Light Bladder Net: Non-invasive Bladder Cancer Prediction by Weighted Deep Learning Approaches and Graphical Data Transformation.
Gespeichert in:
| Titel: | Light Bladder Net: Non-invasive Bladder Cancer Prediction by Weighted Deep Learning Approaches and Graphical Data Transformation. |
|---|---|
| Autoren: | Tung CH; Program of Medical Informatics and Innovative Applications, Fu Jen Catholic University, New Taipei City, Taiwan, R.O.C., Lin SH; Ph.D. Program in Medical Biotechnology, National Chung Hsing University, Taichung, Taiwan, R.O.C., Chang KP; School of Medicine, China Medical University, Taichung, Taiwan, R.O.C.; Department of Pathology, China Medical University Hospital, Taichung, Taiwan, R.O.C., Xu YW; Institute of Genomics and Bioinformatics, National Chung Hsing University, Taichung, Taiwan, R.O.C., Chuang ML; Department of Pathology, Feng Yuan Hospital, Ministry of Health and Welfare, Taichung, Taiwan, R.O.C., Chu YW; Ph.D. Program in Medical Biotechnology, National Chung Hsing University, Taichung, Taiwan, R.O.C.; ywchu@dragon.nchu.edu.tw.; Institute of Genomics and Bioinformatics, National Chung Hsing University, Taichung, Taiwan, R.O.C.; Biotechnology Center, National Chung Hsing University, Taichung, Taiwan, R.O.C.; Agricultural Biotechnology Center, National Chung Hsing University, Taichung, Taiwan, R.O.C.; Institute of Molecular Biology, National Chung Hsing University, Taichung, Taiwan, R.O.C.; Smart Sustainable New Agriculture Research Center (SMARTer), Taichung, Taiwan, R.O.C. |
| Quelle: | Anticancer research [Anticancer Res] 2025 May; Vol. 45 (5), pp. 1953-1964. |
| Publikationsart: | Evaluation Study; Journal Article; Validation Study |
| Sprache: | English |
| Info zur Zeitschrift: | Publisher: International Institute of Anticancer Research Country of Publication: Greece NLM ID: 8102988 Publication Model: Print Cited Medium: Internet ISSN: 1791-7530 (Electronic) Linking ISSN: 02507005 NLM ISO Abbreviation: Anticancer Res Subsets: MEDLINE |
| Imprint Name(s): | Publication: Attiki, Greece : International Institute of Anticancer Research Original Publication: Athens, Greece : Potamitis Press |
| MeSH-Schlagworte: | Urinary Bladder Neoplasms*/diagnosis , Detection Algorithms* , Urinalysis*/methods , Early Detection of Cancer*/methods , Prediction Methods, Machine*, Datasets as Topic ; Area Under Curve ; Humans ; Male ; Female ; Sensitivity and Specificity ; Predictive Value of Tests ; Deep Learning |
| Abstract: | Background/aim: Bladder cancer (BCa) is associated with high recurrence rates, emphasizing the importance of early and accurate detection. This study aimed to develop a lightweight and fast deep learning model, Light-Bladder-Net (LBN), for non-invasive BCa detection using conventional urine data. Materials and Methods: We improved LBN's generalization by applying data transformations, adding uniform noise, and employing feature selection methods (mRMR, PCA, SVD, t-SNE) to extract key vectors from its fully connected layer. These vectors were integrated into the original dataset, and multiple machine learning models were trained to enhance classification accuracy. Lastly, weighted voting was used to assign importance across these models. Results: Our approach achieved an accuracy of 0.83, a sensitivity of 0.85, a specificity of 0.80, and a precision of 0.81, indicating robust performance in detecting BCa from urine data. Conclusion: This non-invasive diagnostic method offers rapid, cost-effective predictions. A free online tool is available for clinicians and patients to conveniently detect BCa using standard urine samples at http://merlin.nchu.edu.tw/LBN/. (Copyright © 2025 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.) |
| Contributed Indexing: | Keywords: Conventional urine examination; data image transformation; deep feature extraction; non-invasive bladder cancer prediction; weighted voting |
| Entry Date(s): | Date Created: 20250428 Date Completed: 20250519 Latest Revision: 20250519 |
| Update Code: | 20250520 |
| DOI: | 10.21873/anticanres.17572 |
| PMID: | 40295062 |
| Datenbank: | MEDLINE |
Full Text Finder 
Nájsť tento článok vo Web of Science | Abstract: | Background/aim: Bladder cancer (BCa) is associated with high recurrence rates, emphasizing the importance of early and accurate detection. This study aimed to develop a lightweight and fast deep learning model, Light-Bladder-Net (LBN), for non-invasive BCa detection using conventional urine data.<br />Materials and Methods: We improved LBN's generalization by applying data transformations, adding uniform noise, and employing feature selection methods (mRMR, PCA, SVD, t-SNE) to extract key vectors from its fully connected layer. These vectors were integrated into the original dataset, and multiple machine learning models were trained to enhance classification accuracy. Lastly, weighted voting was used to assign importance across these models.<br />Results: Our approach achieved an accuracy of 0.83, a sensitivity of 0.85, a specificity of 0.80, and a precision of 0.81, indicating robust performance in detecting BCa from urine data.<br />Conclusion: This non-invasive diagnostic method offers rapid, cost-effective predictions. A free online tool is available for clinicians and patients to conveniently detect BCa using standard urine samples at http://merlin.nchu.edu.tw/LBN/.<br /> (Copyright © 2025 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.) |
|---|---|
| ISSN: | 1791-7530 |
| DOI: | 10.21873/anticanres.17572 |