Case Study: a prestigious health research institute achieves earlier colon cancer relapse detection with Quibim's AI predictive model

Early detection of localized colon cancer relapse using an AI-based predictive model

The customer, a prestigious health research institute, was challenged with identifying colon cancer patients at high risk of relapse. Traditional and newer liquid biopsy methods for detecting minimal residual disease (MRD) were either invasive or offered limited sensitivity. The institute partnered with Quibim to create a more effective prognostic tool by integrating liquid biopsy with standard-of-care CT scan data.

Quibim developed an AI-based predictive model by analyzing retrospective multi-center data, including clinical information, liquid biopsies, and CT scans. Their solution, integrated into the QP-Discovery® platform, combined radiomic, deep, and fractal features with other patient data. The resulting model achieved 87.9% accuracy in predicting relapse, a significant increase over using clinical variables alone (60%), enabling better patient stratification and the potential for personalized treatment plans.