Related Experiment Video
Updated: Aug 29, 2025
Utilizing High Resolution Ultrasound to Monitor Tumor Onset and Growth in Genetically Engineered Pancreatic Cancer Models
Published on: April 7, 2018
Pancreatic Cancer Detection on CT Scans with Deep Learning: A Nationwide Population-based Study
Po-Ting Chen1, Tinghui Wu1, Pochuan Wang1
1From the Department of Medical Imaging (P.T.C., K.L.L.) and Division of Gastroenterology and Hepatology, Department of Internal Medicine (M.S.W., W.C.L.), National Taiwan University Hospital, National Taiwan University College of Medicine, Taipei, Taiwan; Institute of Applied Mathematical Sciences (T.W., D.C., W.W.) and Departments of Computer Science and Information Engineering (P.W.) and Internal Medicine, College of Medicine (M.S.W., W.C.L.), National Taiwan University, No. 1, Section 4, Roosevelt Rd, Taipei 10617, Taiwan; Department of Medical Imaging, National Taiwan University Cancer Center, Taipei, Taiwan (K.L.L.); NVIDIA, Bethesda, Md (H.R.R.); and National Health Insurance Administration, Ministry of Health and Welfare, Taipei, Taiwan (P.C.L.).
A new deep learning tool accurately detects pancreatic cancer on CT scans, improving detection rates for small tumors often missed. This AI-powered approach shows high sensitivity and specificity, aiding early diagnosis.
Area of Science:
- Radiology
- Artificial Intelligence
- Oncology
Background:
- Pancreatic tumors under 2 cm are frequently missed on abdominal CT, impacting early diagnosis.
- Developing advanced tools for pancreatic cancer detection is crucial.
Purpose of the Study:
- To develop and validate a deep learning (DL) tool for detecting pancreatic cancer using CT scans.
- To assess the tool's performance, especially for small tumors.
Main Methods:
- An end-to-end DL tool, combining segmentation and classification convolutional neural networks (CNNs), was developed.
- The tool was trained and validated on retrospective contrast-enhanced CT studies from cancer patients and control subjects.
- Performance was evaluated against radiologist interpretations using the McNemar test.
Main Results:
- The DL tool achieved 89.9% sensitivity and 95.9% specificity in an internal test set (AUC, 0.96).
- In a real-world validation set (1473 studies), sensitivity was 89.7% and specificity 92.8% (AUC, 0.95).
- The tool demonstrated 74.7% sensitivity for detecting pancreatic malignancies smaller than 2 cm.
Conclusions:
- A DL-based tool can accurately detect pancreatic cancer on CT scans.
- The tool shows promise in improving the detection of small pancreatic tumors, addressing a significant clinical challenge.
