Liquid Biopsy Based Bladder Cancer Diagnostic by Machine Learning

Diagnostics, Journal Year: 2025, Volume and Issue: 15(4), P. 492 - 492

Published: Feb. 18, 2025

Background/Objectives: The timely diagnostics of bladder cancer is still a challenge in clinical settings. reliability conventional testing methods does not reach desirable accuracy and sensitivity, it has an invasive nature. present study examines the application machine learning to improve by integrating miRNA expression levels, demographic routine laboratory test results, data. We proposed that merging these datasets would enhance diagnostic accuracy. Methods: This combined molecular biology for liquid biopsy, data, approach acquired data analysis. evaluated urinary exosome combination with patient as well using three models: Random Forest, SVM, XGBoost classifiers. Results: Based solely on SVM model achieved ROC curve area 0.75. Patient analysis' obtained 0.80. Combining both types enhanced performance, resulting F1 score 0.79 0.85. feature importance analysis identified key predictors, including erythrocytes urine, age, several miRNAs. Conclusions: Our findings indicate potential multi-modal diagnosis non-invasive manner.

Language: Английский

A serum three-microRNA panel: promising biomarkers for renal cell carcinoma screening

Journal of Cancer Metastasis and Treatment, Journal Year: 2025, Volume and Issue: unknown

Published: April 27, 2025

Aim: Renal cell carcinoma (RCC) screening is helpful to improve the prognosis of patients. However, existing RCC detection methods are not suitable for large-scale screening. Serum microRNAs (miRNAs) expected be a convenient, economical, and non-invasive tool RCC. This study aimed identify relevant serum miRNAs as diagnostic markers Methods: research included 112 patients with healthy control individuals, carried out in three distinct phases. The objective was diagnosis using quantitative reverse transcription polymerase chain reaction (RT-qPCR). Additionally, bioinformatics analyses were performed predict target genes provide functional annotations. Results: Compared controls, highly expressed miR-221-3p lowly miR-124-3p, let-7b-5p, miR-30a-5p, miR-302d-3p. After multiple rounds combination screening, miR-221-3p, let-7b-5p showed good predictability. panel exhibited 0.838 area under curve (AUC), achieving 75.00% sensitivity 77.68% specificity. Conclusion: Our analysis demonstrates that combining forms non-invasive, remarkably effective indicator renal carcinoma.

Language: Английский