Machine Learning for Recurrence Risk of Pancreatic Cancer After Radical Resection

Multimodal Data Prediction Based on Machine Learning for Recurrence Risk of Pancreatic Cancer After Radical Resection

Recurrence of Pancreatic Cancer(PCa) is a multifactorial event. Based on the clinicopathological characteristics and imaging data of patients with PCa, the investigators used image processing and machine learning algorithms to build a more comprehensive and robust model, and added some unused features to explore its clinical application value.

A retrospective analysis of patients with PCa who underwent radical resection at Zhejiang Cancer Hospital (Hangzhou, China) from January 2013 to December 2020. The database was extracted from the preoperative demographics, blood markers, and surgical pathology information of patients undergoing radical PCa surgery in the investigators' hospital. The investigators used the PyRadiomics platform to extract image features.

Inclusion Criteria:

• Have the results of enhanced CT examination of the pancreas within 1 month before surgery in our hospital;
• Radical resection of pancreatic cancer was performed in our hospital;
• There are follow-up results in our hospital, and the follow-up endpoints include disease recurrence or at least 12 months.
• Complete clinical medical records and imaging data.

Exclusion Criteria:

• non-R0 resection;
• Combined with other malignant tumors
• The patient's imaging data has technical problems or the lesion is too small (less than 1cm), which is not suitable for omics analysis.