Prognostic value of diffusion kurtosis imaging histogram based nomogram model for cervical cancer / 中华放射肿瘤学杂志

ABSTRACT

Objective:To analyze the prognostic value of nomogram model for cervical cancer based on the imaging features of diffusion kurtosis imaging (DKI) histogram.Methods:The DKI and clinical data of 272 patients with cervical cancer who were admitted to Affiliated Hospital of Guangdong Medical University from March 2015 to February 2022 were collected and retrospectively analyzed. All patients were randomly divided into the training group ( n=190) and validation group ( n=82) at a ratio of 7 vs. 3. The parameters of DKI histogram were obtained by GE AW 4.2 MRI software. The best prognostic imaging features were screened by LASSO regression. The DKI radiomics score was calculated by linear combination. The independent risk factors of prognosis were identified by univariate and multivariate regression analyses, and a nomogram model was constructed. The model discrimination was assessed by the area under the receiver operating characteristic (ROC) curve (AUC). The internal consistency of the model was evaluated by the calibration map. Results:Adenocarcinoma ( HR=2.496, 95% CI=1.312-4.749, P=0.005), DKI score ( HR=24.087, 95% CI=6.062-95.711, P<0.001), depth of invasion ≥ 1/2 muscular layer ( HR=2.277, 95% CI=1.156-4.487, P=0.017) and neutrophil to lymphocyte ratio (NLR) ( HR=1.800, 95% CI=1.313-2.468, P<0.001) were the independent risk factors for prognosis of cervical cancer. The AUC of the nomogram model in the training and validation groups were 0.860 and 0.757, respectively. The calibration curve was well fitted with the 45° diagonal. The prediction results of long-term prognosis of this model were in good agreement with the actual situation. Conclusions:Adenocarcinoma, NLR, DKI score and depth of invasion ≥ 1/2 muscular layer are the independent risk factors for the prognosis of patients with cervical cancer. The constructed nomogram model could reliably predict the 3-year survival rate of patients with cervical cancer.