Objective:

To investigate the predictive value of quantitative parameters of dual-energy computed tomography (DECT) scan of gastric cancer primary on preoperative lymph node metastasis (LNM).

Methods:

A retrospective analysis of the clinicopathological and imaging data of 203 patients with gastric adenocarcinoma diagnosed by postoperative pathology in Fujian Provincial Hospital

Fujian Provincial Clinical College

Fujian Medical University

from July 2020 to January 2022 was divided into LNM (+) group (130 cases) and LNM (-) group (73 cases) according to the pathological results. Two physicians measured and calculated the normalized iodine concentration in arterial

venous and delayed phase (nIC

A

/nIC

V

/nIC

D

)

and the corresponding energy spectrum curve (40-110 keV) slope k values of each patient’s lesion in a double-blind manner for univariate analysis. The indicators of P＜0.05 were incorporated into the multivariate logistic regression analysis to build a predictive model. The receiver operating characteristic (ROC) curve was plotted to evaluate the diagnostic performance of the prediction model

and the differences between the curves were compared using DeLong’s test.

Results:

The maximum diameter of primary focus

T stage and Lauren’s classification were statistically significant between the two groups (

P

＜0.05); the nIC

V

nIC

D

k

A

k

V

and k

D

of the LNM (+) group were larger than those in the LNM (-) group (

P

＜0.05)

while the nIC

A

was not statistically significant between the two groups. Multiple logistic regression analysis showed that T stage

Lauren’s classification

nIC

V

and k

V

were independent risk factors forLNM in gastric cancer (

P

＜0.05). The area under curve (AUC) corresponding to the preoperative diagnosis of LNM were 0.839 and 0.858 respectively in the prediction model 1 with T stage

Lauren

’s classification and nIC

V

and model 2 with T stage

Lauren’s classification and k

V

and there was no significant difference between the two (

P

＞0.05).

Conclusion:

The prediction model combined with the quantitative parameters of DECT has high diagnostic power for the preoperative LNM status of gastric cancer

and the quantitative parameters of the venous phase have more diagnostic value than the residual phase.