Deep Learning-Based Classification of Hepatocellular Nodular Lesions on Whole-Slide Histopathologic Images
Na Cheng,Yong Ren,Jing Zhou,Yiwang Zhang,Deyu Wang,Xiaofang Zhang,Bing Chen,Fang Liu,Jin Lv,Qinghua Cao,Sijin Chen,Hong Du,Dayang Hui,Zijin Weng,Qiong Liang,Bojin Su,Luying Tang,Lanqing Han,Jianning Chen,Chunkui Shao
DOI: https://doi.org/10.1053/j.gastro.2022.02.025
IF: 29.4
Gastroenterology
Abstract:Background & aims: Hepatocellular nodular lesions (HNLs) constitute a heterogeneous group of disorders. Differential diagnosis among these lesions, especially high-grade dysplastic nodules (HGDNs) and well-differentiated hepatocellular carcinoma (WD-HCC), can be challenging, let alone biopsy specimens. We aimed to develop a deep learning system to solve these puzzles, improving the histopathologic diagnosis of HNLs (WD-HCC, HGDN, low-grade DN, focal nodular hyperplasia, hepatocellular adenoma), and background tissues (nodular cirrhosis, normal liver tissue). Methods: The samples consisting of surgical and biopsy specimens were collected from 6 hospitals. Each specimen was reviewed by 2 to 3 subspecialists. Four deep neural networks (ResNet50, InceptionV3, Xception, and the Ensemble) were used. Their performances were evaluated by confusion matrix, receiver operating characteristic curve, classification map, and heat map. The predictive efficiency of the optimal model was further verified by comparing with that of 9 pathologists. Results: We obtained 213,280 patches from 1115 whole-slide images of 738 patients. An optimal model was finally chosen based on F1 score and area under the curve value, named hepatocellular-nodular artificial intelligence model (HnAIM), with the overall 7-category area under the curve of 0.935 in the independent external validation cohort. For biopsy specimens, the agreement rate with subspecialists' majority opinion was higher for HnAIM than 9 pathologists on both patch level and whole-slide images level. Conclusions: We first developed a deep learning diagnostic model for HNLs, which performed well and contributed to enhancing the diagnosis rate of early HCC and risk stratification of patients with HNLs. Furthermore, HnAIM had significant advantages in patch-level recognition, with important diagnostic implications for fragmentary or scarce biopsy specimens.