Computer methods and programs in biomedicine · 2025
Predicting clinical prognosis in gastric cancer using deep learning-based analysis of tissue pathomics 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.
Liu Long, Zhang Yu, Zhao Xiao-Hong, Zhu Min, Liang Jia-Wei, Jiang Zhen-Yu, Zhang Jin-Qiu, Ma Jia-Li, Wang Zhen-Zhen, Gan Mei-Fu, Mao Xin-Li, Li Shao-Wei
Аннотация
OBJECTIVE: Evaluate the utility of a machine learning-based pathomics model in predicting overall survival (OS) post-surgery for gastric cancer patients.
METHODS: A retrospective analysis of 160 gastric cancer patients undergoing radical surgery with continuous follow-up was conducted. Pathomic parameters of the primary lesion and surrounding area were assessed. Six machine learning methods were employed to develop an optimal pathomics model and calculate the pathomics score (Pathscore). The Pathscore was integrated with clinical parameters to construct a comprehensive nomogram for accurate prognosis prediction. The TCGA and GEO databases were utilized for model validation and underlying mechanism exploration.
RESULTS: The combined GBM-based pathomics model exhibited superior predictive performance, with 1-, 3-, and 5-year area under the curve (AUC) values of 0.837, 0.970, and 0.963, respectively, in the training set. The pathomics model demonstrated significant incremental value for clinical predictors. The nomogram incorporating age, M stage, TNM stage, and Pathscore achieved even higher AUCs of 0.954, 0.939, and 0.898 for 1-, 3-, and 5-year OS prediction, respectively. Bioinformatics analysis suggested that the pathomics model could predict OS by reflecting tumor immune status and NRP1 expression.
CONCLUSION: Machine learning-based pathomics models can effectively predict OS in gastric cancer patients post-surgery. The integration of pathomics and clinical parameters enhances predictive accuracy compared to clinical factors alone, providing a reliable basis for prognostic prediction and personalized treatment decisions.
METHODS: A retrospective analysis of 160 gastric cancer patients undergoing radical surgery with continuous follow-up was conducted. Pathomic parameters of the primary lesion and surrounding area were assessed. Six machine learning methods were employed to develop an optimal pathomics model and calculate the pathomics score (Pathscore). The Pathscore was integrated with clinical parameters to construct a comprehensive nomogram for accurate prognosis prediction. The TCGA and GEO databases were utilized for model validation and underlying mechanism exploration.
RESULTS: The combined GBM-based pathomics model exhibited superior predictive performance, with 1-, 3-, and 5-year area under the curve (AUC) values of 0.837, 0.970, and 0.963, respectively, in the training set. The pathomics model demonstrated significant incremental value for clinical predictors. The nomogram incorporating age, M stage, TNM stage, and Pathscore achieved even higher AUCs of 0.954, 0.939, and 0.898 for 1-, 3-, and 5-year OS prediction, respectively. Bioinformatics analysis suggested that the pathomics model could predict OS by reflecting tumor immune status and NRP1 expression.
CONCLUSION: Machine learning-based pathomics models can effectively predict OS in gastric cancer patients post-surgery. The integration of pathomics and clinical parameters enhances predictive accuracy compared to clinical factors alone, providing a reliable basis for prognostic prediction and personalized treatment decisions.