Nomogram Based on MRI Radiomics Combined with Clinical Features for Assessing Lymphovascular Invasion in Node-Negative Rectal Cancer

Motivation: Given that lymphovascular invasion (LVI) detection typically relies on invasive pathological biopsy, developing a non-invasive, personalized technique to accurately and effectively predict the preoperative LVI status in rectal cancer (RC) patients is crucial. This would aid clinicians in stratified management and precise treatment of RC patients. Goal(s): To investigate the application value of intra- and peritumoral radiomics in predicting LVI in RC patients. Approach: A predictive model was developed using intra- and peritumoral radiomics combined with clinicopathological features to predict the preoperative LVI status in RC patients. Results: The nomogram model combining intra- and peritumoral radiomics demonstrates enhanced performance in predicting LVI. Impact: The treatment approach for RC without lymph node and distant metastasis but with LVI-positive is equivalent to that for with LNM-positive. The nomogram model offers a non-invasive tool for prediction of LVI, avoiding the risks associated with invasive pathological detection.