Multimodal Artificial Intelligence Fusion Models to Predict Response to Venetoclax and Azacitidine in Patients with Newly Diagnosed Acute Myeloid Leukemia
- 1Department of Internal Medicine I, University Hospital Carl Gustav Carus, TUD Dresden University of Technology, Dresden, Germany
Abstract
Venetoclax and azacitidine (ven-aza) is a standard of care therapy for newly diagnosed acute myeloid leukemia (AML) patients unfit to receive intensive therapy, and may also be appropriate for certain fit patients. Models predicting response to ven-aza currently focus predominantly on genetics. Here, we present two multimodal fusion models: a custom ResNet-34-based model and a pathology foundation model fine-tuned for bone marrow smear (BMS) images that combine clinical and genetic variables with whole slide images. Training on 193 newly diagnosed AML patients who received ven-aza, we obtained an area-under-the-receiver-operating-characteristic (AUROC) of 0.79 and 0.82 for the ResNet-based and fine-tuned foundation models in a heldout test set, and 0.76 and 0.73 in an external validation set, predicting best response to this therapy. Sensitivity analysis showed that fusing image data with clinical and genetic variables outperformed training on either modality alone. In univariable and multivariable analyses, image-based classification was an independent predictive marker, second only in effect size to IDH1/2 mutation status, outperforming other genetic alterations. Image-based explainability showed high model attention on myeloblasts, suggesting previously unrecognized cell-specific patterns associated with response. Our findings suggest BMS images complement genetic biomarkers to refine outcome prediction in ven-aza treatment and incentivize AI-supported outcome prediction, including bone marrow morphology.
Related
- Efficacy of Sorafenib Combined with Venetoclax and Azacitidine in Newly Diagnosed FLT3-ITD-Mutant Acute Myeloid Leukemia: Remission Efficacy and Prognostic Influence of Post-Remission Consolidation Therapy
- Effects of azacitidine combined with venetoclax on inflammatory markers and overall response rate in elderly patients with newly diagnosed acute myeloid leukemia
- Response to Venetoclax and Azacitidine in Induction-Refractory Adult Acute Myeloid Leukemia with UBTF Tandem Duplication and Associated Aberrant Splicing: A Case Report
- Stabilization of the PP2A-B56α complex overcomes venetoclax-azacitidine resistance by impairing OXPHOS in AML
- Patient-tailored azacitidine, venetoclax and gilteritinib triplet therapy appear to improve tolerability while preserving encouraging clinical outcomes in FLT3-mutated relapsed/refractory AML.
BibTeX
@misc{eckardt2026multimodal,
title = {{Multimodal Artificial Intelligence Fusion Models to Predict Response to Venetoclax and Azacitidine in Patients with Newly Diagnosed Acute Myeloid Leukemia}},
author = {J.-N. Eckardt and I. Srivastava and K. Gill and A. A. Osypuk and Z. Pan and M. Benouis and M. M. K. Schneider and F. Schulze and S. WInter and K. Sockel and M. Bornhaeuser and R. Shahswar and M. Heuser and J. M. Middeke and D. Pollyea},
howpublished = {medRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.12.26364938},
url = {https://www.medrxiv.org/content/10.64898/2026.10.12.26364938v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Multimodal Artificial Intelligence Fusion Models to Predict Response to Venetoclax and Azacitidine in Patients with Newly Diagnosed Acute Myeloid Leukemia
AU - J.-N. Eckardt
AU - I. Srivastava
AU - K. Gill
AU - A. A. Osypuk
AU - Z. Pan
AU - M. Benouis
AU - M. M. K. Schneider
AU - F. Schulze
AU - S. WInter
AU - K. Sockel
AU - M. Bornhaeuser
AU - R. Shahswar
AU - M. Heuser
AU - J. M. Middeke
AU - D. Pollyea
PY - 2026
DA - 2026/10/12
PB - medRxiv
DO - 10.64898/2026.10.12.26364938
UR - https://www.medrxiv.org/content/10.64898/2026.10.12.26364938v1
AB - Venetoclax and azacitidine (ven-aza) is a standard of care therapy for newly diagnosed acute myeloid leukemia (AML) patients unfit to receive intensive therapy, and may also be appropriate for certain fit patients. Models predicting response to ven-aza currently focus predominantly on genetics. Here, we present two multimodal fusion models: a custom ResNet-34-based model and a pathology foundation model fine-tuned for bone marrow smear (BMS) images that combine clinical and genetic variables with whole slide images. Training on 193 newly diagnosed AML patients who received ven-aza, we obtained an area-under-the-receiver-operating-characteristic (AUROC) of 0.79 and 0.82 for the ResNet-based and fine-tuned foundation models in a heldout test set, and 0.76 and 0.73 in an external validation set, predicting best response to this therapy. Sensitivity analysis showed that fusing image data with clinical and genetic variables outperformed training on either modality alone. In univariable and multivariable analyses, image-based classification was an independent predictive marker, second only in effect size to IDH1/2 mutation status, outperforming other genetic alterations. Image-based explainability showed high model attention on myeloblasts, suggesting previously unrecognized cell-specific patterns associated with response. Our findings suggest BMS images complement genetic biomarkers to refine outcome prediction in ven-aza treatment and incentivize AI-supported outcome prediction, including bone marrow morphology.
ER -