Deep learning system for lymph node quantification and metastatic cancer identification from whole-slide pathology images

• Published in: Gastric cancer : official journal of the International Gastric Cancer Association and the Japanese Gastric Cancer Association Vol. 24; no. 4; pp. 868 - 877
• Main Authors: Hu, Yajie, Su, Feng, Dong, Kun, Wang, Xinyu, Zhao, Xinya, Jiang, Yumeng, Li, Jianming, Ji, Jiafu, Sun, Yu
• Format: Journal Article
• Language: English
• Published: Singapore Springer Singapore 01.07.2021; Springer Nature B.V
• Subjects: Abdominal Surgery; Algorithms; Cancer; Cancer Research; Deep Learning; Diagnosis; Gastric cancer; Gastroenterology; Humans; Image Processing, Computer-Assisted - methods; Lymph nodes; Lymph Nodes - pathology; Lymphatic Metastasis - diagnosis; Lymphatic system; Medicine; Medicine & Public Health; Metastases; Metastasis; Oncology; Original Article; Predictive Value of Tests; Prospective Studies; Stomach Neoplasms - diagnosis; Stomach Neoplasms - pathology; Surgical Oncology; Deep learning; Lymph node metastasis; Lymph node quantification; Gastric cancer
• ISSN: 1436-3291; 1436-3305; 1436-3305
• DOI: 10.1007/s10120-021-01158-9

Background Traditional diagnosis methods for lymph node metastases are labor-intensive and time-consuming. As a result, diagnostic systems based on deep learning (DL) algorithms have become a hot topic. However, current research lacks testing with sufficient data to verify performance. The aim of this study was to develop and test a deep learning system capable of identifying lymph node metastases. Methods 921 whole-slide images of lymph nodes were divided into two cohorts: training and testing. For lymph node quantification, we combined Faster RCNN and DeepLab as a cascade DL algorithm to detect regions of interest. For metastatic cancer identification, we fused Xception and DenseNet-121 models and extracted features. Prospective testing to verify the performance of the diagnostic system was performed using 327 unlabeled images. We further validated the proposed system using Positive Predictive Value (PPV) and Negative Predictive Value (NPV) criteria. Results We developed a DL-based system capable of automated quantification and identification of metastatic lymph nodes. The accuracy of lymph node quantification was shown to be 97.13%. The PPV of the combined Xception and DenseNet-121 model was 93.53%, and the NPV was 97.99%. Our experimental results show that the differentiation level of metastatic cancer affects the recognition performance. Conclusions The diagnostic system we established reached a high level of efficiency and accuracy of lymph node diagnosis. This system could potentially be implemented into clinical workflow to assist pathologists in making a preliminary screening for lymph node metastases in gastric cancer patients.