Preoperative prediction of early recurrence in pancreatic cancer: A novel clinical-radiomics model.

Journal: International journal of cancer

The study (DOI: 10.1002/ijc.70142) developed and validated a preoperative prediction model for early postoperative recurrence in pancreatic ductal adenocarcinoma (PDAC).

Data from 895 treatment-naïve patients who underwent direct resection were used. Early recurrence was defined as relapse within 6 months post-surgery.

The integrated clinical-radiomics model combines:

  • Four clinical variables: gender, CA125, radiologic N stage, and adjuvant treatment
  • Twenty-nine radiomic features extracted from preoperative CT scans

This model demonstrated strong predictive performance with AUCs of:

  • 0.862 in the training set
  • 0.843 in internal validation
  • 0.848 in external validation

It outperformed models based on clinical or radiomic data alone.

High-risk patients identified by this model had significantly shorter disease-free and overall survival.

This tool enables personalized treatment planning by identifying PDAC patients at high risk for early recurrence prior to surgery.

Leave a Reply