International journal of surgery (London, England) · 2025
Research on the auxiliary diagnosis of male neurogenic lower urinary tract dysfunction based on a deep-learning algorithm.
СтатьяНазвание 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.
Liu Xin, Zhong Ping, Xiang Tianyu, Liao Limin
Аннотация
BACKGROUND: The diagnosis of neurogenic lower urinary tract dysfunction (NLUTD) is complicated and often misdiagnosed by inexperienced physicians. Therefore, we used deep-learning algorithm models combined with video urodynamics to aid in the diagnosis of male NLUTD patients and construct an automated diagnosis and treatment process for such patients.
METHODS: The urodynamic data from two cohorts of patients in our center with NLUTD were used: (1) a total of 284 male patients with NLUTD from 2009 to 2019 were used for model training and validation optimization; and (2) a cohort of 100 male patients with NLUTD from 2020 to 2021 were used for model testing. The deep-learning algorithm models were yolov5 and yolov10. Based on pressure-flow traces and cystourethrography images, the qualitative and localization evaluation model of male NLUTD was constructed.
RESULTS: For the qualitative assessment of bladder outlet obstruction and detrusor contractility in male NLUTD patients, the best model was yolov10x, and the average precision (mAP) value of all categories was 0.89. For the localization evaluation of bladder outlet obstruction, the best model was yolov5x, and the average mAP value of all categories was 0.83.
CONCLUSIONS: The yolov10x model can be used for qualitative evaluation of bladder outlet obstruction and detrusor contractility, and the yolov5x model can be used for localization evaluation of bladder outlet obstruction in male NLUTD patients.
METHODS: The urodynamic data from two cohorts of patients in our center with NLUTD were used: (1) a total of 284 male patients with NLUTD from 2009 to 2019 were used for model training and validation optimization; and (2) a cohort of 100 male patients with NLUTD from 2020 to 2021 were used for model testing. The deep-learning algorithm models were yolov5 and yolov10. Based on pressure-flow traces and cystourethrography images, the qualitative and localization evaluation model of male NLUTD was constructed.
RESULTS: For the qualitative assessment of bladder outlet obstruction and detrusor contractility in male NLUTD patients, the best model was yolov10x, and the average precision (mAP) value of all categories was 0.89. For the localization evaluation of bladder outlet obstruction, the best model was yolov5x, and the average mAP value of all categories was 0.83.
CONCLUSIONS: The yolov10x model can be used for qualitative evaluation of bladder outlet obstruction and detrusor contractility, and the yolov5x model can be used for localization evaluation of bladder outlet obstruction in male NLUTD patients.