Lunit Presents 12 AI-Powered Studies at ASCO 2025, Demonstrating Enhanced Cancer Treatment Selection
核心洞察
Lunit (搜索) will present 12 studies at ASCO 2025 showcasing AI-powered digital pathology solutions for precision oncology applications.
A key study with Japan's National Cancer Center demonstrated AI-enhanced HER2 (搜索) analysis in biliary tract cancer (搜索), identifying additional responders to trastuzumab deruxtecan with 50% objective response rate.
AI-powered PD-L1 (搜索) evaluation showed 70% concordance with expert pathologists and identified 231 additional immunotherapy candidates among 949 lung cancer patients.
Lunit (搜索) (KRX:328130.KQ), a leading AI-powered cancer diagnostics company, announced that 12 studies featuring its digital pathology solutions will be presented at the 2025 American Society of Clinical Oncology (ASCO) Annual Meeting in Chicago from May 30 to June 3. The research demonstrates significant advances in AI-assisted biomarker analysis for precision oncology applications.
HER2 Analysis Expands Treatment Options in Biliary Tract Cancer
A pivotal study conducted with Japan's National Cancer Center Hospital East (NCCE) evaluated HER2 (搜索) expression in biliary tract cancer (搜索) (BTC) patients using Lunit (搜索)'s AI-powered analyzer. The research analyzed a 288-patient screening cohort, with AI scores showing strong agreement with pathologist-assigned immunohistochemistry (IHC) scores.
Among 29 patients treated with trastuzumab deruxtecan (T-DXd), those with higher levels of HER2 (搜索)-intense tumor cells achieved a 50% objective response rate (ORR), along with significantly longer progression-free survival and overall survival. Notably, the study identified that AI-derived "membrane specificity" helped identify additional responders who also achieved a 50% ORR and improved survival outcomes.
The membrane specificity metric expanded the pool of potential T-DXd responders beyond traditional HER2 (搜索)-intense patients to include some classified as HER2-low, suggesting broader therapeutic applications for the antibody-drug conjugate in BTC.
AI-Powered PD-L1 Assessment Identifies Additional Immunotherapy Candidates
A separate prospective study with NCCE evaluated concordance between pathologist- and AI-assessed PD-L1 (搜索) expression in lung cancer patients from LC-SCRUM, one of Japan's largest nationwide observational cohorts. The analysis included 847 non-small cell lung cancer (搜索) (NSCLC) and 102 small cell lung cancer (搜索) (SCLC) patients using Lunit (搜索) SCOPE PD-L1.
The overall concordance between the AI model and three expert pathologists reached 70%, with particularly high agreement in key clinical subgroups: 84% for tumor proportion score (TPS) ≥50% and 94% for TPS 1-49%. Among 416 patients initially classified as TPS <1% by pathologists, the AI identified 231 with higher PD-L1 (搜索) expression, potentially expanding immunotherapy eligibility.
Novel CLDN18.2 Prediction Model for Gastric Cancer
A third highlighted study introduced an AI model to predict CLDN18.2 (搜索) expression in gastric cancer (搜索) using standard H&E slides. CLDN18.2 serves as a therapeutic target for zolbetuximab, but traditional IHC assessment faces limitations in tissue quantity, cost, and time requirements.
The AI model achieved area under the receiver operating characteristic curves (AUROCs) over 0.751 in external validation, demonstrating potential for efficient pre-screening of CLDN18.2 (搜索)-positive patients. The research also incorporated immune phenotype analysis, finding that CLDN18.2-negative patients with an "inflamed" phenotype—characterized by high tumor-infiltrating lymphocyte density—showed significantly better outcomes with immune checkpoint inhibitor plus chemotherapy compared to chemotherapy alone.
Comprehensive Research Portfolio
Beyond the three featured studies, Lunit (搜索) will present nine additional abstracts covering diverse research applications. These include AI-based subcellular profiling to assess drug-targetability of 74 membrane proteins across 34 cancer types, and deep learning analysis of endothelial cells to understand tumor vascular environment influences on immunotherapy response.
"Our ASCO 2025 presentations build on years of work to turn AI into a clinically dependable tool—not just for reading pathology images, but for improving how we select the right treatments," said Brandon Suh, CEO of Lunit (搜索). "From HER2 (搜索) scoring in biliary tract cancer (搜索) to PD-L1 (搜索) evaluation in lung cancer, our models are helping uncover treatment opportunities for patients who might otherwise be overlooked."
Clinical Impact and Implementation
The research demonstrates AI's potential to enhance precision oncology through improved biomarker assessment and patient stratification. The high concordance rates with expert pathologists support the reliability of AI models as clinical decision-support tools, while the identification of additional treatment candidates suggests meaningful clinical impact.
Lunit (搜索)'s AI solutions are currently deployed at over 4,800 medical institutions across more than 55 countries, with FDA-cleared Lunit INSIGHT suite supporting cancer screening globally. The company will exhibit at ASCO Booth #26149, where attendees can learn more about the featured studies and AI solutions.