Skip to main content

ORIGINAL RESEARCH article

Front. Oncol.
Sec. Cancer Imaging and Image-directed Interventions
Volume 14 - 2024 | doi: 10.3389/fonc.2024.1362850
This article is part of the Research Topic Quantitative Imaging: Revolutionizing Cancer Management with biological sensitivity, specificity, and AI integration View all 14 articles

Weakly supervised large-scale pancreatic cancer detection using multiinstance learning

Provisionally accepted
Shyamapada Mandal Shyamapada Mandal 1Keerthiveena B Keerthiveena B 1Hariprasad Kodamana Hariprasad Kodamana 1Chetan Arora Chetan Arora 1Julie Clark Julie Clark 2David Kwon David Kwon 2,3Anurag S. Rathore Anurag S. Rathore 1*
  • 1 Indian Institute of Technology Delhi, New Delhi, National Capital Territory of Delhi, India
  • 2 Henry Ford Pancreatic Cancer Center, Detroit, United States
  • 3 Henry Ford Hospital, Detroit, Michigan, United States

The final, formatted version of the article will be published soon.

    Early detection of pancreatic cancer continues to be a challenge due to the difficulty in accurately identifying specific signs or symptoms that might correlate with the onset of pancreatic cancer. Unlike breast or colon or prostate cancer where screening tests are often useful in identifying cancerous development, there are no tests to diagnose pancreatic cancers. As a result, most pancreatic cancers are diagnosed at an advanced stage, where treatment options, whether systemic therapy, radiation, or surgical interventions, offer limited efficacy.Methods: A two-stage weakly supervised deep learning-based model has been proposed to identify pancreatic tumors using Computed Tomography (CT) images from Henry Ford Health (HFH) and publicly available Memorial Sloan Kettering Cancer Center (MSKCC) data sets. In the first stage, nnU-Net supervised segmentation model was used to crop an area in the location of the pancreas, which was trained on the MSKCC repository of 281 patient image sets with established pancreatic tumors. In the second stage, a multi-instance learning-based weakly-supervised classification model was applied on the cropped pancreas region to segregate pancreatic tumors from normal appearing pancreas. The model was trained, tested, and validated on images obtained from a HFH repository with 463 cases and 2882 controls.The proposed deep learning model, the two-stage architecture, offers an accuracy of 0.907±0.01, sensitivity of 0.905±0.01, specificity of 0.908±0.02, and AUC (ROC) 0.903±0.01. The two-stage framework can automatically differentiate pancreatic tumours from non-tumors pancreas with improved accuracy on HFH dataset.The proposed two-stage deep learning architecture shows significantly enhanced performance for predicting the presence of a tumor in the pancreas using CT images compared to other reported studies in the literature.

    Keywords: Pancreatic Cancer, Multi-instance learning, image segmentation, feature extraction, Medical Image Analysis

    Received: 29 Dec 2023; Accepted: 01 Aug 2024.

    Copyright: © 2024 Mandal, B, Kodamana, Arora, Clark, Kwon and Rathore. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

    * Correspondence: Anurag S. Rathore, Indian Institute of Technology Delhi, New Delhi, 110016, National Capital Territory of Delhi, India

    Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.