Lunit will present an AI study on predicting EGFR mutations in non-small cell lung cancer (NSCLC) at the American Association for Cancer Research (AACR) Annual Meeting 2025, held from April 25 to 30 in Chicago, Illinois. The study, conducted with AstraZeneca, unveils the Lunit SCOPE Genotype Predictor, an AI model that predicts EGFR mutations from standard H&E-stained tissue samples.
The Lunit SCOPE Genotype Predictor stands out for its robust performance in diverse clinical settings. Its capabilities hinge on an extensive and varied dataset, which includes over 12,000 pathology slides sourced from patients across multiple countries, ensuring its real-world applicability. Predicting EGFR mutations is vital for identifying effective treatment pathways. Then, ensuring widespread and accurate testing remains challenging due to logistical and resource constraints within clinical environments.
The dataset comprises over 4,500 EGFR-mutated slides and more than 7,500 wild-type slides from different geographic regions, reinforcing the model’s versatility. Performance remained consistent across various clinical variables:
- Specimen types: The model can be used with different sample types.
- EGFR mutation subtypes: The model performs effectively across mutations.
- Slide scanners: Ability scales with multiple scanner usages.
- Scan magnifications: Does not depend on levels of zoom or close-up.
The tool shows potential as a cost-effective and rapid method to ensure timely diagnosis. Verbatim reported comments from Brandon Suh, CEO of Lunit, on study findings include how validating routine pathology slides for predicting driver mutations mitigates delays in targeted therapy.
Case Status just turned law firms into time-travelling mind readers
Lunit has established a strong global footprint over the past decade. Founded in 2013 in Seoul, South Korea, Lunit is a leading provider of AI-powered cancer diagnostics and therapeutic solutions. Their Lunit INSIGHT suite, FDA-cleared for cancer screening, is in use across over 4,800 medical institutions in 55 countries. Collaborative clinical studies from Lunit appear in journals including the Journal of Clinical Oncology and the Lancet Digital Health. Its research presentations extend to high-impact conferences like ASCO and RSNA. With a diverse and extensive international reach headquartered in South Korea, Lunit’s impact on cancer research and treatment is still growing.
The real kicker here is not just the promise of AI in cancer diagnostics but the hard-nosed execution: training on a mind-boggling 12,000 pathology slides from multiple countries, ensuring it’s ready for the messy realities of global medicine.
Lunit’s deployment of AI is a giant leap from the limited, controlled environments where diagnostic tools often falter; solidifying diverse datasets and consistent performance across different clinical settings.
