European journal of radiology · 2026
Development and validation of a CT-based deep learning radiomics model for differentiating parathyroid adenoma from atypical parathyroid tumor/parathyroid carcinoma.
СтатьяВалидационное 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.
Wen Shiqi, Huang Meijing, Wang Zilin, Tian Yihan, Wang Peng, Li Fuxin, Li Wei
Аннотация
BACKGROUND: Accurate preoperative differentiation between parathyroid adenoma (PA) and atypical parathyroid tumor (APT) or parathyroid carcinoma (PC) is essential for individualized treatment.
METHODS: This retrospective study included 358 datasets with pathologically confirmed parathyroid tumors who underwent CT and surgery between January 2016 and September 2024. Patients were divided into training (n = 250) and testing (n = 108) cohorts. Radiomic and deep learning features were extracted from CT images. Feature selection was conducted using univariate analysis, minimum redundancy maximum relevance (mRMR), and least absolute shrinkage and selection operator (LASSO). Three classifiers including K-nearest neighbors (KNN), Extra Trees (ET), and Random Forest (RF) were evaluated. A combined model incorporating deep learning features, radiomics features and clinical variables was developed.
RESULTS: KNN showed the best performance. Serum parathyroid hormone (PTH) level was an independent predictor of APT/PC (odds ratio = 0.996, 95% confidence intervals: 0.994-0.998, p < 0.05). The combined model achieved area under the curve (AUC) values of 0.976 (training) and 0.878 (testing), with corresponding accuracies of 93% and 82%, sensitivities of 93% and 87%, and specificities of 94% and 77%. Calibration curves, decision curve analysis, and net reclassification index confirmed superior clinical utility.
CONCLUSION: The combined model demonstrates high diagnostic performance in distinguishing PA from APT/PC and may support personalized surgical decision-making.
METHODS: This retrospective study included 358 datasets with pathologically confirmed parathyroid tumors who underwent CT and surgery between January 2016 and September 2024. Patients were divided into training (n = 250) and testing (n = 108) cohorts. Radiomic and deep learning features were extracted from CT images. Feature selection was conducted using univariate analysis, minimum redundancy maximum relevance (mRMR), and least absolute shrinkage and selection operator (LASSO). Three classifiers including K-nearest neighbors (KNN), Extra Trees (ET), and Random Forest (RF) were evaluated. A combined model incorporating deep learning features, radiomics features and clinical variables was developed.
RESULTS: KNN showed the best performance. Serum parathyroid hormone (PTH) level was an independent predictor of APT/PC (odds ratio = 0.996, 95% confidence intervals: 0.994-0.998, p < 0.05). The combined model achieved area under the curve (AUC) values of 0.976 (training) and 0.878 (testing), with corresponding accuracies of 93% and 82%, sensitivities of 93% and 87%, and specificities of 94% and 77%. Calibration curves, decision curve analysis, and net reclassification index confirmed superior clinical utility.
CONCLUSION: The combined model demonstrates high diagnostic performance in distinguishing PA from APT/PC and may support personalized surgical decision-making.