Value of spectral CT based iodine concentration for the preoperative prediction of vascular invasion in gastric cancer
10.3760/cma.j.cn112149-20211206-01078
- VernacularTitle:能谱CT碘基值术前预测胃癌脉管侵犯的价值
- Author:
Jing LI
1
;
Yi WANG
;
Yingying WEI
;
Guanglong CHEN
;
Qiuxia LIAO
;
Jianbo GAO
;
Jinrong QU
Author Information
1. 郑州大学附属肿瘤医院 河南省肿瘤医院医学影像科，郑州 450008
- Keywords:
Stomach neoplasms;
Tomography, X-ray computed;
Spectral imaging;
Iodine concentration;
Lymphovascular invasion
- From:
Chinese Journal of Radiology
2023;57(2):181-186
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To investigate the value of spectral CT based iodine concentration (IC) parameters for preoperative prediction of lymphovascular invasion (LVI) in gastric cancer.Methods:Between January 2021 and November 2021, 266 patients diagnosed as gastric adenocarcinomas by endoscopy and undergoing gastrectomy at the Affiliated Cancer Hospital of Zhengzhou University were recruited prospectively. They were divided into LVI and non-LVI groups according to pathological reports. Triple phase contrasted enhanced CT scans, including arterial phase (AP), venous phase (VP) and delayed phase (DP) were performed on a spectral CT platform within one week before surgery. The IC of gastric cancer lesions at three enhanced phases were measured based on iodine maps, and the normalized IC (nIC) was calculated. The thickness of the tumor was measured. Clinicopathological features were collected, including ulceration, pathological tumor staging (pT), pathological node staging (pN), histodifferentiation, Lauren subtype, perineural invasion (PNI), positive node numbers and positive node ratio. Student′s t tes t or Mann-Whitney U test were used to compare the differences of continuous variables between the two groups, while Chi-square test or Fisher′s exact test was used for categorical data. Multivariable logistic regression analysis was used to screen independent risk factors of LVI, and to build a combined parameter based on risk factors. The receiver operating characteristic curve analysis was performed to determine the predictive efficacy of IC parameters and the combined parameter for LVI. DeLong′s test was used to compare the differences among different area under the curve (AUC). Results:There were statistical differences in tumor thickness, ulceration, pT, pN, histodifferentiation, positive node numbers, positive node ratio, Lauren subtype and PNI between LVI and non-LVI groups ( P<0.05). The values of IC VP, IC DP, nIC VP, nIC DP in LVI group were statistically higher than those in non-LVI group ( t=3.77, 4.23, 4.25, 6.12, all P<0.001), with the AUC (95%CI) of 0.674 (0.610-0.738), 0.677 (0.614-0.741), 0.731 (0.671-0.792), 0.700 (0.636-0.764) for predicting LVI, respectively. Multivariable logistic regression analysis revealed that tumor thickness (OR=1.148, 95%CI 1.085-1.237, P<0.001) and nIC VP (OR=209.904, 95%CI 14.874-644.362, P<0.001) were independent predictors for LVI, the combined parameter incorporating these two factors yielded an AUC (95%CI) of 0.790 (0.736-0.937), which was statistically higher than any single parameter of IC VP, IC DP, nIC VP and nIC DP ( Z=3.07, 3.29, 2.10, 2.60, P=0.002, 0.001, 0.036, 0.009). Conclusion:The IC and nIC values of gastric cancer lesions derived from the VP and DP on spectral CT can effectively predict LVI status in gastric adenocarcinomas, and the combination of nIC VP and tumor thickness can further improve the predictive efficacy.
