Response to Treatment Among SF3B1 Mutated Myelodysplastic Syndromes (MDS): A Case-Control Study from the MDS Clinical Research Consortium (MDS CRC)

Introduction

Somatic mutations in SF3B1,a gene encoding a core component of RNA splicing machinery, have been identified in patients (pts) with myelodysplastic syndrome (MDS). The SF3B1 mutation (MT) is more commonly detected in pts with ring sideroblasts (RS) morphology and is associated with favorable outcome. The pattern of response among SF3B1 mutated MDS pts to available treatment options, including erythropoiesis stimulating agents (ESA), hypomethylating agents (HMA) and  lenalidomide is not known. The distinct underlying disease biology among such pts may alter response to treatment.

Methods

Pts treated at MDS CRC institutions with MT vs wild-type SF3B1 (WT) controls were matched 1:2.  Matching criteria were age at diagnosis, year of diagnosis and International Prognostic Scoring System (IPSS) category at diagnosis.  IPSS category was split into two groups (Low or Int-1 vs. Int-2 or High).  Matching was performed using the R package by calculating a propensity score, which was then used to determine the two most similar WT SF3B1 patients for each SF3B1-mutated pt, without replacement.  Additionally, to be included in the population, pts also had to have been treated with one of the following: ESAs, HMA, or lenalidomide. Response to treatment was evaluated by international Working Group criteria (IWG 2006) and classified as response if hematological improvement or better was achieved (HI+). Survival was calculated from date of treatment until date of death or last known follow-up, unless otherwise noted.

Results:

We identified 48 Pts with MT and 96 matched controls. Table 1 summarizes baseline characteristics comparing MT vs WT SF3B1 cohorts. SF3B1 MT was detected more often in association with RS, as expected.  The majority of pts had lower-risk disease by IPSS and revised IPSS (IPSS-R). Pts with MT had higher platelets than controls. The most common concomitant somatic mutations observed were TET2 (30%), DNMT3A (21%), and ASXL1 (7%). Median follow-up time from diagnosis was 35 months (mo). Median overall survival (OS) from diagnosis was significantly longer for patients with SF3B1 MT (108.5 mo (68.8, NA)) than wild-type controls (28.3 mo (22.3, 36.4); p < 0.001). Patients with an SF3B1 MT had a decreased hazard of death (hazard ratio [HR]: 0.49 (95% confidence limits [95% CL]: 0.29, 0.84); p = 0.009)

ESA was the first line therapy for 43 pts (88%) with MT and 55 WT Pts (56%). For ESA treated pts, 14 out 40 MT Pts responded (35%) compared to 9/56 among WT Pts (16%), p 0.032. Among those treated with HMA therapy, 5 out 21 (24%) MT pts responded compared to 11/46 (24%) WT Pts (p 0.99). Finally, for Pts treated with lenalidomide 4/16 (25%) and 4/21 (19%) responded among SF3B1 MT and WT Pts respectively, p 0.7.

Conclusions

SF3B1 somatic mutation in MDS is commonly associated with RS, lower risk disease, and better OS. Pts with SF3B1 mutation had higher response to ESA compared WT SF3B1. No difference in response to HMA or lenalidomide was observed compared to WT patients. Response rates to lenalidomide and HMA were low in both MT patients and controls. Biologically rational therapies are needed that target this molecular disease subset.

Table-1: Baseline characteristics

		SF3B1 MT (n=48)	SF3B1 WT (n=96)	P value
Age	median	65	67	0.6
Gender	male	29 (60%)	64(67%)	0.5
Race	White	44/45 (98%)	83/90 (92%)	0.34
WHO classification	RA RARS RCMD RARS-T Del5 q RAEB-I RAEB-II MDS-U MDS/MPN CMML	3 24 8 4 1 3 3 2 0 0	6 9 17 2 6 10 9 3 11 9
IPSS	Low Int-1 Int-2 High	29 (60%) 16 (33%) 3 (6%) 0	21 (22%) 69 (72%) 4 (4%) 2 (2%)	< 0.001
IPSS-R	Very low Low Intermediate High Very High	15 (31%) 26 (54%) 5 (10%) 2 (4%) 0	11 (11%) 37 (39%) 26 (27%) 18 (19%) 4 (4%)	<0.001
Lab values (mean)	Hgb Platelets ANC myeloblasts	9.7 274 2.63 1	9.6 108 1.92 2	0.46 <0.001 0.04 0.05