1. Expression of Salivary miRNAs, Clinical, and Demographic Features in the Early Detection of Gastric Cancer: A Statistical and Machine Learning Analysis.
- Author
-
Koopaie M, Arian-Kia S, Manifar S, Fatahzadeh M, Kolahdooz S, and Davoudi M
- Subjects
- Humans, Female, Male, Middle Aged, Case-Control Studies, ROC Curve, Aged, Adult, Gene Expression Regulation, Neoplastic, Stomach Neoplasms diagnosis, Stomach Neoplasms genetics, MicroRNAs analysis, MicroRNAs genetics, Saliva chemistry, Saliva metabolism, Machine Learning, Early Detection of Cancer methods, Biomarkers, Tumor analysis, Biomarkers, Tumor genetics, Biomarkers, Tumor metabolism
- Abstract
Objective: Gastric cancer ranks as one of the top five deadliest cancers worldwide and is often diagnosed at late stages. Analysis of saliva may provide a non-invasive approach for detection of malignancies in organs associated with the oral cavity. This research aims to analyze salivary microRNA expression together with clinical and demographic features with the aim of diagnosing gastric cancer., Materials: The study included 19 patients with early-stage gastric cancer and 19 healthy controls. Saliva samples were collected and processed for RNA isolation. Salivary expression of miR-223-3p and miR-21-5p were measured using quantitative reverse-transcription polymerase chain reaction (RT-qPCR). Receiver operating characteristic (ROC) curves were generated to evaluate the accuracy of diagnostic models. Machine learning algorithms, multiple logistic regression, and principal component analysis (PCA) were used to assess the predictive power of miRNAs in conjunction with clinical-demographic features., Results: Significant upregulation of miR-223-3p and downregulation of miR-21-5p in saliva were observed in patients with gastric cancer. The area under ROC curve (AUC) values for salivary miR-21-5p, salivary miR-223-3p, and their multiple logistic regression were determined to be 0.723, 0.791, and 0.850, respectively. The AUC for multiple logistic regression model was 0.919. The PCA model led to the highest diagnostic odds ratio (DOR) of 134.33 (sensitivity = 0.785, specificity = 1.00, AUC = 903). Application of machine learning methods, and in particular a random forest algorithm, showed high accuracy in diagnosing patients with gastric cancer (sensitivity = 1.00, specificity = 0.857, AUC = 0.93)., Conclusion: The application of validated salivary diagnostics in clinical practice could help facilitate earlier diagnosis of gastric cancer and improve medical outcome. Expression of miR-21 and miR-223-3p in saliva together with clinical and demographic features, appears promising in screening for GC., (© 2024. The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature.)
- Published
- 2024
- Full Text
- View/download PDF