Detection of pancreatic cancer with two- and three-dimensional radiomic analysis in a nationwide population-based real-world dataset

Q: Why is an automated CAD tool needed for pancreatic cancer detection?

CT can miss a substantial proportion of early pancreatic cancers, especially tumors smaller than 2 cm, and diagnostic performance depends on the radiologist’s interpretation and workload.

Q: How were the datasets used to develop and test the CAD tool?

Contrast-enhanced CT studies from 2005–2019 were retrospectively collected for training (546 pancreatic adenocarcinoma cases and 733 controls). The model was then tested on a nationwide dataset comprising 1,477 CT studies from multiple institutions in Taiwan.

Q: What was the diagnostic performance of the CAD tool?

Overall sensitivity was 0.918 and specificity was 0.822 for differentiating pancreatic cancer from non-cancer studies, with an AUC of 0.947. Sensitivity for tumors under 2 cm was 0.707.

Sep 24, 2026

Uploaded by nayy☆

Contrast-enhanced CT remains the primary modality for pancreatic cancer detection, yet up to ~40% of tumors under 2 cm can be missed by radiologists, motivating reliable computer-aided detection. Using retrospectively collected CT studies from 2005–2019 (546 pancreatic adenocarcinoma patients and 733 controls; tested on 1,477 nationwide scans from Taiwan), a machine-learning CAD system combined 2D and 3D radiomic features to distinguish cancer from normal pancreas. The model showed high accuracy overall, with reduced sensitivity for small tumors, supporting robust clinical potential.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages13

Reading Time33 min

AccessFree to download

Background

Problem: CT misses early pancreatic cancer

Radiomics and machine learning motivation

Methods

Study design and datasets

2D and 3D radiomic CAD approach

Results

Diagnostic performance and operating points

Performance for tumors < 2 cm

Conclusions

Chang etal. BMC Cancer (2023) 23:58 [https://doi.org/10.1186/s12885-023-10536-8](https://doi.org/10.1186/s12885-023-10536-8) BMC Cancer RESEARCH Open Access Detection of pancreatic cancer with twoand three-dimensional radiomic analysis in a nationwide population-based real-world dataset Dawei Chang1, Po‑Ting Chen2, Pochuan Wang3, Tinghui Wu4, Andre Yanchen Yeh5, Po‑Chang Lee6, Yi‑Hui Sung6, Kao‑Lang Liu2,7, Ming‑Shiang Wu8,9, Dong Yang10, Holger Roth10, Wei‑Chih Liao8,9*† and Weichung Wang4*† Abstract Background CT is the major detection tool for pancreatic cancer (PC) . However, approximately 40% of PCs < 2 cm are missed on CT, underscoring a pressing need for tools to supplement radiologist interpretation. Methods Contrast‑enhanced CT studies of 546 patients with pancreatic adenocarcinoma diagnosed by histology/ cytology between January 2005 and December 2019 and 733 CT studies of controls with normal pancreas obtained between the same period in a tertiary referral center were retrospectively collected for developing an automatic end‑ to‑end computer‑aided detection (CAD) tool for PC using two‑dimensional (2D) and three‑dimensional (3D) radiomic analysis with machine learning. The CAD tool was tested in a nationwide dataset comprising 1,477 CT studies (671 PCs, 806 controls) obtained from institutions throughout Taiwan. Results The CAD tool achieved 0.918 (95% CI, 0 .895–0. 938) sensitivity and 0.822 (95% CI, 0 .794–0. 848) specificity in differentiating between studies with and without PC (area under curve 0 . 947, 95% CI, 0 .936–0. 958), with 0.707 (95% CI, 0 .602–0. 797) sensitivity for tumors < 2 cm. The positive and negative likelihood ratios of PC were 5.17 (95% CI, 4.45–6. 01) and 0.10 (95% CI, 0 .08–0. 13), respectively. Where high specificity is needed, using 2D and 3D analyses in series yielded 0.952 (95% CI, 0 .934–0. 965) specificity with a sensitivity of 0.742 (95% CI, 0 .707–0. 775), whereas using 2D and 3D analyses in parallel to maximize sensitivity yielded 0.915 (95% CI, 0 .891–0. 935) sensitivity at a specificity of 0.791 (95% CI, 0 .762–0. 819) . Conclusions The high accuracy and robustness of the CAD tool supported its potential for enhancing the detection of PC. Keywords Pancreatic cancer, Pancreatic ductal adenocarcinoma, Radiomics, Machine learning, Computer‑aided detection †Wei‑Chih Liao and Weichung Wang contributed equally to this work as corresponding authors. *Correspondence: Wei‑Chih Liao [david.ntuh@gmail.com](david.ntuh@gmail.com)[ ](david.ntuh@gmail.com)Weichung Wang [wwang@ntu.edu.tw](wwang@ntu.edu.tw) Full list of author information is available at the end of the article © The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit [http://creativecommons.org/licenses/by/4.0/](http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver ()[. The Creative Commons Public Domain Dedication waiver (](http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver ()[http://creativeco](http://creativeco)[mmons.org/publicdomain/zero/1.0/](mmons.org/publicdomain/zero/1.0/)) applies to the data made available in this article, unless otherwise stated

FAQs

Why is an automated CAD tool needed for pancreatic cancer detection?

CT can miss a substantial proportion of early pancreatic cancers, especially tumors smaller than 2 cm, and diagnostic performance depends on the radiologist’s interpretation and workload.

How were the datasets used to develop and test the CAD tool?

Contrast-enhanced CT studies from 2005–2019 were retrospectively collected for training (546 pancreatic adenocarcinoma cases and 733 controls). The model was then tested on a nationwide dataset comprising 1,477 CT studies from multiple institutions in Taiwan.

What was the diagnostic performance of the CAD tool?

Overall sensitivity was 0.918 and specificity was 0.822 for differentiating pancreatic cancer from non-cancer studies, with an AUC of 0.947. Sensitivity for tumors under 2 cm was 0.707.

Sep 24, 2026

Uploaded by nayy☆