Related Experiment Video
Updated: Sep 29, 2025
10:17
Guidelines and Experience Using Imaging Biomarker Explorer IBEX for Radiomics
Published on: January 8, 2018
13.3K
Baseline CT-based Radiomic Features Aid Prediction of Nodal Positivity after Neoadjuvant Therapy in Pancreatic Cancer
Sherif B Elsherif1, Sanaz Javadi1, Ott Le1
1From the Departments of Diagnostic Radiology (S.B.E., S.J., O.L., E.P.T., P.R.B.) and Surgical Oncology (M.H.G.K.), The University of Texas MD Anderson Cancer Center, Houston, Tex; and MIM Software, Cleveland, Ohio (N.L.).
Radiology. Imaging Cancer
|March 25, 2022
Summary
Dual-energy CT textural features can predict pancreatic cancer lymph node metastasis and patient survival. A specific feature, integral total (∫ T) (HU·mL) (PPP), showed a significant association with outcomes, aiding in treatment response assessment.
Area of Science:
- Radiology
- Oncology
- Medical Imaging Analysis
Background:
- Pancreatic cancer poses a significant challenge due to late diagnosis and poor prognosis.
- Accurate prediction of treatment response and patient outcomes is crucial for effective management.
Purpose of the Study:
- To investigate the association between CT-derived textural features of pancreatic cancer and patient outcomes.
- To identify radiomic signatures predictive of lymph node metastasis and survival in pancreatic cancer patients.
Main Methods:
- Retrospective analysis of 54 pancreatic cancer patients treated with chemoradiation, surgery, and lymph node dissection.
- Extraction of radiomic features from dual-energy CT images (70-keV PPP and iodine material density images).
- Logistic regression and ROC analysis for feature selection and performance evaluation; Kaplan-Meier method for survival analysis.
Main Results:
- The integral total (∫ T) (HU·mL) (PPP) feature was inversely associated with post-therapy lymph node (ypN) category.
- A threshold of ∫ T (HU·mL) (PPP) < 507.85 predicted ypN1-2 with 96% sensitivity.
- Patients with lower ∫ T (HU·mL) (PPP) values exhibited significantly decreased overall survival (P = .006) and progression-free survival (P = .001).
Conclusions:
- CT-based radiomic features, particularly ∫ T (HU·mL) (PPP), can serve as a valuable tool for predicting ypN category in pancreatic cancer.
- These radiomic signatures may aid in assessing treatment response and prognosis, potentially guiding clinical decision-making.
- Further validation of this CT-derived radiomic signature is warranted for broader clinical application in pancreatic cancer management.
