Validation of the international working group (IWG) 2023 criteria for clinical benefit in higher risk Myelodysplastic Syndromes (HR-MDS) using a large, international, randomized Phase 3 clinical trial dataset

Abstract Introduction: The IWG 2023 criteria for HR-MDS (Zeidan Blood 2023) emphasize clinically meaningful blood count recovery. While these criteria have been validated in several retrospective datasets, validation in a large prospective clinical trial has been lacking. Further, the FDA issued a draft guidance for MDS where durable complete remission (CR) and partial remission (PR) were acceptable endpoints for drug approval in HR-MDS. The hemoglobin (Hgb) threshold for CR was lowered to 10 g/dL in the IWG 2023 criteria from 11 g/dL in IWG 2006 criteria (Cheson Blood 2006). Here we sought to examine the prognostic impact of the IWG 2023 criteria, especially CR and composite (cCR), in the phase III PANTHER trial dataset which compared azacitidine (AZA) + pevonedistat vs. AZA alone in patients with HR-MDS, CMML, and oligoblastic AML (Ades L Blood Advances 2022). As only 4% of all patients in the trial received a subsequent allogeneic stem cell transplant (allo-SCT), it allowed examination of association of IWG 2023 with overall survival (OS) in a population largely treated with non-curative, but homogenous, therapy. Methods: We included patients from PANTHER with HR-MDS who received AZA alone. Two investigators separately assessed IWG 2023 responses using blood counts and bone marrow blast percentages. cCR was defined in IWG 2023 as CR + CR with bilineage count recovery (CRbi) + CR with unilineage count recovery (CRuni) + CR with limited count recovery (CRh) + CR equivalent. We compared cCR + PR versus no response (NR) + progressive disease (PD) using the best response achieved for any patient. We calculated overall survival (OS) from randomization to death or last follow up. We then conducted a landmark analysis starting at 6 months post randomization to assess the impact of best response on OS while avoiding immortal-time bias. Log rank tests were used to compare OS for different response categories. In parallel, ongoing work is integrating molecular genetic analyses with the IWG 2023 clinicopathologic criteria using ultrasensitive serial DNA sequencing at multiple baseline and on-treatment timepoints to identify molecular correlates of response durability and resistance. Results: Of 454 patients on the trial, 163 patients with HR-MDS were treated with AZA monotherapy, and sufficient primary data to adjudicate IWG 2023 responses were available in 113 (69%) of these patients. Median age was 74 years (range 41-92) and 37% were female. The IPSS-R at baseline was intermediate, high and very high in 25%, 39% and 36% of patients respectively. Mutations in TP53 and/or deletion of chromosome 17p were present in 25% of patients. Of this cohort, 7.4% of patients received a subsequent allo-SCT. Per IWG 2023 response criteria, 51% of patients achieved a cCR and best responses included CR (27%), CRbi (11%), CRuni (12%), CRh (1%), PR (0%), no response (42%), PD (7%). Of note, among the 31 patients who achieved CR per IWG 2023, only 2 patients had a Hgb of 10-10.9 g/dL and would not have qualified for a CR per IWG 2006. Median OS for the cohort was 18 (95% CI, 15 -21) months. Based on a landmark analysis at 6 months from randomization, patients who achieved a cCR+PR had an improved median OS of 15 (95% CI 13-not reached [NR]) months compared to patients who did not achieve cCR+PR (9.3, 95% CI 8.4-17 months, p=0.008). OS for patients with cCR+PR vs. no cCR+PR at 6 months was 89% vs. 78%, at 12 months was 65% vs. 40%, and at 24 months was 33% vs. 9.1%, respectively. Excluding patients who received a subsequent allo-SCT, patients who achieved a cCR+PR had an improved median OS of 16 (95% CI 14 - NR) months, compared to patients who did not achieve cCR+PR with 9.3 (95% CI 8.4-17) months, p=0.006. The OS for these patients with cCR+PR vs. no cCR+PR at 6 months was 92% vs. 77%; at 12 months was 68% vs. 41%; and at 24 months was 33% vs. 9.3%, respectively. Conclusions: We show that achieving cCR per IWG 2023 response criteria is associated with a significant improvement in OS in a large prospective clinical trial for patients treated homogenously with azacitidine monotherapy, with only 7.4% of patients receiving subsequent allo-SCT. Importantly, no patients had PR as their best response, which is in line with prior observations that PR is rare and often a transient response that converts to CR with continued therapy. Our data supports the inclusion of cCR as a measure of meaningful clinical benefit in clinical trials of HR-MDS.