Lunit Presents Study on Standardizing Immune Phenotyping Analysis at U.S. Immunotherapy Conference

Medical artificial intelligence (AI) company Lunit announced on the 10th that it presented three research findings utilizing its AI biomarker platform, Lunit Scope, at the 2025 Society for Immunotherapy of Cancer (SITC 2025) held in Maryland, USA, from November 5 to 9.

First, Lunit, in collaboration with global contract research organization (CRO) CellCarta, presented a study demonstrating the potential for standardizing AI-based immune phenotyping analysis. The research team analyzed hematoxylin and eosin (H&E) slides from a total of 93 patients with non-small cell lung cancer, colorectal cancer, and urothelial carcinoma using Lunit Scope IO, and compared the AI analysis results with those of pathologists. The findings showed a very high level of concordance between the AI analysis and the pathologists' interpretations.

In particular, the Spearman's correlation coefficient-a statistical measure of rank correlation between two variables-was 0.91 for tumor cell nests (TCN) and 0.86 for tumor-associated stroma (TAS), indicating a very high degree of agreement. This coefficient approaches 1 as the correlation increases.

This study once again demonstrated that AI can interpret the immune environment of cancer tissue with precision and at a level comparable to specialists, using only basic H&E slides without additional staining or complex experimental procedures. It reaffirmed the advantages of utilizing AI in evaluating cancer immune response and analyzing immune phenotypes.

Lunit also demonstrated the potential of AI utilization in the development of antibody-based new drugs. The research team used Lunit Scope uIHC to analyze 47,591 IHC images across 34 cancer types, evaluating spatial interactions between 74 membrane protein targets-being studied as candidates for antibody-drug conjugates (ADC) and bispecific T-cell engagers (BiTE)-and tumor-infiltrating lymphocytes (TIL).

The analysis showed that for most protein targets, TIL density decreased in the corresponding regions. However, two targets-PD-L1 and TNFRSF4-showed a strong positive correlation with TIL infiltration. Notably, FGFR4 (Fibroblast Growth Factor Receptor 4) exhibited a marked increase in intratumoral lymphocyte infiltration in colorectal cancer, non-squamous non-small cell lung cancer, and uterine cancer, confirming its potential as a promising target for bispecific antibody therapeutics.

This study demonstrated that Lunit Scope uIHC can precisely analyze the spatial relationship between immune cells and proteins within cancer tissue, enabling the rapid and accurate identification of new drug candidate targets.

Seo Bumseok, CEO of Lunit, stated, "The studies presented at SITC 2025 show that Lunit Scope has established expertise in predicting immunotherapy response and is expanding its scope into the field of antibody drug development. We will strengthen collaboration with global partners such as CellCarta to create tangible medical value and establish a new standard for AI-powered precision medicine."

This content was produced with the assistance of AI translation services.