IDH mutant MDS: Proposal for disease subset recognition based on molecular and clinical features, and the availability of targeted therapies

Abstract

Abstract Introduction IDH1 & IDH2 somatic mutations (MT) are described in in 5-10 % of myelodysplastic syndromes (MDS) cases. Recognition of molecular entities is crucial, particularly in the presence of active therapy for that subset. For example, del5q MDS accounts for <5% of MDS, but the impact of lenalidomide treatment emphasizes the need to identify those patients (pts). We explored the molecular and clinical phenotype, outcomes, and therapeutic implications of IDH MT MDS. Methods We identified pts with IDH1 & IDH2 MT and compared both to wildtype (WT) IDH cohort. We examined baseline characteristics. We reported overall survival (OS) from time of diagnosis, AML transformation, and explored response to therapy. Results Among 3513 molecularly annotated MDS pts, 107 pts (3%) harbored IDH1 MT, 154 pts (4.4%) had IDH2 MT, and 3 cases harbored both. Baseline characteristics comparing IDH1, IDH2, and WT groups were similar except 45.7% IDH2, 32% IDH1, and 23.4% WT were classified as MDS-IB (> 5% blasts) (IDH2 vs WT p<0.005, IDH1 vs WT p=0.067, IDH2 vs IDH1 p=0.04). Both IDH2 & IDH1 were higher risk by IPSS-M than WT (IDH1 57%, IDH2 73.9%, WT 44.4%) (IDH1 vs WT p=0.02, IDH2 vs WT p<0.005, IDH2 vs IDH1 p=0.001). Both IDH 1 & 2 MT had less complex karyotype than WT, 9% vs 3% vs 20% (p <0.005). Both IDH MThad significantly lower ANC compared to WT (p <0.001). Hgb was higher among IDH1 than WT (p=0.013). Autoimmune-related diseases (AIRD) were observed more in IDH1 compared to IDH2 and WT, 28%, 18.8%, and 19.5%, respectively (p=0.036). SRSF2 MT were more common along both IDH 1 & 2 compared to WT (33.6%, 43.5%, and 11.3%, respectively, p<0.001) and RUNX1 (24.3%, 18.2%, and 11.1%, respectively, p< 0.001). Both IDH 1 & 2 had fewer concomitant TP53 MT (17.8%, 7.1%, and 23.1%, respectively, p<0.001) and fewer TET2 (15.9%, 9.1%, and 23.6%, respectively, p<0.001). For IDH2 compared to WT, ASXL1 MT were increased (37% vs 21.5%) and SF3B1 decreased (9.1% vs 20.7%) (p<0.001). Reponses to azacitidine were higher among IDH1 & 2 MT pts compared to WT (ORR (CR+PR+HI) 49.1%, 48.1%, and 37.5%, respectively, and CR was 14.3%, 14% and 10.4%, respectively) (p=0.024 compared to WT and p=1.0 between IDH1 & 2). Among IDH1 MT pts, 16 pts received ivosidenib at any timepoint including AML. Three pts received treatment for LRMDS, 8 pts for HRMDS, and 5 pts after AML transformation. Six pts received ivosidenib as 1st line, 5 pts as 2nd line, 4 pts as 3rd line, and 1 pt as 4th line. Treatment was mostly given as monotherapy (13/16 pts). Response by IWG 2006 was CR 25%, mCR 6% and HI 37% (ORR 68%). By IWG 2022, response was 25% CR, 6% CRh, 12.5% HI, and 25% CReq (ORR 68.5%). The median OS was 41.2 mo for those who received ivosidenib compared to 30.4 mo for those who did not (p=0.41). Three pts received olutasidenib after ivosidenib, no response was observed. Among IDH2 MT pts, 42 received enasidenib at any timepoint including AML. Four pts received treatment for LRMDS, 14 pts for HRMDS, and 24 pts after AML transformation. Four pts received as 1st line, 17 pts as 2nd line, 19 pts as 3rd line, and 2 pts as 4th line. Treatment was mostly monotherapy (33/42 pts). Response (n=40 pts) by IWG 2006 was CR 10%, mCR 17.5%, and HI 25% (ORR 52.5%). By IWG 2022, response was CR 12.5%, CRL 7.5%, CRbi 5%, CRh 6%, HI 20%, and CReq 5% (ORR 56%). The median OS was 32.9 mo for those treated with enasidenib compared to 23.8 mo for those who did not (p=0.1). The rate of AML transformation was 53.9%, 45.8%, and 29.1% for IDH2, IDH1, and WT, respectively (p <0.005 compared to WT and p=0.2 between IDH1 and IDH2). IDH1 & IDH2 MT did not impact OS compared to WT, but IDH2 had worse OS compared to IDH1. The median OS was 30.9, 26.7, and 25.4 mo for IDH1, IDH2 and WT, respectively (IDH1 vs WT p=0.16, IDH2 vs WT p=0.2, IDH1 vs IDH2 p=0.04). In Cox regression analysis adjusting for IPSS-M, both IDH MT had better OS compared to WT (IDH1 HR 0.62 p<0.001, IDH2 HR 0.7 p=0.001). Conclusions IDH MT MDS have unique molecular and clinical features. Pts have higher blasts and higher risk IPSS-M but less complex cytogenetics compared to WT. SRSF2 and RUNX1 are common concomitant MT. Despite higher AML transformation, the OS is better than WT when adjusted for IPSS-M, either because of different biology or response to treatments. Treatment with IDH inhibitors during the disease course suggests survival benefit among those pts. Our findings suggest the recognition of IDH MT MDS as a unique disease subset.