Prospective clinical validation of MAC-scoring as an affordable, real-time biomarker for HMA/venetoclax response in AML

Abstract The treatment of acute myeloid leukemia (AML) has been transformed by incorporating hypomethylating agents (HMA) like 5-Azacytidine and the BCL-2 inhibitor Venetoclax (VEN) as the first line regimen for most unfit AML patients. With response rates around 66%, remission rates in this patient group are similar to the ones observed with standard induction therapy regimens. Importantly, HMA/VEN has also been successfully used to salvage patient refractory to induction regimens before allogeneic stem cell transplantation, and is currently investigated in first-line trials for younger patients, raising the question which patient should be induced with which regimen. Recently, multiple biomarkers of HMA/VEN response based on mutational profiles been proposed. However, genomic profiling remains costly and can take 10 days or longer to be completed. Therefore, these scores might be useful for estimating response duration but not to guide treatment decision. We have recently described the “Mediators of Apoptosis combinatorial” (“MAC”)-Score, a flowcytometry-based tool to assess the BCL-2 family protein ratio: BCL-2/ BCL-xL+MCL-1 in GPR56+ leukemic stem cells (Waclawiczek et al, 2023). We extended our initial retrospective cohort from 60 to now 114 patients and validated that above median MAC-Scores predicted remission (CR, CRi, MLFS) with a positive predictive value of >95% leading to a 3.5-times increase in event-free survival (EFS). When comparing MAC-Scores of this cohort with the recently proposed 4-gene score, patients in the “high benefit” group displayed higher MAC-Scores then those in the “low” or “intermediate benefit” groups, correlating with the respective EFS of 86 (low benefit), 104 (intermediate benefit) and 253 days (high benefit). Importantly, within the ELN2024 “higher” and “intermediate” risk groups, MAC-Scores below the median of 0,4 clearly identified patient subgroups with dismal outcome similar to TP53 mutant patients, some of them with NPM1-mutations classically linked to good prognosis. These results highlight the clear need for a biomarker that can guide treatment decisions in real-time. Based on these retrospective results, we have established MAC-Scoring within the diagnostic flowcytometry facility at Heidelberg University hospital and have routinely performed MAC-Scoring in diagnostic AML patients from Heidelberg and Mannheim University hospitals since November 2023. The median MAC-Score of the 85 prospectively assessed fresh, unfrozen patient samples was again 0,4. MAC-Scores were stable for up to 72 hours and comparable between leukemic blasts harvested from the bone marrow or peripheral blood. Of the 85 patients, 36 received HMA/VEN based first-line treatment with 19 displaying a below and 17 an above median MAC-Score. So far, no patient with a MAC-Score >0.4 was refractory to induction treatment. Median EFS was 106 days (<0,4) and not reached (>0,4) with only one event recorded during a median follow-up of 136 days. Similar results were obtained after including 10 additional patients receiving HMA/VEN as a salvage regimen. Low MAC-Scores correlated with poor risk genetic features like complex karyotypes or TP53 mutations, and high MAC-Scores with IDH1/2 or Core-binding factor mutations. However, MAC-Scoring allowed the identification of individual patients defying genetic predictors. At relapse, MAC-Scores were generally lower with clear shifts in BCL2 family protein levels. In summary, MAC-Scoring provides an affordable, quickly accessible real-time tool to risk stratify patients for individualizing treatment decisions. This might become even more relevant when HMA/VEN moves into the first line for younger, fit patients. MAC-Scoring can longitudinally assess changes in BCL-2 family proteins, which may guide future therapies with novel drugs targeting other BCL-2 family proteins emerging.