Research demonstrates scalable AI frameworks for resolving incomplete clinical data and dissecting complex tumor biology
BostonGene, developer of an AI model for tumor and immune biology, today announced its participation in ESMO AI Digital Oncology Congress 2026 taking place November 16-18, 2026 in Berlin, Germany. The international forum unites oncologists, data scientists, and industry leaders to showcase advances in AI and computational biology across cancer care.
BostonGene will demonstrate how its multimodal AI platform addresses critical oncology challenges, including accurately identifying tumor origin in cancers of unknown primary from fragmented diagnostic profiles and deconvolving complex transcriptomic data at scale. These applications highlight the power of biology-first AI to support the characterization of patient populations and inform treatments that match their disease biology.
Details of BostonGene's presentations are below:
Oral presentation
Abstract number: 252
Title: Resolving tumor-of-origin from incomplete clinical data: a multimodal AI framework integrating histopathology, transcriptomics, and genomics
Date time: November 16 3:15 4:00 PM
Location: Alpha 1
Presenter: Aleksandr Sarachakov, PhD, BostonGene
BostonGene has clinically validated its proprietary multimodal AI framework to address tumor-of-origin classification bottlenecks in cancer of unknown primary (CUP). Conducted in collaboration with researchers at Weill Cornell Medicine, the study demonstrated that this tri-modal fusion framework-integrating histopathology, transcriptomics, and genomics-accurately predicted tumor origin even from partial diagnostic profiles. This approach rescues fragmented clinical data, accelerates CUP diagnostic workflows, and streamlines precision oncology trials by expanding patient eligibility and matching patients to targeted therapies more effectively.
Research done in collaboration with Weill Cornell Medicine
Poster presentation
Abstract number: 376
Title: Deconvolving bulk tumor transcriptomes: A Pan-Cancer Deep Learning Approach
Date time: November 17 4:30 5:30 PM
Presenter: Alexander Bagaev, PhD, BostonGene
BostonGene developed a proprietary pan-cancer deep learning platform that dissects bulk RNA-seq data across 20 tumor microenvironment cell types and 13 tumor diagnoses, yielding over 14,000 compartment-specific gene expression profiles. In collaboration with the Inova Schar Cancer Institute and Endeavor Health, the platform was evaluated to confirm its ability to distinguish gene expression originating from tumor cells versus the tumor microenvironment directly from bulk RNA-seq. This capability reduces target ambiguity and provides biopharma partners with a validated framework to de-risk target selection, refine biomarker-driven patient stratification, and accelerate precision oncology drug development pipelines.
Research done in collaboration with Inova Schar Cancer Institute and Endeavor Health
About BostonGene Corporation
BostonGene powers an AI model of tumor and immune biology to deliver disease-level insights and enable precision decision-making across oncology and immune-mediated diseases, spanning drug development and clinical care. By integrating multimodal data, including genomic, transcriptomic, immune, and clinical signals, BostonGene generates biologically grounded intelligence to optimize patient selection, trial design, and therapeutic strategy. This creates a scalable AI decision layer that improves development outcomes and clinical management. BostonGene partners with leading biopharmaceutical organizations and academic institutions to accelerate the delivery of precision therapies while continuously expanding its capabilities across new diseases and complex biological systems. For more information, visit www.BostonGene.com.
View source version on businesswire.com: https://www.businesswire.com/news/home/20261116629434/en/
Contacts:
BostonGene
Erin Keleher
+1-617-283-2285
Erin.Keleher@BostonGene.com