European journal of radiology · 2026
Spectral CT-based multiparametric model for preoperative lymph node metastasis prediction in colorectal cancer.
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.
Yang Hui, Chen Biyun, Jiang Linling, Yang Yuanzhong, Zheng Shuwen, Jiang Dongping, Fu Jin, Liu Lizhi, Zhuo Shuiqing
Аннотация
OBJECTIVE: Preoperative diagnosis of lymph node metastasis (LNM) is critical for colorectal cancer (CRC) staging and treatment planning. This study aimed to develop and validate a spectral CT-derived multiparametric model to predict regional LNM in CRC.
METHODS: In this retrospective study, 311 patients with surgically confirmed CRC who underwent preoperative spectral CT were included. Eligible participants were chronologically stratified into a primary cohort (n = 249) and an independent temporal test cohort (n = 62), with the primary cohort randomly partitioned into a training subset (n = 186) and an internal validation subset (n = 63). Pathologically matched peritumoral lymph nodes (>3 mm) were analyzed on venous-phase spectral base images. Quantitative parameters included conventional Hounsfield units at 120 kV images (HU120kV), CT values of virtual monoenergetic images at 40 keV and 70 keV (HU40keV, HU70keV), iodine density (ID), effective atomic number (Zeff), and spectral curve slope (K). Four logistic regression models were constructed: Model 1 (serum tumor markers), Model 2 (conventional CT parameters), Model 3 (spectral CT parameters), and Model 4 (spectral CT parameters combined with Node-RADS Score). Model performance was compared using ROC curves, with statistical significance evaluated by DeLong's test.
RESULTS: The spectral CT model demonstrated significantly enhanced diagnostic efficacy, with AUC values of 0.925 (95%CI: 0.877-0.959) in the training cohort. This model significantly outperformed the serum tumor marker (AUC = 0.542, P < 0.0001), conventional CT (AUC = 0.800, P < 0.0001), and Node-RADS Score (AUC = 0.812, P < 0.0001) models. Furthermore, integrating this spectral CT model with the Node-RADS score (Model 4) led to an additional improvement, achieving an AUC of 0.961 in the training cohort. At the optimal threshold, Model 3 achieved 83.33% sensitivity, 90.20% specificity, and 86.56% accuracy.
CONCLUSION: A multiparametric model incorporating quantitative spectral CT parameters, particularly when combined with the Node-RADS score, facilitates accurate preoperative LNM prediction in CRC. This tool may improve staging precision and help optimize treatment strategies.
METHODS: In this retrospective study, 311 patients with surgically confirmed CRC who underwent preoperative spectral CT were included. Eligible participants were chronologically stratified into a primary cohort (n = 249) and an independent temporal test cohort (n = 62), with the primary cohort randomly partitioned into a training subset (n = 186) and an internal validation subset (n = 63). Pathologically matched peritumoral lymph nodes (>3 mm) were analyzed on venous-phase spectral base images. Quantitative parameters included conventional Hounsfield units at 120 kV images (HU120kV), CT values of virtual monoenergetic images at 40 keV and 70 keV (HU40keV, HU70keV), iodine density (ID), effective atomic number (Zeff), and spectral curve slope (K). Four logistic regression models were constructed: Model 1 (serum tumor markers), Model 2 (conventional CT parameters), Model 3 (spectral CT parameters), and Model 4 (spectral CT parameters combined with Node-RADS Score). Model performance was compared using ROC curves, with statistical significance evaluated by DeLong's test.
RESULTS: The spectral CT model demonstrated significantly enhanced diagnostic efficacy, with AUC values of 0.925 (95%CI: 0.877-0.959) in the training cohort. This model significantly outperformed the serum tumor marker (AUC = 0.542, P < 0.0001), conventional CT (AUC = 0.800, P < 0.0001), and Node-RADS Score (AUC = 0.812, P < 0.0001) models. Furthermore, integrating this spectral CT model with the Node-RADS score (Model 4) led to an additional improvement, achieving an AUC of 0.961 in the training cohort. At the optimal threshold, Model 3 achieved 83.33% sensitivity, 90.20% specificity, and 86.56% accuracy.
CONCLUSION: A multiparametric model incorporating quantitative spectral CT parameters, particularly when combined with the Node-RADS score, facilitates accurate preoperative LNM prediction in CRC. This tool may improve staging precision and help optimize treatment strategies.