukemia (GVL) effect. The aim of this study was to characterize the incidence and parameters predicting relapse in pts with advanced MDS treated with TCD allo-HSCT.

One hundred and eight adult pts with advanced MDS (RAEB I and higher at diagnosis or at progression) underwent TCD transplants at MSKCC between 1/2001-4/2012. This included 52 females and 56 males with a median age of 57.6 years (18-73). Etiologies included: de-novo 85, therapy-related 16 and post aplastic anemia or MPD 7. MDS subtypes at diagnosis by WHO criteria was: RA/RCMD 30, RAEB-I 34 & RAEB-II 44; and risk by IPSS-R in 103 pts (data missing in 5 pts) was: very low 7, low 15, intermediate 26, high 31 & very high 24. In 73 pts the disease progressed prior to transplant (44 AML and 1 developed myelofibrosis). One hundred and one pts were treated before transplant, 27 with hypomethylating agents, 73 with induction chemotherapy and one with syngeneic HSCT following high dose melphalan. Disease status at time of transplant by WHO criteria was: CR 40, RA/RCMD 49, RAEB-I 16 & RAEB-II 3; and using IPSS-R risk criteria: very low 7, low 37, intermediate 25, high 18 & very high 3. All pts underwent conditioning with a myeloablative regimen, 91 with busulphan, melphalan, and fludarabine; and 17 with a TBI-based regimen. ATG to prevent graft rejection was given to 104 pts. The source of stem cells was PB in 102 pts and BM in 6 pts. The BM grafts were depleted of T-cells by soybean agglutination method followed by sheep RBC rosetting, and the G-CSF mobilized PB stem cell using immunomagnetic CD34+ selection (Isolex initially and CliniMACS after 09/2011). Donors were HLA- matched (79; 39 related & 40 unrelated) or mismatched (29).

One hundred and six pts engrafted (2 died prior to engraftment). The OS (with 95% confidence interval) at 1 year was 70.2% (60.5%-77.9%) & at 3 years 50% (39%-59.2%). Cumulative incidence (CI) at 1 year of grade III-IV acute GvHD was 12.1% (6.8%-19.1%) & of chronic GvHD 2.8% (0.8%-7.4%). The CI of NRM at 1 year was 23.3% (15.8%-31.7%) & at 2 years 31.8 % (23.2%-40.7%). CI of relapse at 1 -year was 11.2 % (6.1%-18%) & at 2 years 16% (9.7%-23.6%). Two relapses occurred prior to day 100 (day 47& 98); 60% within the first year, 25% in the second year and 15% later. Cytogenetics at diagnosis as defined by IPSS and IPSS-R criteria were strongly associated with risk for relapse. The CI of relapse for pts with poor risk cytogenetics as per IPSS at 1 year was 25.8% (11.9-42.2%) & at 2 years 32.3% (16.6%-49.1%). For pts with intermediate risk it was 9.1% (1.5%-25.6%) at 1 year & 18.2% (5.4%-36.9%) at 2 years and for pts with low risk 2% (<1%-9.4%) at 1 year & 4% (<1%-12.3%) at 2 years (p=0.008). The same pattern was seen when pts were analyzed according to IPSS-R (p=0.03). When pts with poor risk cytogenetics achieved cytogenetic remission by the time of transplant, the CI of relapse was lower as compared to those who did not; CI at 1 year for the remissions was 22.2% (6.5%-43.7%) & at 2 years 27.8% (9.5%-49.8%), versus 33.3% (9.3%-60.2%) at 1 year & 41.7% (13.6%-68.1%) at 2 years in the nonresponders. These differences were not statistically significant (p=0.6). Mixed BM chimerism (defined by <95% donor chimerism) and mixed T cell chimerism (defined by <50% donor chimerism) on day 100 and 6 months post transplant were not associated with relapse.

In this cohort of patients, where the majority of the patients received therapy before cytoreduction to induce remission or to decrease disease burden, relapse rates were not higher compared to reported rates for unmodified transplant. Poor risk cytogenetic abnormalities at diagnosis were the most important predictors of relapse. While clearance of cytogenetic abnormalities pre-transplant was associated with lower relapse rates compared to persistence of cytogenetic abnormalities, the risk remained high in both groups. Such patients would be ideal candidates for early post transplant intervention either with adoptive cell therapy, with NK or T cells exhibiting specific cytotoxic activity against leukemia antigens or other agents that can target residual leukemic stem cells or modify the recipient immune system.