Independent Multi-Vendor Study Validates BostonGene’s AI for Precision HER2 Scoring

Publication in Modern Pathology Highlights High Agreement Rates for BostonGene’s Foundation Model in Assessing Complex Breast Cancer

WALTHAM, Mass.--(BUSINESS WIRE)--BostonGene, the developer of the leading AI foundation model for tumor and immune biology, today announced another major independent validation of its AI and machine learning (ML) capabilities in a landmark blinded, multi-vendor HER2 benchmarking study. The results, published in the Modern Pathology article, “Agreement Across 10 Artificial Intelligence Models in Assessing Human Epidermal Growth Factor Receptor 2 (HER2) Expression in Breast Cancer Whole-Slide Images” were conducted in collaboration with Friends of Cancer Research and supported by leading global pharmaceutical companies and patient advocacy stakeholders.

This publication adds to a growing body of external evidence validating the technical rigor, performance, and real-world relevance of BostonGene’s AI-driven approach. By evaluating algorithmic and AI-based methods under an independent framework, the study mirrors the stringent benchmarks that drug developers and regulators increasingly use to assess advanced analytics technologies. Such multi-site validation is critical for increasing confidence in AI-driven biomarkers and de-risking regulatory, clinical, and companion diagnostic development. These blinded, ecosystem level evaluations confirm BostonGene’s position among a select group of companies operating at the highest threshold of scientific and technical excellence for clinical-grade AI.

BostonGene’s AI and machine learning capabilities are built on a foundation model of cancer and immune system that integrates multiomic, RNA, DNA, TCR, spatial, and clinical data at scale. This multidimensional approach enables deep characterization of tumors and the immune microenvironment, supporting critical decisions across the drug development lifecycle, from early-stage target discovery to research to patient stratification and trial optimization.

As HER2 remains one of the most critical biomarkers in oncology, the terminology and evaluation frameworks outlined in this publication are expected to set the benchmark for how AI and ML tools are assessed by drug developers going forward. BostonGene’s inclusion in this initiative reflects the industry’s sustained confidence in the company’s ability to deliver transparent, reproducible and clinically meaningful AI.

“This is not an isolated result,” said Nathan Fowler, MD, Chief Medical Officer at BostonGene. “We continue to see independent, external validation of the AI and ML algorithms that power our foundation model. These blinded, real-world evaluations provide the high-stakes certainty that drug developers trust when accelerating life-saving therapies.”

Pharmaceutical organizations supporting the study included AstraZeneca, Bristol Myers Squibb, Amgen, Merck, and GlaxoSmithKline. BostonGene continues to partner with these and other leading pharmaceutical companies on strategic programs where its AI-driven insights directly inform biomarker strategy and clinical execution.

About BostonGene Corporation

BostonGene is redefining cancer patient care and drug development through the integration of omnimodal data and artificial intelligence. Built and validated through an extensive real-world clinical testing network, BostonGene’s Foundation Model of cancer and the immune system integrates genomic, transcriptomic, and immune data with clinical outcomes to generate biologically grounded, actionable insights. These insights enable biopharma partners to design and de-risk trials, identify novel targets, and optimize therapeutic response prediction across all stages of development while simultaneously improving patient care through clinically integrated innovation. For more information, visit www.BostonGene.com.