Correlation between MDCT-based arterial radiomics score and the resection margin of superior mesenteric vein in patients with pancreatic head cancer
10.3760/cma.j.cn115667-20211016-00182
- VernacularTitle:多排螺旋CT动脉期影像组学分值与胰头癌肠系膜上静脉切缘的关系
- Author:
Yun BIAN
1
;
Mengmeng ZHU
;
Jian ZHOU
;
Na LI
;
Haiyan ZHAO
;
Jianping LU
Author Information
1. 海军军医大学第一附属医院放射诊断科，上海 200433
- Keywords:
Pancreatic neoplasms;
Resection margin;
Tomography, X-ray computed;
Radiomics
- From:
Chinese Journal of Pancreatology
2021;21(6):411-417
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To accurately identify the relationship between the arterial radiomics score (rad-score) and pathologic superior mesenteric vein (SMV) resection margin in patients with pancreatic head cancer.Methods:The clinical data of 181 patients with pathologically confirmed pancreatic head cancer, who underwent multi-slice computed tomography (MDCT) within one month of resection in the First Affiliated Hospital of Naval Medical University between January 2016 and December 2018 were collected. Based on the pathology of SMV resection margin, the patients were divided into SMV negative margin group ( n=127) and SMV positive margin group ( n=54). The clinical, pathological and radiological features were compared between two groups. 3D slicer software was used to draw the region of interest in each layer of the primary CT arterial images for tumor segmentation. Rython package was applied to extract the radiomics features of pancreatic tumors after segmentation and the extracted features were reduced and chosen using the least absolute shrinkage and selection operator (Lasso) logistic regression algorithm. Lasso logistic regression formula was applied to calculated the arterial rad-score. Univariate and multivariate logistic regression models were used to analyze the association between the arterial rad-score and SMV resection margin. ROC was drawn and AUC, sensitivity, specificity and accuracy for diagnosing the SMV resection margin were calculated. The clinical usefulness of arterial rad-score for diagnosing SMV resection margin was determined by decision curve analysis (DCA). Results:There were statistical differences on LVSI and the touching angle of tumor and SMV/portal vein (PV) between SMV negative margin group and SMV positive margin group (all P<0.001). A total of 1 029 arterial radiomics CT features were obtained, and 14-selected arterial phase features associated with SMV resection margin were determined after being reduced by the Lasso logistic regression algorithm. Univariate analysis showed that the arterial radiomics score, LVSI, the touching angle of tumor and SMV/PV were all correlated with SMV resection margin (all P<0.001). Multivariate analyses confirmed that patients with high arterial radiomics score had a 3.63-fold risk of positive resection margin compared with that with low arterial radiomics score, and a higher arterial rad-score was associated with a higher risk of SMV positive resection margin ( P<0.0001). At the cut-off value of -0.711, AUC of the arterial rad-score for diagnosing SMV resection margin was 0.838, and the sensitivity, specificity and accuracy was 77.8%, 75.6% and 76.24%. Decision curve analysis demonstrated that the percentage of the arterial radiomics score for predicting the positive SMV resection margin was >0.02, and the application of the arterial radiomics score could benefit the patients. Conclusions:The arterial rad-score was strongly correlated with SMV resection margin of pancreatic cancer, and can accurately predict SMV resection margin and provide a new tool for preoperative noninvasive evaluation of the SMV resection margin.
