International urology and nephrology · 2025
Liquid biopsy based on multi-targeted capture of urinary tumor DNA combined with machine learning to detect urothelial carcinoma: a multicenter prospective study.
СтатьяМногоцентровое исследованиеНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Tang Chaozhi, Wang Tianlong, Luo Huarong, Xue Jingdong, Zhu Miaojun, Hong Zhe, Ding Fang, Zhang Fengwu, Zhu Yihao, Tan Ruoying, Wu Denglong
Аннотация
PURPOSE: Diagnostics for urothelial carcinoma have low sensitivity, thereby negatively impacting diagnostic outcomes. Herein, we present BiovueUro, a machine learning-based urine DNA biomarker tool for urothelial carcinoma detection. We developed BiovueUro to monitor DNA mutation and methylation in urine samples.
METHODS: The study involved 63 patients with malignant urothelial carcinoma, 13 patients with benign lesions, 115 patients with urinary symptoms, and 106 healthy participants undergoing physical check-ups. Detection was performed using real-time fluorescence quantitative polymerase chain reaction. BiovueUro can detect six single nucleotide polymorphism sites in FGFR3, PIK3CA, and TERT, and aberrant methylation in HIST1H4F, NRN1, and POU4F2. Ten machine-learning algorithms were employed to train the data and classify patients and healthy controls.
RESULTS: The optional model demonstrated high diagnostic efficacy in distinguishing patients with high-grade urothelial carcinoma from non-patients. In the validation set, the receiver operating characteristic curve, sensitivity, specificity, accuracy, and positive and negative predictive values were 96.1% (95% CI 89.3-100), 94.7%, 99.1%, 98.5%, 94.7%, and 99.1%, respectively. For distinguishing patients with early-stage urothelial carcinoma from non-patients, the values in the training set were 91.7% (95% CI 84.2-99.1), 80%, 98.3%, 95.1%, 90.9%, and 95.8%, and those in the validation set were 92.5% (95% CI 86.6-98.4%), 80%, 95.7%, 93%, 80%, and 95.7%, respectively. The study's limitations include a relatively small cohort size, restricted regional diversity and a lack of long-term follow-up data to assess disease progression and treatment outcomes.
CONCLUSIONS: In summary, BiovueUro, a machine learning-enhanced multi-analyte liquid biopsy integrating mutation polymorphisms and methylation profiling, demonstrates superior diagnostic accuracy for urothelial carcinoma compared to conventional urinary tests and methylation-based assays. Its robust sensitivity for early-stage tumors, high specificity for high-grade lesions, and non-invasive nature underscore its potential as a clinically actionable tool for urothelial carcinoma detection. Longer-term outcomes and regional diversity should be evaluated.
METHODS: The study involved 63 patients with malignant urothelial carcinoma, 13 patients with benign lesions, 115 patients with urinary symptoms, and 106 healthy participants undergoing physical check-ups. Detection was performed using real-time fluorescence quantitative polymerase chain reaction. BiovueUro can detect six single nucleotide polymorphism sites in FGFR3, PIK3CA, and TERT, and aberrant methylation in HIST1H4F, NRN1, and POU4F2. Ten machine-learning algorithms were employed to train the data and classify patients and healthy controls.
RESULTS: The optional model demonstrated high diagnostic efficacy in distinguishing patients with high-grade urothelial carcinoma from non-patients. In the validation set, the receiver operating characteristic curve, sensitivity, specificity, accuracy, and positive and negative predictive values were 96.1% (95% CI 89.3-100), 94.7%, 99.1%, 98.5%, 94.7%, and 99.1%, respectively. For distinguishing patients with early-stage urothelial carcinoma from non-patients, the values in the training set were 91.7% (95% CI 84.2-99.1), 80%, 98.3%, 95.1%, 90.9%, and 95.8%, and those in the validation set were 92.5% (95% CI 86.6-98.4%), 80%, 95.7%, 93%, 80%, and 95.7%, respectively. The study's limitations include a relatively small cohort size, restricted regional diversity and a lack of long-term follow-up data to assess disease progression and treatment outcomes.
CONCLUSIONS: In summary, BiovueUro, a machine learning-enhanced multi-analyte liquid biopsy integrating mutation polymorphisms and methylation profiling, demonstrates superior diagnostic accuracy for urothelial carcinoma compared to conventional urinary tests and methylation-based assays. Its robust sensitivity for early-stage tumors, high specificity for high-grade lesions, and non-invasive nature underscore its potential as a clinically actionable tool for urothelial carcinoma detection. Longer-term outcomes and regional diversity should be evaluated.