PMNet: A probability map based scaled network for breast cancer diagnosis. (April 2021)

• Record Type: Journal Article
• Authors: Ahmed, Salman; Tariq, Maria; Naveed, Hammad
• Is Part Of: Computerized medical imaging and graphics. Volume 89(2021)
• Publication Date: 2021-04
• Subjects: Deep learning -- Digital pathology -- Tumor detection -- PMNet -- Whole slide image analysis; Diagnostic imaging -- Periodicals; Imaging systems in medicine -- Periodicals; Diagnosis, Radioscopic -- Data processing -- Periodicals; Diagnostic Imaging -- Periodicals; Imagerie pour le diagnostic -- Périodiques; Diagnostic imaging; Periodicals; Electronic journals; Electronic journals; 616.0754
• DOI: 10.1016/j.compmedimag.2021.101863
• Languages: English
• ISSNs: 0895-6111

Abstract

Graphical abstract: Highlights: PMNet is a novel two-stage pipeline that can integrate local and global features in a WSI unlike previous studies for breast cancer classification. We demonstrate the superior performance of our methodology on two publicly available datasets (BACH and Dryad) as compared to previous state-of-the-art methodologies. We proposed a method to generate patch level annotations for the image level TCGA breast cancer database. This labelled data will be useful for future deep learning methods. Abstract: The mortality rate of Breast Cancer in women has increased, both in west and east. Early detection is important to improve the survival rate of cancer patients. The manual detection and identification of cancer in whole slide images are critical and difficult tasks for pathologists. In this work, we introduce PMNet, a pipeline to detect regions with invasive characteristics in whole slide images. Our method employs scaled networks for detecting breast cancer in whole slide images. It classifies whole slide images on patch level into normal, benign, in situ and invasive tumors. Our approach yielded f1-score of 88.9( ± 1.7)% that outperforms the benchmark f1-score of 81.2( ± 1.3)% on patch level and achieved an average dice coefficient of 69.8% on 10 whole slide images compared to the benchmark average dice coefficient of 61.5% on BACH dataset. Similarly, on the dryad test dataset that comprises of 173 whole slide images, we achieved an average dice coefficient of 82.7% as compared to the previous state-of-art of 76% without fine-tuning on this dataset. We further proposed a method to generate patch level annotations for the image level TCGA breast cancer database that will be useful for future deep learning methods.