[Song Young Doo, Edaily Reporter] Medical AI company Lunit announced on April 20 that it will present six research abstracts at the AACR 2026, held in San Diego, U.S. from April 17 to 22. The studies utilize its AI biomarker platform, Lunit SCOPE.

The most notable study to be presented is a collaborative research project conducted with Agilent Technologies and Ajou University College of Medicine. The study explores the relationship between c-MET (hepatocyte growth factor receptor) expression and the tumor microenvironment in non-small cell lung cancer(NSCLC).

c-MET is a protein known to promote tumor growth and metastasis, and has recently gained attention following approvals of antibody-drug conjugates(ADCs) targeting this pathway. However, the relationship between c-MET expression, tumor microenvironment, and response to immunotherapy remains insufficiently understood.

The research team analyzed 25,674 NSCLC samples using Lunit SCOPE IO and Lunit SCOPE uIHC to assess immune cell distribution surrounding tumor cells based on c-MET expression levels.

The results showed that tumors with high c-MET expression—particularly those with higher membrane expression relative to cytoplasmic expression—were associated with a significant decrease in tumor-infiltrating lymphocytes(TILs), which are immune cells responsible for attacking cancer cells.

These findings suggest a link between high c-MET expression and immune evasion. They further support the potential for improved treatment outcomes by combining MET-targeted therapies to modulate the immune-suppressive environment with immunotherapy.

In addition, Lunit will present findings from a Phase 2 clinical study involving HER2-positive metastatic colorectal cancer.

The study analyzed tumor samples from 30 patients treated with a combination of Tukysa(tucatinib) and Herceptin(trastuzumab), using AI to measure both the proportion of HER2-high tumor cells and the density of surrounding immune cells, and comparing these metrics with actual treatment outcomes.

The overall objective response rate(ORR) was 43.4%. However, response rates increased stepwise with higher proportions of HER2-high tumor cells. Notably, patients with ≥50% HER2-high tumor cells had an 83% lower risk of disease progression compared to those with <50%.

In contrast, patients with stromal tumor-infiltrating lymphocyte(sTIL) density in the lowest 25% showed no treatment response(ORR 0.0%), even when HER2-high cell proportions were elevated. This group also exhibited a 4.4-fold higher risk of disease progression.

These results indicate that accurately predicting treatment response requires not only HER2 expression levels but also consideration of the tumor immune microenvironment.

CEO Beomseok Seo of Lunit stated, “These studies meaningfully demonstrate the real-world applicability of AI biomarkers,” adding, “We will continue to expand collaborations with leading global institutions and companies to accelerate the commercialization of Lunit SCOPE as an essential tool in clinical oncology practice worldwide.”