Bioinformatics advances · 2026
DeepBCTPred: deep learning-based prediction of bladder cancer tissues from endoscopic images.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
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.
Nafi Md Muhaiminul Islam, Md Rahat Hossain Khandokar
Аннотация
MOTIVATION: Bladder cancer is one of the most prevalent malignancies worldwide, affecting the tissues of the urinary bladder and posing a significant threat to patient survival and quality of life. Accurate classification of bladder cancer tissue is critical for early diagnosis and patient survival, yet conventional methods suffer from subjective interpretation and human error.
RESULTS: We propose DeepBCTPred, a novel deep learning framework that integrates handcrafted and learned features through a dual-branch architecture combining MobileNetV3 with a Feedforward Neural Network. Our approach incorporates Recursive Feature Elimination (RFE) for feature selection and employs a genetic algorithm-based image generation pipeline for optimal data selection. DeepBCTPred achieved superior performance with 98.74% recall, 99.45% specificity, and 97.96% F1-score on the test dataset, significantly outperforming existing state-of-the-art methods, achieving improvements ranging from 2% to 15% in recall, 1.3%-13.1% in F1-score, and 1.5%-16% in Matthews Correlation Coefficient (MCC). This framework demonstrates strong potential for clinical implementation in bladder cancer diagnosis and may be extensible to other cancer types for enhanced precision medicine applications.
AVAILABILITY AND IMPLEMENTATION: The training, validation, and test scripts are freely available at https://github.com/nafcoder/DeepBCTPred.
RESULTS: We propose DeepBCTPred, a novel deep learning framework that integrates handcrafted and learned features through a dual-branch architecture combining MobileNetV3 with a Feedforward Neural Network. Our approach incorporates Recursive Feature Elimination (RFE) for feature selection and employs a genetic algorithm-based image generation pipeline for optimal data selection. DeepBCTPred achieved superior performance with 98.74% recall, 99.45% specificity, and 97.96% F1-score on the test dataset, significantly outperforming existing state-of-the-art methods, achieving improvements ranging from 2% to 15% in recall, 1.3%-13.1% in F1-score, and 1.5%-16% in Matthews Correlation Coefficient (MCC). This framework demonstrates strong potential for clinical implementation in bladder cancer diagnosis and may be extensible to other cancer types for enhanced precision medicine applications.
AVAILABILITY AND IMPLEMENTATION: The training, validation, and test scripts are freely available at https://github.com/nafcoder/DeepBCTPred.