Researchers have developed a deep learning method to automatically segment regions for the radiological Peritoneal Cancer Index (rPCI) from CT scans. The study evaluated nnU-Net and Swin UNETR on 62 CT scans, with nnU-Net achieving a Dice Similarity Coefficient of 0.82, which is close to human interobserver agreement. This approach aims to provide a non-invasive, imaging-based alternative to the current invasive diagnostic laparoscopy for assessing peritoneal metastases. AI
Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →
IMPACT Automated segmentation of rPCI regions could enable non-invasive, imaging-based assessment of peritoneal metastases, potentially improving clinical workflows.
RANK_REASON Academic paper detailing a new deep learning approach for medical image segmentation.