Presentations highlight the clinical utility of AI-powered models and multimodal profiling for biomarker discovery, treatment response prediction, and patient stratification across diverse solid tumors and hematologic malignancies

WALTHAM, Mass.: BostonGene, developer of the leading AI model for tumor and immune biology, today announced that nine abstracts have been accepted at the 2026 ASCO Annual Meeting (ASCO), scheduled to take place May 30 – June 2, 2026, at McCormick Place Convention Center in Chicago, IL. BostonGene will also exhibit at booth #33137.

The accepted abstracts underscore the broad clinical utility of BostonGene’s AI-powered models and advanced analytical platforms, including the Tumor Portrait® test and Kassandra™ cell deconvolution. Spanning both tissue and peripheral blood analysis, the featured research showcases how integrating deep genomic, transcriptomic, and immunomic profiling can identify novel actionable biomarkers, predict immunotherapy response and toxicity, and map complex resistance mechanisms. Developed in collaboration with leading institutions such as the UT MD Anderson Cancer Center, Weill Cornell Medicine, and the Parker Institute for Cancer Immunotherapy, these studies highlight BostonGene's pivotal role in accelerating biomarker discovery and optimizing patient stratification for clinical trials.

Details about the abstracts selected for presentation can be found below:

Poster presentations
Poster number: 273
Title: TROP2 as an actionable biomarker for anal cancer
Date & time: May 30 | 9:00 AM – 12:00 PM
Speaker: Mir Lim, MD, UT MD Anderson

BostonGene and MD Anderson Cancer Center collaborated to perform comprehensive genomic profiling of localized and metastatic anal cancers, assessing TACSTD2 (TROP2) expression as a potential therapeutic biomarker. The analysis found high TACSTD2 expression that correlated with TROP2 immunohistochemistry, supporting its potential use in identifying patients who may benefit from anti-TROP2 antibody-drug conjugates. The study underscores the clinical utility of BostonGene’s Tumor Portrait® test in biomarker discovery and validation.

Research conducted in collaboration with UT MD Anderson

Poster number: 257
Title: Developing harmonized tumor microenvironment subtypes for patient stratification in clinical trials
Date & time: May 30 | 1:30 PM – 4:30 PM
Speaker: Sofya Kust, PhD, BostonGene

BostonGene developed an AI-driven harmonized tumor microenvironment (HTME) classification system to improve patient stratification for immunotherapy and targeted therapies across solid tumors. By applying machine learning approaches-including agglomerative clustering, K-nearest neighbors classification, and transcriptomic pathway analysis-to RNA-seq data from more than 40,000 cancer samples, the study identified nine reproducible TME subtypes capturing key biological features such as immune activity, fibrosis, vascularization, and hypoxia. The AI-based framework outperformed existing methods by integrating multiple biological dimensions simultaneously and accurately predicting treatment response and progression-free survival across different cancer types and therapies.

Poster number: 323
Title: Multimodal immunoprofiling of peripheral blood using foundation models of the immune system for predicting immunotherapy response and toxicity in the RADIOHEAD pan-cancer cohort
Date & time: May 30 | 1:30 PM – 4:30 PM
Speaker: Evgeny Barykin, PhD, BostonGene

In collaboration with Parker Institute for Cancer Immunotherapy, BostonGene applied AI-based models of the immune system to analyze blood samples from more than 1,000 cancer patients receiving immune checkpoint inhibitor (ICI) therapy in the RADIOHEAD pan-cancer cohort. By combining RNA sequencing, flow cytometry, and machine learning models trained on 45,000 patient profiles, the study identified immune signatures linked to both treatment response and severe immune-related toxicities. The AI-driven embeddings and multimodal immunoprofiling approach successfully stratified patients into responder, non-responder, and high-risk toxicity groups, highlighting the potential of AI-powered peripheral blood profiling to improve immunotherapy patient selection.

Research conducted in collaboration with Parker Institute for Cancer Immunotherapy

Poster number: 360
Title: Effects of a multimodal TCR/BCR repertoire foundation model on blood RNA-seq–based prediction of severe adverse event risk and rheumatoid arthritis
Date & time: May 30 | 1:30 PM – 4:30 PM
Speaker: Alexander Bagaev, PhD, BostonGene

BostonGene utilized its immune model leveraging multimodal TCR/BCR repertoire data to address two critical clinical challenges: predicting severe immune-related adverse events (irAEs) in cancer patients receiving immune checkpoint inhibitors and detecting rheumatoid arthritis from peripheral blood. By integrating adaptive immune receptor repertoire embeddings with gene expression signatures, the model significantly outperformed traditional immune repertoire analysis approaches-such as clonality, diversity, and V(D)J gene usage metrics-in both predictive accuracy and cross-cohort robustness. These advances enabled earlier risk stratification and more precise immunotherapy management than conventional biomarker methods.

Poster number: 533
Title: Immune priming with the EZH2 inhibitor tazemetostat in B-cell lymphomas receiving CART cell therapy
Date & time: June 1 | 9:00 AM – 12:00 PM
Speaker: Samuel Yamshon, MD, Weill Cornell Medicine

BostonGene applied its multimodal molecular profiling platform to characterize the systemic immune remodeling driven by tazemetostat priming in patients with B-cell lymphomas receiving CART therapy. By integrating high-parameter flow cytometry, RNA sequencing, immune cell deconvolution, and functional gene signature analysis, BostonGene’s analysis revealed systemic immune remodeling associated with enhanced antigen presentation, T-cell activation, and reduced immunosuppressive signaling-supporting the use of EZH2 inhibition for improved CART efficacy and durability.

Research conducted in collaboration with Weill Cornell Medicine

Poster number: 268b
Title: IvoLoC: A phase II trial of ivonescimab (IVO) in endocrine-refractory hormone receptor (HR)-positive or triple-negative (TN) metastatic invasive lobular carcinoma (mILC)
Date & time: June 1 | 1:30 PM – 4:30 PM
Speaker: Jason Mouabbi, MD, UT MD Anderson

In a trial in progress with UT MD Anderson Cancer Center, BostonGene will perform multimodal longitudinal blood and tissue analysis for mILC patients receiving the bispecific antibody targeting PD-1 and VEGF, ivonescimab. Moving beyond traditional efficacy metrics, complex correlates linking clinical outcomes with molecular features are ongoing. By integrating real-time ctDNA dynamics with predictive responder scores, BostonGene aims to identify critical markers of response and resistance needed to optimize treatment durability and refine future trial strategies.

Research conducted in collaboration with UT MD Anderson

Online only
Abstract number: e12594
Title: Examining the tumor microbiology and microenvironment of patients with triple negative inflammatory breast cancer receiving Keynote 522

Stage III triple‑negative inflammatory breast cancer patients receiving KN522 therapy were analyzed with BostonGene’s integrative multimodal platform. By combining intrinsic subtype classification with immune deconvolution, BostonGene successfully mapped the molecular landscape of treatment response. Non-responders were marked by angiogenic and endothelial remodeling and potential biomarkers of brain metastasis risk, demonstrating the platform’s unique capability to predict both therapeutic outcomes and disease progression.

Research conducted in collaboration with UT MD Anderson

Abstract number: e13080
Title: Real-world clinical utility of comprehensive genomic and transcriptomic profiling in metastatic breast cancer (mBC)
Description: Clinicians at MD Anderson Cancer Center utilized BostonGene’s Tumor Portrait® test to validate the impact of comprehensive genomic profiling in guiding treatment decisions and predicting outcomes in a heterogeneous real-world mBC cohort. The findings demonstrated that BostonGene’s integrated approach uncovered clinically actionable insights and predictive signatures that directly guided treatment selection and improved patient outcomes.

Research conducted in collaboration with UT MD Anderson

Abstract number: e16004
Title: Association of angiogenic, fibrotic, and immunosuppressive tumor microenvironment with immunotherapy outcomes in metabolic disease–associated biliary tract cancers

Researchers at MD Anderson Cancer Center leveraged BostonGene’s transcriptomic classifiers and KassandraTM cell deconvolution to characterize the tumor microenvironment (TME) of metabolic disease-associated biliary tract cancers. The analysis uncovered a distinct TME in metabolic disease-associated cholangiocarcinoma that likely impairs response to immune checkpoint inhibitors, demonstrating the value of BostonGene’s AI-powered solutions for therapeutic response prediction and trial design.

Research conducted in collaboration with UT MD Anderson

For more information, please visit the 2026 ASCO Annual Meeting website. The abstracts will be published online in the Journal of Clinical Oncology supplement for the ASCO Annual Meeting Proceedings.

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.

Fonte: Business Wire