Search

This article discusses a study that aimed to develop a novel artificial intelligence model for predicting the risk of thromboembolic (TE) events after colonoscopy in patients on antithrombotic agents. The model, called PTRP, was trained using a cohort of 1182 patients and then validated against two additional cohorts. The results showed that the PTRP model was more accurate than the conventional CHA2DS2-VASc score in predicting TE events, with a lower number needed to screen to pick up one TE event. The study concludes that the PTRP model can be a useful tool in determining the need for prophylactic measures and the timely resumption of antithrombotic agents after colonoscopy. [Extracted from the article]