MRI-based clinical-radiomics model predicts tumor response before treatment in locally advanced rectal cancer.

摘要: Neoadjuvant chemo-radiotherapy (CRT) followed by total mesorectal excision (TME) represents the standard treatment for patients with locally advanced (≥ T3 or N+) rectal cancer (LARC). Approximately 15% of LARC shows a complete response after CRT. The use pre-treatment MRI as predictive biomarker could help to increase chance organ preservation tailoring neoadjuvant treatment. We present novel machine learning model combining MRI-based clinical and radiomic features early prediction in patients. scans (3.0 T, T2-weighted) 72 were included. Two readers independently segmented each tumor. Radiomic extracted from both “tumor core” (TC) border” (TB). Partial least square (PLS) regression was used multivariate, learning, algorithm choice leave-one-out nested cross-validation optimize hyperparameters PLS. MRI-Based “clinical-radiomic” properly predicted (AUC = 0.793, p = 5.6 × 10–5). Importantly, improved when features, latter from both TC TB. Prospective validation studies randomized trials are warranted better define role radiomics development precision medicine.