Longitudinal Analysis of Manually and Automatically Classified Circulating Tumor Biomarkers and their Prediction of Survival in Metastatic Colorectal Cancer.

Circulating tumor cells (CTCs) and tumor-derived extracellular vesicles (tdEVs) are promising biomarkers for predicting survival and informing treatment decisions in metastatic colorectal cancer (mCRC); yet their clinical application remains limited. In this study, we analyzed CellSearch image archives from 446 patients with mCRC treated in the CAIRO2 study to evaluate the predictive value of CTCs and tdEVs and explore their utility in guiding personalized therapy. Using pharmacodynamic modeling, we examined longitudinal changes in manually and automatically classified CTC and tdEV counts, assessing their relationship with tumor size dynamics and overall survival. Automated tdEV counts demonstrated the strongest association with survival, followed by automated CTCs, which outperformed manually assessed counts. The combination of automated tdEVs and manual CTCs further improved predictive performance. Simulations predict transitioning high-risk patients (CTC/tdEV score >1.75) to second-line FOLFIRI at week 4 or 10 improves median survival from 15.1 months (12.8-18.4 CI) to 22.7 months (17.1-35.8 CI), and from 13.3 months (10.5-21.7 CI) to 23.3 months (17.8-31.2 CI), respectively. Additionally, biomarker monitoring demonstrated reduced cost (€3,913 vs. €7,802) and environmental burden (10 kg vs. 167 kg CO2e per patient). These findings suggest that tdEVs, alone or in combination with CTCs, may help optimize treatment timing and outcomes in mCRC. The integration of CTCs and tdEVs into clinical practice could offer a personalized, cost-effective, and more sustainable alternative to routine imaging in managing advanced colorectal cancer.