EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

CT-Based Radiomics for Preoperative Differentiation of Pancreatic Cancer From Periampullary Non-pancreatic Cancer

Yu Guo iD, Yang-Jie Chen, Wen-Jing Cui, Li Zhu, Zhong-Qiu Wang, Shuai Ren iD

Cancer Control · 2026-09 · 33

DOI: 10.1177/10732748261511508 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Cancer Control, online erschienen am 20.11.2026. Verfasst von Yu Guo und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Radiology, Affiliated Hospital of Nanjing University of Chinese Medicine, Nanjing, China. Hauptthemen laut MEDLINE: Pancreatic Neoplasms, Tomography, X-Ray Computed, Common Bile Duct Neoplasms. Die Arbeit nennt 42 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: Jiangsu Provincial Key Research and Development Program, Developing Program for High-level Academic Talent in Jiangsu Hospital of Chinese Medicine, Postgraduate Research & Practice Innovation Program of Jiangsu Province.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Accurate preoperative differentiation between pancreatic cancer and non-pancreatic periampullary carcinoma is important for treatment selection. This study aimed to develop a predictive model integrating CT-based radiomics and conventional imaging features for this task.

Methods

This single-center retrospective study developed and internally validated prediction models using data from 72 patients with pathologically confirmed periampullary carcinoma, including 37 pancreatic and 35 non-pancreatic cancers. Clinical data and conventional CT findings were collected, and radiomics features were extracted from three-dimensional tumor volumes on preoperative portal venous phase CT images. Model development and evaluation used stratified five-fold nested cross-validation, with data-dependent preprocessing and predictor selection restricted to the corresponding training folds. Radiomics feature selection involved Pearson correlation filtering followed by LASSO logistic regression, with the regularization parameter selected using inner five-fold cross-validation and the one-standard-error rule. Clinical-imaging, radiomics, and combined models were constructed using logistic regression. Diagnostic performance was evaluated using pooled out-of-fold predictions from the outer test folds and summarized by the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and accuracy. AUCs were compared using DeLong’s test.

Results

The AUCs of the clinical-imaging and radiomics models were 0.825 (0.715-0.920) and 0.856 (0.753-0.944), respectively. The combined model achieved the best performance, with an AUC of 0.924 (0.876-0.983), sensitivity of 85.71%, specificity of 94.59%, and accuracy of 90.28%. Lesion enhancement degree and radiomics score were the key predictors in the combined model. DeLong tests showed that the combined model significantly outperformed both the clinical-imaging model ( P = 0.011) and the radiomics model ( P = 0.021).

Conclusions

The combined model integrating CT-based radiomics and conventional imaging features showed promising discriminative performance for the preoperative differentiation between pancreatic cancer and non-pancreatic periampullary carcinoma in this single-center cohort. Further prospective validation in larger, independent cohorts is required before clinical implementation.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Pancreatic NeoplasmsTomography, X-Ray ComputedCommon Bile Duct NeoplasmsHumansRadiomicsRetrospective StudiesFemaleMaleDiagnosis, DifferentialMiddle AgedAged

Stichwörter der Autor:innen

Computed tomographyDifferential diagnosisPancreatic cancerPeriampullary carcinomaRadiomics

Autor:innen und Institutionen

  • Department of Radiology, Affiliated Hospital of Nanjing University of Chinese Medicine, Nanjing, China.

ORCID-Profile: Yu Guo, Shuai Ren

Förderung

  • Jiangsu Provincial Key Research and Development Program
  • Developing Program for High-level Academic Talent in Jiangsu Hospital of Chinese Medicine
  • Postgraduate Research & Practice Innovation Program of Jiangsu Province
  • National Natural Science Foundation of China
  • Open Project of Jiangsu Provincial Science and Technology Resources (Clinical Resources) Coordination Service Platform
  • Foundation of Excellent Young Doctor of Jiangsu Province Hospital of Chinese Medicine
  • Pancreatic Innovation Fund under 45 (U45) of the Chinese Association of Pancreatology
  • Young Elite Scientists Sponsorship Program by China Association of Chinese Medicine

Zitieren

APA

Guo, Y., Chen, Y. J., Cui, W. J., Zhu, L., Wang, Z. Q., & Ren, S. (2026). CT-Based Radiomics for Preoperative Differentiation of Pancreatic Cancer From Periampullary Non-pancreatic Cancer. Cancer Control, 33. https://doi.org/10.1177/10732748261511508

Vancouver

Guo Y, Chen YJ, Cui WJ, Zhu L, Wang ZQ, Ren S. CT-Based Radiomics for Preoperative Differentiation of Pancreatic Cancer From Periampullary Non-pancreatic Cancer. Cancer Control. 2026;33. doi:10.1177/10732748261511508

BibTeX
@article{Guo2026,
  author = {Guo, Yu and Chen, Yang-Jie and Cui, Wen-Jing and Zhu, Li and Wang, Zhong-Qiu and Ren, Shuai},
  title = {CT-Based Radiomics for Preoperative Differentiation of Pancreatic Cancer From Periampullary Non-pancreatic Cancer},
  journal = {Cancer Control},
  year = {2026},
  volume = {33},
  doi = {10.1177/10732748261511508}
}
Kostenlos starten →