Prognostic Value of HLA Evolutionary Divergence in AML Patients Treated with Venetoclax and Hypomethylating Drugs

BACKGROUND

The well-known polymorphism of class I human leukocyte antigen (HLA) molecules enables the presentation of a wide repertoire of peptides, including tumor-associated antigens. The HLA evolutionary divergence (HED) is a surrogate for immunopeptidome diversity and has been recently established as a novel prognostic factor for AML patients undergoing allogeneic HSCT (Pagliuca et al. Nature Communications 2023). Preclinical evidence supports the role of HED as a possible prognostic factor also in AML patients treated with venetoclax and hypometinating agents (VH), given their immunostimulating synergic action (Wang et al. PLoS One 2013; Lee et al Blood 2021). On these premises, we started a multicentric pilot study to investigate the possible role of HED as a predictor of outcome in AML patients treated with VH.

METHODS

From February 2021 to April 2024 we prospectively and retrospectively enrolled consecutive AML patients with de novo or secondary AML (sAML) treated with first-line VH. Exclusion criteria were refractory/relapsed AML, having received an AML therapy different from VH (excluding hydroxyurea), or unavailability of post-treatment disease assessment by bone marrow reevaluation. The primary endpoint was the complete remission (CR) rate, while overall survival (OS) was set as secondary endpoint. Here we present the data from the first 15 patients analyzed. HED was calculated according to published literature (Chowell et al. Nat Med 2019; Pierini et al. Mol Biol and Evol 2018) and values were expressed as median for locus A, B and C and mean for the class I taken together. Patients with greater than the median HED for locus A, B and C and greater than the mean for class I were compared against those with lower than the median HED for locus A, B and C and lower than the mean for class I.

RESULTS

Median age at diagnosis was 74 years (range 61-85), 10/15 patients (67%) were males. Eight patients (53%) had de novo AML and 7 (47%) sAML. According to ELN 2022 classification, one patient was classified as favourable risk, 10 as intermediate risk, and 4 as adverse risk. Patients received a median of 5 (range 1-30) cycles of VH. Median follow-up was 9.2 months (interquartile range: 5-25). Overall, one-year OS and PFS were respectively 67% and 58%. At the last follow-up, 7/15 patients (47%) were alive; disease-related death occurred in 4 out of 8 patients.

Sex and age were comparable between the groups; patients with lower HED had a trend toward higher platelet count at diagnosis (p 0.09). 2022 ELN risk was intermediate in 6/7 and low in 1/7 patients with high HED vs high in 4/7 high and intermediate in 3/7 patients with low HED; 6/7 patients whit high HED had a primary AML as compared to 2/7 in low HED.

Interestingly, while class I HED did not correlate with the CR rate, patients with a HED-B value higher than median showed superior CR rate (7/8 vs 3/7), though not statistically significant due to the low samples size. Of note, class I HED and HED-B significantly impacted on OS, since patients with class I HED and HED-B above the medians showed longer median OS (24 vs 9 months for both factors, p=0.052 and p=0.02 respectively).

CONCLUSIONS

To our knowledge, this is the first study reporting the role of HLA polymorphism, throughout the analysis of HED, in the setting of VH-treated AML patients. According to our preliminary results, class I HED and HED-B may be prognostically significant in terms of OS. A larger number of patients will be analyzed in this ongoing study to verify these findings.

If confirmed, HED may be integrated with the existing predictive test at diagnosis to orient therapeutic decision for elderly and/or unfit AML patients in the expanding landscape of low intensity treatments.

Onida:Menarini-stemline: Speakers Bureau; kyowa: Speakers Bureau; MEDAC: Speakers Bureau; takeda: Speakers Bureau.