Dataset Information

Clinical application of machine learning-based pathomics signature of gastric atrophy.

ABSTRACT:

Background

The diagnosis of gastric atrophy is highly subjective, and we aimed to establish a model of gastric atrophy based on pathological features to improve diagnostic consistency.

Methods

We retrospectively collected the HE-stained pathological slides of gastric biopsies and used CellProfiler software for image segmentation and feature extraction of ten representative images for each sample. Subsequently, we employed the Least absolute shrinkage and selection operator (LASSO) to select features and different machine learning (ML) algorithms to construct the diagnostic models for gastric atrophy.

Results

We selected 289 gastric biopsy specimens for training, testing, and external validation. We extracted 464 pathological features and screened ten features by LASSO to establish the diagnostic model for moderate-to-severe atrophy. The range of area under the curve (AUC) for various machine learning algorithms was 0.835-1.000 in the training set, 0.786-0.949 in the testing set, and 0.689-0.818 in the external validation set. LR model had the highest AUC value, with 0.900 (95% CI: 0.852-0.947) in the training set, 0.901 (95% CI: 0.807-0.996) in the testing set, and 0.818 (95% CI: 0.714-0.923) in the external validation set. The atrophy pathological score based on the LR model was associated with endoscopic atrophy grading (Z=-2.478, P=0.013) and gastric cancer (GC) (OR=5.70, 95% CI: 2.63-12.33, P<0.001).

Conclusion

The ML model based on pathological features could improve the diagnostic consistency of gastric atrophy, which is also associated with endoscopic atrophy grading and GC.

SUBMITTER: Lan Y

PROVIDER: S-EPMC10929611 | biostudies-literature | 2024

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Clinical application of machine learning-based pathomics signature of gastric atrophy.

Lan Yadi Y Han Bing B Zhai Tianyu T Xu Qianqian Q Li Zhiwei Z Liu Mingyue M Xue Yining Y Xu Hongwei H

Frontiers in oncology 20240227

<h4>Background</h4>The diagnosis of gastric atrophy is highly subjective, and we aimed to establish a model of gastric atrophy based on pathological features to improve diagnostic consistency.<h4>Methods</h4>We retrospectively collected the HE-stained pathological slides of gastric biopsies and used CellProfiler software for image segmentation and feature extraction of ten representative images for each sample. Subsequently, we employed the Least absolute shrinkage and selection operator (LASSO) ...[more]

PMID: 38476364

Similar Datasets

Project description:Traditional histopathology performed by pathologists through naked eyes is insufficient for accurate survival prediction of bladder cancer (BCa). In addition, how neutrophil to lymphocyte ratio (NLR) could be used for prognosis prediction of BCa patients has not been fully understood. In this study, we collected 508 whole slide images (WSIs) of hematoxylin-eosin strained BCa slices and NLR value from the Shanghai General Hospital and The Cancer Genome Atlas (TCGA), which were further processed for nuclear segmentation. Cross-verified prediction models for predicting clinical prognosis were constructed based on machine learning methods. Six WSIs features were selected for the construction of pathomics-based prognosis model, which could automatically distinguish BCa patients with worse survival outcomes, with hazard ratio value of 2.19 in TCGA cohort (95% confidence interval: 1.63-2.94, p <0.0001) and 3.20 in General cohort (95% confidence interval: 1.75-5.87, p = 0.0014). Patients in TCGA cohort with high NLR exhibited significantly worse clinical survival outcome when compared with patients with low NLR (HR = 2.06, 95% CI: 1.29-3.27, p <0.0001). External validation in General cohort also revealed significantly poor prognosis in BCa patients with high NLR (HR = 3.69, 95% CI: 1.83-7.44 p <0.0001). Univariate and multivariate cox regression analysis proved that both the MLPS and the NLR could act as independent prognostic factor for overall survival of BCa patients. Finally, a novel nomogram based on MLPS and NLR was constructed to improve their clinical practicability, which had excellent agreement with actual observation in 1-, 3- and 5-year overall survival prediction. Decision curve analyses both in the TCGA cohort and General cohort revealed that the novel nomogram acted better than both the tumor grade system in prognosis prediction. Our novel nomogram based on MLPS and NLR could act as an excellent survival predictor and provide a scalable and cost-effective method for clinicians to facilitate individualized therapy. Nevertheless, prospective studies are still needed for further verifications.

Project description:Clear cell renal cell carcinoma (ccRCC) is a highly lethal urinary malignancy with poor overall survival (OS) rates. Integrating computer vision and machine learning in pathomics analysis offers potential for enhancing classification, prognosis, and treatment strategies for ccRCC. This study aims to create a pathomics model to predict OS in ccRCC patients. In this study, data from ccRCC patients in the TCGA database were used as a training set, with clinical data serving as a validation set. Pathological features were extracted from H&E-stained slides using PyRadiomics, and a pathomics model was constructed using the non-negative matrix factorization (NMF) algorithm. The model's predictive performance was assessed through Kaplan-Meier (KM) survival curves and Cox regression analysis. Additionally, differential gene expression, gene ontology (GO) enrichment analysis, immune infiltration, and mutational analysis were conducted to investigate the underlying biological mechanisms. A total of 368 pathomics features were extracted from H&E-stained slides of ccRCC patients, and a pathomics model comprising two subtypes (Cluster 1 and Cluster 2) was successfully constructed using the NMF algorithm. KM survival curves and Cox regression analysis revealed that Cluster 2 was associated with worse OS. A total of 76 differential genes were identified between the two subtypes, primarily involving extracellular matrix organization and structure. Immune-related genes, including CTLA4, CD80, and TIGIT, were highly expressed in Cluster 2, while the VHL and PBRM1 genes, along with mutations in the PI3K-Akt, HIF-1, and MAPK signaling pathways, exhibited mutation rates exceeding 40% in both subtypes. The machine learning-based pathomics model effectively predicts the OS of ccRCC patients and differentiates between subtypes. The critical roles of the immune-related gene CTLA4 and the PI3K-Akt, HIF-1, and MAPK signaling pathways offer new insights for further research on the molecular mechanisms, diagnosis, and treatment strategies for ccRCC.

Project description:BackgroundCTLA4, an immune checkpoint, plays an important role in tumor immunotherapy. The purpose of this study was to develop a pathomics signature to evaluate CTLA4 expression and predict clinical outcomes in clear cell renal cell carcinoma (ccRCC) patients.MethodsA total of 354 patients from the TCGA-KIRC dataset were enrolled in this study. The patients were stratified into two groups based on the level of CTLA4 expression, and overall survival rates were analyzed between groups. Pathological features were identified using machine learning algorithms, and a gradient boosting machine (GBM) was employed to construct the pathomics signatures for predicting prognosis and CTLA4 expression. The predictive performance of the model was subsequently assessed. Enrichment analysis was performed on diferentially expressed genes related to the pathomics score (PS). Additionally, correlations between PS and TMB, as well as immune infiltration profiles associated with different PS values, were explored. In vitro experiments, CTLA4 knockdown was performed to investigate its impact on cell proliferation, migration, invasion, TGF-β signaling pathway, and macrophage polarization.ResultsHigh expression of CTLA4 was associated with an unfavorable prognosis in ccRCC patients. The pathomics signature displayed good performance in the validation set (AUC = 0.737; P < 0.001 in the log-rank test). The PS was positively correlated with CTLA4 expression. We next explored the underlying mechanism and found the associations between the pathomics signature and TGF-β signaling pathways, TMB, and Tregs. Further in vitro experiments demonstrated that CTLA4 knockdown inhibited cell proliferation, migration, invasion, TGF-β expression, and macrophage M2 polarization.ConclusionHigh expression of CTLA4 was found to correlate with poor prognosis in ccRCC patients. The pathomics signature established by our group using machine learning effectively predicted both patient prognosis and CTLA4 expression levels in ccRCC cases.

Dataset Information

Clinical application of machine learning-based pathomics signature of gastric atrophy.

Background

Methods

Results

Conclusion

Publications

Clinical application of machine learning-based pathomics signature of gastric atrophy.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets