Objective: To investigate the feasibility of the apparent diffusion coefficient (ADC) value in the preoperative non-invasive prediction of Ki-67 expression level in patients with endometrial cancer (EC), and to construct a nomogram prdiction model by combining clinical data.
Methods: We performed the retrospective analysis of 280patients with EC who underwent preoperative magnetic resonance imaging (MRI) diffusion weighted imaging (DWI) and were confirmed by pathology after operation from January 2017 to February 2023 in our hospital. Two independent radiologists measured the ADC values (ADCmax, ADCmean, and ADCmin) of EC on the ADC image by comparing the MRI enhancement and DWI images, respectively. Potential non-linear relationships between ADC values and Ki-67 expression were assessed using a 3-node restricted cubic spline plot. The included patients were randomized into a training set (n = 186) and a validation set (n = 84). Using a combination of Logistic regression and LASSO regression results, from which the four best predictors were identified for the construction of the nomogram. The accuracy and clinical applicability of the nomogram were assessed using the receiver operating characteristic curve (ROC), calibration curve, and decision curve Analysis (DCA).
Results: ADCmax, ADCmean and ADCmin were significantly correlated with Ki-67 expression levels in EC patients(P=0.001). Among them, there was a non-linear correlation between ADCmin and Ki-67 expression (Nonlinear P = 0.019); and between ADCmean and Ki-67 expression (Nonlinear P = 0.048), specifically. The final obtained ADCmax, FIGO, Chemotherapy and grade were combined to construct the nomogram. The area under the curve (AUC) values of ROC for Nomogram, ADCmax, FIGO, Chemotherapy and Grade in the training set were 0.783, 0.718, 0.579, 0.636 and 0.654, respectively. In the validation set, the AUC values for Nomogram, ADCmax, FIGO, Chemotherapy and Grade were 0.820, 0.746, 0.558, 0.542 and 0.738, respectively. In addition, the calibration curves and the DCA curves suggested a better predictive efficacy of the model.
Conclusion: A nomogram prediction model constructed on the basis of ADCmax values combined with clinical data can be used as an effective method to noninvasively assess Ki67 expression in EC patients before surgery.