AI Pathology Platform Predicts Which Bile Duct Cancer Patients Will Respond to Targeted Therapy

A clinical study using artificial intelligence to analyze tumor biomarkers has demonstrated improved treatment response prediction in patients with HER2-positive biliary tract cancer, one of the most aggressive and difficult-to-treat malignancies. The research, published in Nature Medicine and led by Professor Lee Choong-kun of Yonsei University College of Medicine, employed Lunit's SCOPE AI platform to identify which patients would benefit most from adding targeted therapy to standard treatment regimens.

AI Analysis Reveals Treatment Response Patterns

The study enrolled 40 patients with advanced HER2-positive biliary tract cancer receiving first-line treatment combining the HER2-targeted drug Herceptin (trastuzumab) with standard chemotherapy and immunotherapy. Results showed a 55% objective response rate—meaning more than half of patients experienced significant tumor shrinkage—and a median progression-free survival of 10.6 months.

Lunit SCOPE analyzed whole slide images to assess HER2 expression patterns and characterize the tumor immune microenvironment. The AI platform revealed that patients with higher HER2 expression levels experienced greater tumor reduction. Perhaps more significantly, the analysis identified that immune-excluded patients—those whose tumors typically respond poorly to immunotherapy—showed longer progression-free survival when HER2-targeted therapy was added to their treatment regimen.

This finding addresses a critical clinical challenge in biliary tract cancer, where approximately 15% of patients are HER2-positive and typically face worse outcomes than their HER2-negative counterparts. Standard treatment for biliary tract cancer generally yields a life expectancy of around one year.

From Diagnostic Tool to Drug Development Partner

The Nature Medicine publication represents a significant validation milestone for Lunit (328130.KQ), positioning the company's SCOPE platform as more than a diagnostic tool. By demonstrating utility in clinical trial patient stratification and biomarker-driven treatment selection, Lunit is establishing itself as a potential partner for pharmaceutical companies pursuing precision medicine approaches in oncology.

The study was previously presented at the 2024 American Society of Clinical Oncology (ASCO) annual meeting in a rapid oral session, indicating the research community's recognition of its clinical significance. For patients with HER2-positive biliary tract cancer, the AI-guided approach may enable more personalized treatment selection based on both HER2 expression levels and immune microenvironment characteristics—potentially extending survival beyond historical norms in this challenging cancer type.

Clinical Implications for Precision Oncology

The research validates AI-driven biomarker analysis as a practical tool for treatment stratification in rare, aggressive cancers with limited therapeutic options. By identifying which patients are most likely to benefit from specific drug combinations, computational pathology platforms like SCOPE could help oncologists move beyond one-size-fits-all approaches toward truly personalized cancer care.

As pharmaceutical companies increasingly seek AI partners for biomarker-driven patient selection in clinical trials, publications in high-impact journals like Nature Medicine serve as critical proof points for the clinical utility of digital pathology artificial intelligence platforms.

Sources

1. Lunit AI Biomarker Study on Bile Duct Cancer Published in Nature Medicine — Seoul Economic Daily