Value of dual-energy CT quantitative parameters combined with clinical features in diagnosis of stages T2 and T3 colorectal cancer

Objective‍ ‍To investigate the diagnostic value of our regression model based on quantitative parameters of dual-energy CT and clinical features for stages T2 and T3 colorectal cancer. Methods‍ ‍A cross-section study was performed on 91 patients with colorectal cancer confirmed by postoperative pathology in our hospital from January 2022 to November 2023. All of them underwent dual-energy CT examination. According to the pathological T staging criteria of Chinese Colorectal Cancer Diagnosis and Treatment Standard （2020 Edition）， they were divided into T2 group （n=43） and T3 group （n=48）. Univariate analysis was used to compare the differences in quantitative CT parameters and clinical features between the 2 groups， and the obtained significant variables were employed to construct diagnosis models by univariate or multivariate logistic regression analysis. The area under receiver operating characteristic curve （AUC） of the CT parametric model and the model combined with clinical features was compared to evaluate the efficacy of diagnosing T2 and T3 stages. Results‍ ‍Univariate analysis showed that carcinoembryonic antigen （CEA）， N stage， tumor location， tumor longest diameter （LD）， CT value of virtual noncontrast （CT-VNC）， fat fraction， electron density （Rho） and dual energy index （DEI） were significantly different between the T2 and T3 groups （P<0.05）. Multivariate logistic regression analysis found that N stage， tumor location， LD， fat fraction and DEI were independent risk factors for the diagnosis of stage T3. The AUC value of the model of above CT parameters in diagnosing stage T3 colorectal cancer was 0.671 （95%CI： 0.558~0.783）， and the AUC value of the combined model of above CT parameters and clinical features was 0.886 （95%CI： 0.815~0.957）， and statistical difference was observed in the AUC value between the combined model and the CT parametric model （P<0.01）. Conclusion‍ ‍The regression model constructed with dual-energy CT quantitative parameters combined with clinical features has high value in the preoperative diagnosis of stages T2 and T3 colorectal cancer before surgery.