Long-Term Outcome of Lower-Risk MDS Patients after Immunosuppressive Therapy (IST) with Anti-Thymocyte Globulin (ATG)+/- Cyclosporin A (CsA)

Introduction:Long-term outcome of patients with IPSS low or int 1 (LR) MDS treated by IST with horse or rabbit (hATG or rATG) +/-CsA is not well described and patient selection criteria for such a therapy remain ill-defined in daily practice, including after resistance to erythropoiesis stimulating agents (ESA).

Methods: We report the outcome of 15 consecutive LR-MDS patients treated in our institution between 2005 and 2013, with h or rATG (15 mg/kg/d for 5 days and 3.5 mg/kg/d for 5 days, respectively), +/- CsA 2 mg/kg BID for 6 months (m). The use of h or rATG formulation was based on institutional practice and availability. Patients with altered renal function or vascular contra-indication to CsA (mostly due to age) only received ATG. Bone marrow (BM) blasts were counted over 500 cells in cases of very low BM cellularity and over 1,000 cells otherwise. FISH analysis of chromosomes 5 and 7 was performed in case of cytogenetic failure. IST response probability score (ISTRPS) was calculated according to Saunthararajah, Blood 2003. Overall response rate (ORR), including complete response (CR), assessed at 12 m, and progressive disease (PD) were defined according to IWG 2006 criteria. Transfusion independence (TI) was defined as independence from RBC transfusion for at least 8 weeks.

Results: Median age was 60 years (y) (range 46-77). Median follow-up was 8.1 y. All patients were RBC-transfusion dependent (RCTD). Median RCTD duration was 4.9 m (0-42). Platelet counts were <50 10⁹/L and <20 10⁹/L in 85% and 50% of patients, respectively. Median ANC was 0.95 10⁹/L (0.18-4.5). WHO classification was RA, RCMD, RAEB-1 and non assessable due to very low BM cellularity in 5, 6, 2 and 2 patients. Median BM blast percentage was 1% (0-8). BM cellularity was reduced in 9/14 (64%) evaluable patients. Cytogenetics according to IPSS-R was very good, good and intermediate, in 2 (13%), 10 (67%) and 1 (7%) respectively and failed in 2 (13%). IPSS-R, assessable in 14 patients, was low, int and poor in 7, 6, 1. 7/14 patients had a DRB15 allele. ISTRPS was high in 4/15 patients. A PNH clone was detected by flow in 4 patients. A large granular lymphocyte population, a TCR clonal rearrangement and clinical or serological features of an associated immune disorder were present in 7/11 (63%), 4/10 (40%) and 5/14 (36%) patients, respectively. 6/14 patients had received at least one prior therapy for MDS including ESA in 5 and CsA, androgens, steroids, thalidomide, lenalidomide in 2, 3, 3, 1, 1 patients, respectively. 10 patients received hATG (66%), 5 (34%) rATG, combined to CsA in 7/15 (46%). In the 14 patients evaluable for response (early death, n=1), ORR including TI was 7/14 (50%) (4 CR and 3 HI; 6 TI). Median time to TI was 5.5 m (1-8). Median TI duration was 6.2 y. The only significant prognostic factors of TI were BM blast counts (60% if BM blasts ≤2% versus 0 if BM blasts >2%, P=0.04) and RCTD duration (62.5% if RCTD ≤6 m versus 0 if RTCD >6 m, P=0.02). There was a trend for a higher TI rate with lower platelet counts (P=0.13) and high ISTRPS (P=0.14). By contrast age, using a 60y cut-off, and presence of a DRB15 allele were not predictive of TI (P=0.58 and 0.31) in this cohort. Two relapsed patients were retreated with hATG +CsA and achieved a second durable response (1 CR (45m+) and 1 TI (9m+)). One non-responding patient, who subsequently developed full-blown PNH, responded to CsA and eculizumab. 5-y cumulative incidence of PD and AML were 42% and 6% respectively and median OS was 5.6 y. Median OS was not reached in responders versus 22 m in non-responders (P=0.004 by log-rank test). A single toxic death from a hemorrhagic stroke occurred very early in a patient heavily pretreated for MDS-associated vasculitis. PNH clone expansion and cytogenetic evolution occurred in 2 (both responders) and 3 patients (2 responders).

Conclusion: Lower-risk MDS patients responding to IST have a good long-term outcome. After relapse, durable responses to a second course of IST can be achieved. In addition to RCTD ≤6 m, BM blasts ≤2% emerged as an IPSS-R component strongly predicting response to IST. Age >60 y was not associated with worse response. Therefore, IST should be considered as a treatment option for selected lower-risk MDS patients with very low blast counts, irrespective of age.