Lunit's SCOPE IO artificial intelligence platform has demonstrated significant predictive capabilities for immunotherapy response across multiple cancer types, with three studies presented at the European Society for Medical Oncology (ESMO) Congress 2025 in Berlin, Germany, from October 17-21.
The South Korean company's AI-powered histopathology platform analyzed routine tissue slides to generate biomarkers that successfully stratified patients based on their likelihood of responding to immune checkpoint inhibitor therapy across colorectal, kidney, and lung cancers.
Breakthrough Results in Colorectal Cancer
The most significant findings emerged from collaborative studies with the University of Pisa, Italy, examining patients with proficient mismatch repair (pMMR) metastatic colorectal cancer enrolled in the AtezoTRIBE and AVETRIC trials. SCOPE IO analyzed pre-treatment histology slides and quantified multiple cell types within the tumor microenvironment to generate biomarkers that divided patients into "biomarker-high" and "biomarker-low" groups.
In the AtezoTRIBE trial, biomarker-high patients treated with atezolizumab plus FOLFOXIRI/bevacizumab demonstrated significantly improved progression-free survival and overall survival compared to biomarker-low patients, while no such benefit was observed in the control arm. The AVETRIC cohort validation confirmed these improved survival outcomes for biomarker-high patients receiving ICI-based therapy.
These results address an urgent unmet medical need, as the findings suggest that an AI-derived tumor microenvironment biomarker could help identify patients with pMMR mCRC who are most likely to benefit from immunotherapy combinations.
Promising Outcomes in Kidney Cancer
A collaborative study with Yonsei University College of Medicine, Korea, evaluated AI-defined immune phenotypes in patients with advanced clear cell renal cell carcinoma (ccRCC). The research compared outcomes between patients treated with nivolumab plus ipilimumab (NIVO+IPI) versus sunitinib (SUN).
SCOPE IO classified tumors as inflamed or non-inflamed based on the density and spatial distribution of tumor-infiltrating lymphocytes. Patients with inflamed tumors treated with NIVO+IPI demonstrated significantly longer progression-free survival, overall survival, and notably higher response rates of 60.5% versus 23.2% compared to those with non-inflamed tumors. No such benefit was observed in the sunitinib treatment arm.
The findings were validated in an independent ccRCC cohort and aligned with inflamed gene expression signatures from The Cancer Genome Atlas, supporting AI-based immune phenotyping as a promising biomarker for guiding treatment selection between immunotherapy combinations and targeted therapies in first-line ccRCC treatment.
Validation in Lung Cancer
The third study, conducted with Japan's National Cancer Center Hospital East (NCCHE), further validated SCOPE IO's predictive power in non-small cell lung cancer (NSCLC) patients. In this multicenter prospective study, tumors classified as inflamed showed significantly better responses and longer survival with ICI therapy compared to non-inflamed tumors. Importantly, this difference was not observed among patients treated with cytotoxic chemotherapy, strengthening the evidence supporting SCOPE IO as a predictive biomarker specifically for immunotherapy benefit in NSCLC.
Clinical Implications and Future Directions
"These findings show the potential of Lunit SCOPE IO to help identify patients who will truly benefit from immunotherapy – whether in colorectal or kidney cancer – and to guide treatment strategies that can make cancer care more precise and effective," said Brandon Suh, CEO of Lunit.
The AI platform has previously demonstrated capabilities in identifying specific patterns in rare tumor samples that correlate with better treatment outcomes and improving HER2 biomarker evaluation in metastatic colorectal cancer patients undergoing HER2-targeted therapy.
SCOPE IO's ability to generate predictive biomarkers directly from routine pathology slides represents a significant advancement in precision oncology, potentially enabling more personalized treatment selection across multiple cancer types without requiring additional specialized testing.
