European journal of radiology · 2025
MRI based fractal analysis of esophageal squamous cell carcinoma to preoperatively predict lymphnode metastasis.
СтатьяМногоцентровое 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.
Liao Xin-Yi, Su Yan-Xia, Liu Xu-Rui, Liao Wen-Han, Li Jing-Ke, Ou Jing, Zhou Hai-Ying, Zhang Xiao-Ming, Fu Mao-Yong, Chen Tian-Wu
Аннотация
PURPOSE: To develop an MRI fractal-based quantitative model of resectable esophageal squamous cell carcinoma (ESCC) to predict regional lymphnode metastasis (LNM).
METHODS: 132 consecutive ESCC patients undergoing preoperative MRI scans including T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), T1-weighted contrast-enhancement (T1CE) and diffusion-weighted imaging (DWI) at b-values of 200, 500, 800, 1000 and 1200 s/mm2 were prospectively enrolled from two centers, among which 110 from center 1 were randomly stratified into training (n = 88) and internal validation (n = 22) cohorts, and 22 from center 2 formed the external validation cohort (n = 22). Fractal parameters including fractal dimension (FD) and lacunarity (LAC), and gross tumor volume (GTV) were measured. A predictive model combining fractal and volumetric parameters was developed using logistic regression to predict LNM, with postoperative histopathological evaluation serving as the reference standard. Area under the receiver operating characteristic curve (AUC) was used to evaluate prediction performance in cohorts.
RESULTS: In training cohort, lower FD on T1WI, T2WI and DWI (b = 200, 500, 800, 1000, 1200 s/mm2), and higher FD, lacunarity and GTV on T1CE were associated with regional LNM (Estimate: -8.35 to 16.98, all p-values < 0.05). A model by combining GTV, FD, and LAC on T1CE with a cutoff of 0.47 to predict LNM showed superior performance (AUC: 0.895) compared with individual predictors. The performance was validated in internal (AUC: 0.862) and external (AUC: 0.836) validation cohorts.
CONCLUSION: This study developed an MRI based fractal-analysis model to well predict regional LNM secondary to resectable ESCC.
METHODS: 132 consecutive ESCC patients undergoing preoperative MRI scans including T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), T1-weighted contrast-enhancement (T1CE) and diffusion-weighted imaging (DWI) at b-values of 200, 500, 800, 1000 and 1200 s/mm2 were prospectively enrolled from two centers, among which 110 from center 1 were randomly stratified into training (n = 88) and internal validation (n = 22) cohorts, and 22 from center 2 formed the external validation cohort (n = 22). Fractal parameters including fractal dimension (FD) and lacunarity (LAC), and gross tumor volume (GTV) were measured. A predictive model combining fractal and volumetric parameters was developed using logistic regression to predict LNM, with postoperative histopathological evaluation serving as the reference standard. Area under the receiver operating characteristic curve (AUC) was used to evaluate prediction performance in cohorts.
RESULTS: In training cohort, lower FD on T1WI, T2WI and DWI (b = 200, 500, 800, 1000, 1200 s/mm2), and higher FD, lacunarity and GTV on T1CE were associated with regional LNM (Estimate: -8.35 to 16.98, all p-values < 0.05). A model by combining GTV, FD, and LAC on T1CE with a cutoff of 0.47 to predict LNM showed superior performance (AUC: 0.895) compared with individual predictors. The performance was validated in internal (AUC: 0.862) and external (AUC: 0.836) validation cohorts.
CONCLUSION: This study developed an MRI based fractal-analysis model to well predict regional LNM secondary to resectable ESCC.