Predicting peritoneal recurrence in gastric cancer with serosal invasion using a pathomics nomogram

Area of Science:

• Oncology
• Digital Pathology
• Biomarker Discovery

Background:

• Peritoneal recurrence is a major challenge in gastric cancer (GC) with serosal invasion.
• Current methods inadequately predict peritoneal recurrence risk in these patients.
• Pathomics analysis shows promise for improved risk stratification.

Purpose of the Study:

• To develop and validate a pathomics signature for predicting peritoneal recurrence in GC.
• To create a nomogram integrating pathomics and clinical factors for personalized risk assessment.

Main Methods:

• Extraction of pathomics features from digital hematoxylin and eosin-stained GC images.
• Development of a pathomics signature based on identified features.
• Construction of a competing-risk pathomics nomogram incorporating clinical variables and the pathomics signature.

Main Results:

• The pathomics signature was significantly associated with peritoneal recurrence.
• The developed pathomics nomogram demonstrated favorable discrimination and calibration.
• The nomogram effectively integrated pathomics and clinical data for risk prediction.

Conclusions:

• A novel pathomics signature serves as a valuable predictive indicator for peritoneal recurrence in GC.
• The pathomics nomogram offers a reliable tool for assessing individual risk of peritoneal recurrence in serosal-invasive GC.