The value of nomogram analysis in prediction of cerebral spread of hepatic alveolar echinococcosis

Abstract

Background

The prognosis after brain metastasis of alveolar echinococcosis is inferior, but there is currently no effective method to predict brain metastasis.

Purpose

To explore the value of a nomogram constructed based on a CT plain scan and enhanced imaging features combined with clinical indicators in predicting brain metastasis of hepatic alveolar echinococcosis (HAE).

Materials and Methods

The imaging characteristics and clinical indicators of 116 patients diagnosed with HAE in the Affiliated Hospital of Qinghai University from 2015 to 2022 were retrospectively collected. The data were randomly divided into a training set and a validation set according to 7:3, and the difference between the two groups was analyzed. Binary logistic regression analysis was used to obtain independent predictors of brain metastasis in HAE, and a prediction model was constructed based on this and expressed in the form of a nomogram. Receiver operating characteristic (ROC) curve and calibration curve (CRC) were used to evaluate model performance, and decision curve analysis (DCA) was used to assess the clinical value of the predictive model.

Result

A total of 116 HAE patients were included (average age 38.07±15.09 years old, 54 males and 62 females, 81 patients (70 %) in the training set, and 35 patients (30 %) in the validation set). There was no statistically significant difference between CT plain scan and enhanced imaging features combined with clinical indicators between the training set and the validation set (p > 0.05). After statistical analysis, it was found that whether there is invasion of the inferior vena cava, whether there is invasion of the hepatic artery, and whether there is metastasis to other organs are independent predictors of brain metastasis in HAE. A prediction model was built based on these three variables. The area under the ROC curve (AUC), cutoff value, sensitivity, and specificity of the training set and validation set were 0.922 and 0.886, 0.6934 and 0.6643, 75.00 and 84.62, 94.34 and 81.82, respectively. CRC shows good consistency between the predicted probability and the actual value of the sample. DCA showed that the clinical value of the model was high.

Conclusion

The nomogram constructed based on imaging features combined with clinical indicators can effectively predict whether HAE will develop brain metastasis, which is helpful for clinicians to quickly screen out high-risk patients with HAE developing brain metastases, evaluate patient prognosis, and is more conducive to the realization of individualized and precise medical decisions.