Clinically applicable histopathological diagnosis system for gastric cancer detection using deep learning

Song, Z.; Zou, S.; Zhou, W.; Huang, Y.; Shao, L.; Yuan, J.; Gou, X.; Jin, W.; Wang, Z.; Chen, X.; Ding, X.; Liu, J.; Yu, C.; Ku, C.; Liu, C.; Sun, Z.; Xu, G.; Wang, Y.; Zhang, X.; Wang, D.; Wang, S.; Xu, W.; Davis, R. C.; Shi, H.

doi:10.1101/2020.01.30.927749

bioRxiv · 10.1101/2020.01.30.927749

Clinically applicable histopathological diagnosis system for gastric cancer detection using deep learning

Song, Z.·Zou, S.·Zhou, W.·Huang, Y.·Shao, L.·Yuan, J.·Gou, X.·Jin, W.·Wang, Z.·Chen, X.·Ding, X.·Liu, J.·Yu, C.·Ku, C.·Liu, C.·Sun, Z.·Xu, G.·Wang, Y.·Zhang, X.·Wang, D.·Wang, S.·Xu, W.·Davis, R. C.·Shi, H.

View original source ↗Export metadata ↓

Abstract

Gastric cancer is among the most malignant tumours with the highest incidence and mortality rates. The early detection and accurate histopathological diagnosis of gastric cancer are essential factors that can help increase the chances of successful treatment. While the worldwide shortage of pathologists has imposed burdens on current histopathology services, it also offers a unique opportunity for the use of artificial intelligence assistance systems to alleviate the workload and increase diagnostic accuracy. To the best of our knowledge, there has not been a clinically applicable histopathological assistance system with high accuracy that can generalize to whole slide images created with diverse digital scanner models from different hospitals. Here, we report a clinically applicable artificial intelligence assistance system developed at the Chinese PLA General Hospital, China, using a deep convolutional neural network trained with 2,123 pixel-level annotated H&E-stained whole slide images. The model achieved a sensitivity near 100% and an average specificity of 80.6% on a real-world test dataset, which included 3,212 whole slide images digitalized with three scanner models. We showed that the system could aid pathologists in improving diagnostic accuracy and preventing misdiagnoses. Moreover, we demonstrated that our system could perform robustly with 1,582 whole slide images from two other medical centres. Our study suggests the feasibility and benefits of using histopathological artificial intelligence assistance systems in routine practice scenarios.Competing Interest StatementXiaoqing Zhang is the founder of Thorough Images. Shuhao Wang is the co-founder and chief technology officer (CTO) of Thorough Images. Calvin Ku, Cancheng Liu, Zhuo Sun, Gang Xu, Yuefeng Wang are algorithm researchers of Thorough Images. All remaining authors have declared no conflicts of interest.View Full Text

Source connections

Published article ↗

Explore related subjects

pathology Preprint

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Song, Z., Zou, S., Zhou, W., Huang, Y., Shao, L., Yuan, J., Gou, X., Jin, W., Wang, Z., Chen, X., Ding, X., Liu, J., Yu, C., Ku, C., Liu, C., Sun, Z., Xu, G., Wang, Y., Zhang, X., Wang, D., Wang, S., Xu, W., Davis, R. C., Shi, H.. 2020-01-31. Clinically applicable histopathological diagnosis system for gastric cancer detection using deep learning. https://doi.org/10.1101/2020.01.30.927749

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

NAE1-Dependent Protein Neddylation Preserves Endothelial Identity and Vascular Integrity

Background: Endothelial dysfunction is a central driver of cardiovascular and inflammatory diseases, yet the post-translational mechanisms that preserve endothelial homeostasis remain incompletely understood. Protein neddylation, the covalent conjugation of a ubiquitin-like modifier, regulates diverse cellular processes, yet its physiological role in the vascular endothelium remains unknown. This study investigated whether protein neddylation is required to preserve endothelial identity and vascular homeostasis. Methods: We generated tamoxifen-inducible endothelial-specific Nae1 knockout mice to inhibit neddylation and combined bulk RNA sequencing, single-cell and single-nucleus transcriptomics, quantitative proteomics, biochemical analyses, and gain- and loss-of-function approaches to define the role of endothelial neddylation in vascular homeostasis and inflammatory injury. Results: Endothelial-specific Nae1 deletion caused rapid mortality associated with vascular leakage, platelet accumulation, inflammation, and multi-organ injury. Multi-omics analyses demonstrated profound loss of endothelial identity, characterized by suppression of core endothelial programs and activation of inflammatory, procoagulant, and pyroptotic pathways. Single-cell analyses revealed progressive endothelial dysfunction culminating in depletion of the endothelial population and remodeling of the vascular niche. Mechanistically, endothelial neddylation deficiency activated gasdermin D (GSDMD)- and gasdermin E (GSDME)-dependent pyroptosis, whereas dual inhibition of GSDMD and GSDME markedly attenuated inflammatory transcriptomic remodeling, vascular injury, hepatocyte death, immune cell infiltration, and platelet accumulation. Translational analyses demonstrated reduced endothelial neddylation in experimental endotoxemia and decreased expression of neddylation pathway components in human atherosclerosis and COVID-19 datasets. Conversely, restoration of endothelial neddylation partially reversed inflammatory endothelial transcriptomic reprogramming in vivo. Conclusions: NAE1-dependent protein neddylation is an essential regulator of endothelial identity and vascular integrity. Loss of endothelial neddylation promotes gasdermin-dependent pyroptosis and thrombo-inflammatory vascular injury, whereas restoration of the neddylation pathway mitigates inflammatory endothelial dysfunction. These findings identify endothelial neddylation as a fundamental mechanism maintaining vascular homeostasis and a potential therapeutic target for cardiovascular and inflammatory diseases.

pathology↗

bioRxiv2020

The NAD Metabolome is Functionally Depressed in Patients Undergoing Liver Transplantation for Alcohol-related Liver Disease

Nicotinamide adenine dinucleotide (NAD+) and related coenzymes play critical roles in liver function. Though hepatic alcohol metabolism depresses NAD+, current understanding of the NAD+ metabolome in alcohol-related liver disease (ArLD) is based on animal models. We used human liver samples to quantify the NAD+ metabolome in ArLD with samples obtained at the time of liver transplantation or resection at University Hospitals Birmingham NHS Foundation Trust (UHB). The severity of steatohepatitis in liver from patients with ArLD was assessed with standard liver function tests (LFT) and histology. NAD-targeted quantitative metabolomic analysis of liver tissue was performed by liquid chromatography-tandem mass spectrometry (LC-MS). Seventy-two human liver specimens were analyzed including 43 with ArLD. The NAD+ metabolome differed significantly between different types of liver disease (two-way ANOVA p = 0.001). ArLD liver tissue showed markedly depressed concentrations of NAD+ (432 M vs. 616 M in NL) and precursor molecules nicotinic acid and nicotinamide riboside. There was a significant overall difference in the NAD+ metabolome between ArLD samples with and without steatohepatitis (two-way ANOVA p = 0.018). After correcting for multiple comparisons, a significant difference for individual components of the metabolome was observed for the concentration of NAD+ (mean 451 M vs. 381 M, p = 0.045). NAD+ concentration was inversely related to serum bilirubin concentration (r2 -0.127, p = 0.04) and positively correlated with myeloperoxidase activity (r2 0.31, p = 0.003). The concentration of NAD+ and its precursor molecules are significantly reduced in ArLD and are associated with disease activity. Conclusion: Liver samples from people with ArLD show depressed NAD+ and precursor levels as well as depressed myeloperoxidase activity.

pathology↗

bioRxiv2020

Deficient ER acetyl-CoA import in acinar cells leads to chronic pancreatitis

Maintaining endoplasmic reticulum (ER) proteostasis is essential for pancreatic acinar cell function. Under conditions of severe ER stress, activation of pathogenic unfolded protein response pathways play a central role in the development and progression of pancreatitis. A key event in this pathogenic response is a loss of the transcription factor spliced XBP1 (XBP1s) and activation of the PERK pathway. Less is known of the consequence of perturbing ER-associated post-translational protein modification during pancreatitis. Here we show that expression of the ER acetyl-CoA transporter AT-1, necessary for ER protein acetylation, lies downstream of XBP1s and is significantly downregulated during the onset of pancreatitis. Genetic deletion of AT-1 in acinar cells of adult pancreas induces chronic ER stress marked by activation of both the XBP1s and PERK pathways, leading to mild/moderate chronic pancreatitis evidenced by accumulation of intracellular trypsin, immune cell infiltration, and fibrosis, but little pancreatic degeneration. Two-day induction of acute on chronic pancreatitis in AT-1 acinar specific knockout mice results in a severe CP phenotype with pronounced pancreatic atrophy. These findings uncover a new layer of complexity of the pathological ER stress response and its impact on pancreatic disease.

pathology↗