JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Novel CT radiomics models for the postoperative prediction of early recurrence of resectable pancreatic adenocarcinoma: A single-center retrospective study in China

  • 0Department of Hepatobiliary and Pancreatic Surgery, Peking University First Hospital, Beijing, China.
Journal of Applied Clinical Medical Physics +

|

April 12, 2025

|

April 12, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

Deep learning radiomics from CT scans effectively predict early recurrence in pancreatic cancer patients. This advanced approach significantly outperforms traditional methods for improved patient risk stratification.

Area Of Science

  • Oncology
  • Radiology
  • Artificial Intelligence

Background

  • Pancreatic ductal adenocarcinoma (PDAC) has a high rate of early recurrence after surgery.
  • Accurate prediction of early recurrence is crucial for timely intervention and improved patient outcomes.
  • Current prediction models often lack sufficient accuracy.

Purpose Of The Study

  • To evaluate the predictive capability of CT-based radiomics features for early recurrence (ER) in PDAC patients.
  • To compare the performance of a deep-radiomics model against clinical and conventional radiomics models.

Main Methods

  • Retrospective analysis of 250 postoperative PDAC patients with preoperative CT imaging.
  • Development of TNM staging, clinical, conventional radiomics, and deep-radiomics models for ER prediction (within 9 months).
  • Performance evaluation using ROC-AUC, PR-AUC, DCA, NRI, and IRI.

Main Results

  • The deep-radiomics model achieved the highest ROC-AUC (0.895) and PR-AUC (0.834), significantly outperforming other models.
  • TNM staging showed an ROC-AUC of 0.673, clinical model 0.640, and conventional radiomics 0.722.
  • Decision curve analysis and reclassification metrics confirmed the superior clinical utility of the deep-radiomics model.

Conclusions

  • Deep features extracted from CT images demonstrate strong predictive performance for early recurrence in PDAC.
  • The deep-radiomics model offers a significant advancement in predicting ER compared to existing methods.
  • This AI-driven approach holds promise for enhancing risk stratification and guiding treatment decisions in PDAC management.

Keywords: