Treatments and Outcomes for Patients at Academic Cancer Centers with Myelodysplastic Syndrome (MDS) By Revised International Prognostic Scoring System (IPSS-R) Scores

Abstract

BACKGROUND: IPSS-R classifications are used to estimate risk of disease progression to acute myeloid leukemia (AML), overall survival (OS), and to guide treatment decisions for patients with MDS. Recent data show mutational profiles improve prognostic stratification of MDS. This study aims to describe the distribution of IPSS-R scores in patients with MDS and assess treatment patterns and clinical outcomes by IPSS-R scores. METHODS: A retrospective cohort study assessed real-world, patient-level data from adults diagnosed with MDS between 2010-2020 at the Huntsman Cancer Institute and Moffitt Cancer Center, both NCCN, NCI-designated Comprehensive Cancer Centers. Data were obtained via chart review from electronic medical records. IPSS-R scores were calculated at or within 30 days of diagnosis. Patients with an intermediate to very-high IPSS-R score (3.5-6+) comprised the higher risk (HR) cohort; those with low or very low-risk IPSS-R (0-3) comprised the lower risk (LR) cohort. Study variables were compared across cohorts. Progression-free survival (PFS), defined as physician documented disease progression, AML transformation, or death, was evaluated from initiation of medication therapy and was censored at lost to follow-up or initiation of second-line therapy. OS was also assessed from treatment initiation. Clinical outcomes, specifically PFS and OS, were assessed within the hypomethylating agent (HMA)-treated subgroup. The influence of common genetic alterations on clinical outcomes was evaluated. RESULTS: There were 369 patients included in this analysis. Distribution of IPSS-R scores was: 5% (n=18) very low, 18% (n=68) low, 26% (n=97) intermediate, 25% (n=94) high, and 25% (n=92) very high. The mean age of patients was 66 years (SD: 10.6). Males were 56% (n=208) of patients, 93% (n=342) were white, and 41% (n=141) were transfusion independent at diagnosis. Transfusion independence was more common in the LR cohort (47%, n=35) compared to the HR cohort (39%, n=106). Among NGS-tested patients, the TP53 gene was most frequently altered (30%, n=86), followed by ASXL1 (20%, n=58), TET2 (17%, n=50), and DNMT3A (15%, n=44). The HR cohort harbored more TP53 mutations (HR: n=78 [35%]; LR: n=8 [12%]) and fewer SF3B1 mutations (HR: n=11 [5%]; LR: n=14 [22%]). In the HR cohort, first-line HMAs were used to treat 81% (n=230) of patients for a median of 4 cycles, while 11% (n=32) received no MDS-directed therapy, and 7% (n=21) received other medications for treatment of MDS (e.g., clinical trial drug, hydroxyurea). In the LR cohort, HMAs were used to treat 51% (n=44) of patients for a median 5 cycles, 26% (n=22) received no MDS-directed therapy, & 23% (n=20) received other medications for treatment of MDS (e.g., clinical trial drug, hydroxyurea). Second-line treatment was received by 24% of patients (n=88) with HMAs as the most frequent therapy choice for both HR (19%, n=53) and LR (16%, n=14) patients. Lenalidomide was the second most common second-line therapy for both HR (1%, n=3) and LR (5%, n=4) patients. Stem-cell transplant (SCT) was pursued within the HR cohort both more frequently (n=157 [55%] vs. n=23 [27%]) and sooner (7.2 vs. 9.3 median months from diagnosis) compared to the LR cohort. Among HR patients treated with first-line HMA complete response (CR) was achieved by 11% (n=26) with a median duration of CR of 23.8 months that included 24 patients (92%) who received SCT. Among first-line HMA-treated patients (n=274), median PFS was shorter within the HR cohort (median 9.5 months) compared to the LR cohort (18.3 months). Similarly, among first-line HMA-treated patients (n=274) median OS was shorter for the HR cohort (18.8 months) compared to the LR cohort (25.6 months. Among HMA-treated HR patients (n=227), OS was shorter for TP53 mutated patients (median 14.3 vs. 26.8 months) compared to TP53 wildtype patients, and longer for ASXL1 mutated patients (median 51.8 vs. 16.3 months) compared to ASXL1 wildtype patients. CONCLUSIONS: In the first-line setting, HR patients have low likelihood of achieving and maintaining a complete response. Both HR and LR patients have a high likelihood of experiencing disease progression. In addition, few patients received second-line therapy. Therefore, a large unmet need exists for first and subsequent lines of therapy. Specific genetic profiles may influence treatment decisions. Future research will describe the IPSS-M classification system within the same cohort.