Accurate diagnosis and prognosis prediction of gastric cancer using deep learning on digital pathological images: A retrospective multicentre study

Abstract

Background: To reduce the high incidence and mortality of gastric cancer (GC), we aimed to develop deep learning-based models to assist in predicting the diagnosis and overall survival (OS) of GC patients using pathological images. Methods: 2333 hematoxylin and eosin-stained pathological pictures of 1037 GC patients were collected from two cohorts to develop our algorithms, Renmin Hospital of Wuhan University (RHWU) and the Cancer Genome Atlas (TCGA). Additionally, we gained 175 digital pictures of 91 GC patients from National Human Genetic Resources Sharing Service Platform (NHGRP), served as the independent external validation set. Two models were developed using artificial intelligence (AI), one named GastroMIL for diagnosing GC, and the other named MIL-GC for predicting outcome of GC. Findings: The discriminatory power of GastroMIL achieved accuracy 0.920 in the external validation set, superior to that of the junior pathologist and comparable to that of expert pathologists. In the prognostic model, Cindices for survival prediction of internal and external validation sets were 0.671 and 0.657, respectively. Moreover, the risk score output by MIL-GC in the external validation set was proved to be a strong predictor of OS both in the univariate (HR = 2.414, P < 0.0001) and multivariable (HR = 1.803, P = 0.043) analyses. The predicting process is available at an online website (https://baigao.github.io/Pathologic-Prognostic-Analysis/). Interpretation: Our study developed AI models and contributed to predicting precise diagnosis and prognosis of GC patients, which will offer assistance to choose appropriate treatment to improve the survival status of GC patients. Funding: Not applicable.

Accurate preoperative staging and HER2 status prediction of gastric cancer by the deep learning system based on enhanced computed tomography

“…Different from other studies, most of the deep learning research in the field of gastric cancer focuses on the classification and prognostic analysis of endoscopic images or pathological images (45)(46)(47)(48). Compared with endoscopy and tissue biopsy, enhanced CT is a non-invasive preoperative routine test with few risks (49).…”

Current Developments of Artificial Intelligence in Digital Pathology and Its Future Clinical Applications in Gastrointestinal Cancers

“…By implying such a straightforward assumption, considerable success has been achieved with a variety of techniques, such as recalibrated multi-instance learning [ 127 ] for classification of cancer and dysplasia in GC, graph convolutional network-based multi-instance learning [ 37 ] for identification of the presence of tumours in gastric biopsies, S3TA multi-instance learning method [ 38 ] for prediction of the presence of epithelial cell nuclei in CRC and GastroMIL [ 39 ] for GC diagnosis and risk prediction. Other than GI cancer, multi-instance learning also shows its strong adaptability for cancer tissue classification, such as dual-stream multiple instance learning network [ 40 ] and TransMIL [ 128 ] for recognition of tumours in lung or breast cancer and cluster-to-conquer [ 129 ] for identification of celiac cancer.…”

“…The pressure to provide fast and accurate diagnoses, with overloaded histopathology workforces, has forced the transformation of practice from conventional light microscopy to digital pathology with AI analytical applications. Studies have focused on improving clinical workflow by diagnosing various GI cancers, including oesophageal cancer [ 34 ] ( Figure 3 E–H), GC [ 30 , 39 ] ( Figure 3 C,I), etc. Pertinently, Gehrung et al [ 34 ] established an H&E-based trained deep learning system to identify and triage oesophageal cases with Barrett’s oesophagus (BO) for early detection of oesophageal adenocarcinoma.…”

“…This could create variability within datasets [ 22 , 23 ], thus increasing the difficulty and cost of model training in a hospital setting. Machine learning [ 24 , 25 , 26 , 27 ] and deep learning [ 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 ] are common methods adopted for image analysis and both have unique advantages and disadvantages. Currently, there is a variety of existing models that can be selected and applied easily.…”

Pathological diagnosis and prognosis of Gastric cancer through a multi-instance learning method

“…Endoscopic images based deep learning models for early GC detecting have achieved a performance equal with experienced endoscopists [1] and radiomic methods on CT images have proposed clinically significant imaging biomarkers with diagnostic and prognostic values [2À4]. In this article of Ebiomedicine, Huang et al [5] proposed a simple multiinstance learning (MIL) approach for GC diagnosis and prognosis analysis on whole slide imaging (WSI) pathological images. Experiments on three datasets achieved good performances for both tasks.…”