Clinical Predictors for Relapse Among Patients with AML Who Responded to Venetoclax-Based Treatment - a Real-World Prospective Analysis from the Revive Study Group

Background: Venetoclax (VEN)-based low intensity combinations are approved for the treatment of patients with acute myeloid leukemia (AML) ineligible for intensive chemotherapy. Despite the encouraging composite complete remission rates demonstrated in the VIALE-A trial, disease relapse remains a frequent event with dismal outcome. The aim of the current analysis is to characterize patterns and clinical predictors of relapse in a real-world setting. Methods: The REVIVE study is a prospective, real-life, observational study that assesses patterns of patient selection, efficacy and toxicity. 189 newly diagnosed AML patients from 12 medical centers in Israel were enrolled at initiation of VEN-based low-intensity therapy, between August 12, 2019, and April 15, 2022 (data cut). Demographic, clinical and patient-related baseline characteristics, as well as treatment patterns, safety and efficacy outcomes were collected. For the current analysis, remission was defined as either achieving complete remission (CR), CR with incomplete count recovery (CRi) or morphological leukemia-free state (MLFS). Relapse-free survival (RFS) was calculated from the date of first response to the date of progression of disease or death from any cause, whichever occurs first or lost follow-up in the population of responding patients (CR, CRi, MLFS). A cox proportional hazards model was used to compare relapse-free survival between subgroups of baseline characteristics. Results: Of the 189 patients recruited to the study, 110 (58.2%) achieved a response (CR- 55.5%, CRi- 39.1%, MLFS- 5.5%). The median age of these patients was 75 years, 58.2% were males and 69.1% had ECOG performance status of 0-1. The median time to remission was 35 (0-193) days. As of data cut, with a median follow-up of 278 (31-878) days, 50 patients (45% of responding patients) who achieved response have relapsed. Median RFS was 425 (275-588) days. 1-year after achieving remission 53.3% (42.3%-64.2%) of the patients remained in remission. The estimated OS of the patients who responded was 509 (302-680) days. In univariate analysis, there was a statistically significant association of age (p=0.005) and baseline neutrophil counts (p=0.001) with RFS. The expected hazard of relapse in patients <75 of age was 0.4 times the hazard in patients ≥75 (median RFS of 588 vs. 275 days, respectively). The expected hazard of relapse in patients with baseline neutrophils ≥1,000 cells/µl was 2.7 times the expected hazard in patients <1,000 cells/µl. Patients with a normal BMI (<25) had a hazard of relapse that was higher than those with an overweight BMI (≥25) (p=0.06). There was no statistical association of gender, ECOG performance status, type of AML, cytogenetic risk group, and prior HMA exposure with RFS. On a multivariate analysis, only age and baseline neutrophil counts remained significant in the cox proportional hazards model. A comparison of responding patients with prior myeloproliferative neoplasms (MPNs) (n=14), patients who were excluded from registrational clinical studies, to patients with therapy-related AML/ prior MDS (n=40) demonstrated a significant association with RFS. The median RFS was 154 days (range 76-not reached) in the MPN group compared to not reached in the other group (range 298-not reached, KM estimation for 1 year of 25% vs. 63.8%, respectively p=0.0119). Conclusions: In a real-world, prospective analysis, newly diagnosed AML patients who responded to treatment demonstrated a median relapse free survival of more than 1 year (425 days). Clinical factors significantly associated with shorter RFS were age≥75 and baseline neutrophil counts (≥1,000 cells/µl). AML patients with prior MPN have a shorter RFS compared to other secondary AML patients. Analysis of patterns of post-relapse salvage approaches and survival is planned for future analysis. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal