Cancer reports (Hoboken, N.J.) · 2026
Ultrasound and Clinicopathological Features-Based Machine Learning Model for Predicting Neoadjuvant Therapy Efficacy in Breast Cancer.
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.
Hao Tongtong, Ji Xiaohui, Zhao Qianying, Wei Mengying
Аннотация
BACKGROUND: Accurate prediction of pathological complete response (pCR) after neoadjuvant therapy (NAT) remains challenging in breast cancer. Conventional imaging modalities, such as ultrasound and magnetic resonance imaging (MRI), have limited accuracy when used alone.
AIMS: To develop and validate a machine learning model integrating ultrasound imaging features and clinicopathological information for non-invasive and individualized prediction of pCR following NAT in breast cancer patients.
METHODS AND RESULTS: This retrospective study included 609 breast cancer patients who underwent NAT. Ultrasound imaging features and clinicopathological variables were collected and analyzed. Data preprocessing was performed using Python and R. The diagnostic performance of ultrasound and MRI for predicting pCR was evaluated as a baseline. Significant predictors were identified through univariate and multivariate analyses. Three machine learning models-Random Forest, Logistic Regression, and Support Vector Machine-were developed and validated. Model performance was assessed using receiver operating characteristic (ROC) curves and decision curve analysis, while SHAP analysis and feature importance rankings were used to evaluate variable contributions. The Random Forest model achieved the best performance, with an AUC of 0.85 and an accuracy of 84.7%, outperforming conventional imaging assessments. Key predictors included early NAT tumor volume reduction ≥ 80%, increased echogenicity, HER2 positivity, and higher tumor-infiltrating lymphocyte levels.
CONCLUSION: The Random Forest model substantially improved prediction of pCR after NAT in breast cancer and may provide a practical, non-invasive tool to support individualized treatment planning, and clinical decision-making.
AIMS: To develop and validate a machine learning model integrating ultrasound imaging features and clinicopathological information for non-invasive and individualized prediction of pCR following NAT in breast cancer patients.
METHODS AND RESULTS: This retrospective study included 609 breast cancer patients who underwent NAT. Ultrasound imaging features and clinicopathological variables were collected and analyzed. Data preprocessing was performed using Python and R. The diagnostic performance of ultrasound and MRI for predicting pCR was evaluated as a baseline. Significant predictors were identified through univariate and multivariate analyses. Three machine learning models-Random Forest, Logistic Regression, and Support Vector Machine-were developed and validated. Model performance was assessed using receiver operating characteristic (ROC) curves and decision curve analysis, while SHAP analysis and feature importance rankings were used to evaluate variable contributions. The Random Forest model achieved the best performance, with an AUC of 0.85 and an accuracy of 84.7%, outperforming conventional imaging assessments. Key predictors included early NAT tumor volume reduction ≥ 80%, increased echogenicity, HER2 positivity, and higher tumor-infiltrating lymphocyte levels.
CONCLUSION: The Random Forest model substantially improved prediction of pCR after NAT in breast cancer and may provide a practical, non-invasive tool to support individualized treatment planning, and clinical decision-making.