During this year's congress,
A collaborative study indicated that AI-based immune phenotyping can predict therapy outcomes in advanced biliary tract cancer (BTC) patients who are planning to be treated with anti-PD-1 therapy. A total of 337 H&E-stained whole slide images (WSI) were acquired for assessment. In the study, the research team defined the immune phenotype of the WSI samples via Lunit SCOPE IO,
In another study, it was found that TIL density in tumor microenvironment is highly correlated with favorable treatment response to immune checkpoint inhibitor (ICI) in head and neck squamous cell carcinoma (HNSCC). Assessed by Lunit SCOPE IO, patients with high-TIL showed a higher objective response rate (21.6% vs 5.7%) and more favorable median progression-free survival (3.2 vs 1.6 months).
In another study, HER2 (human epidermal growth factor receptor-2) scoring in biliary tract cancer was evaluated using Lunit SCOPE HER2. The analysis showed a substantial concordance of 75.3% between AI and human pathologists' assessments.
Another study aimed to predict multiple druggable mutations in non-small cell lung cancer (NSCLC) based on AI analysis of H&E images. More than 3,000 NSCLC samples were used as training data to develop an AI-powered predictive model. In validation in an independent dataset, the model showed robust performance in predicting six types of mutations (EGFR-mt, KRAS-mt, ALK-tr, ROS1-tr, RET-tr, MET-ex). Notably, for MET exon skipping mutations, the model achieved a high positive predictive value (PPV), showing that test-positive patients were three times more likely to have true-MET-ex positive mutations compared to the overall patient population. Moreover, specificity and PPV for identifying patients without mutations (All-WT) were 99.2% and 95.2% respectively, which means with AI assistance unnecessary tests can be avoided. Following the results, it is expected that the newly developed AI genotype predictor, available for multiple genotypes in non-small cell lung cancer, holds immense potential for widespread adoption by clinicians and pharmaceutical industry leaders globally.
'We are thrilled to be at this year's ESMO with nine groundbreaking study results that prove the effectiveness of the Lunit SCOPE AI WSI analyzer and biomarker platform,' said
Contact:
Email: contact@lunit.io
(C) 2023 Electronic News Publishing, source