Preoperative prediction of p53 overexpression in pituitary neuroendocrine tumors using MRI radiomics

The expression of p53 protein is closely related to tumor prognosis and plays an important role in patients with pituitary neuroendocrine tumors (PitNETs). However, its evaluation currently relies on postoperative histopathological analysis. Developing a non-invasive method to predict p53 overexpression preoperatively may help support clinical judgment and facilitate individualized treatment strategies.

Clinical and imaging data from 186 patients with pathologically confirmed PitNETs were retrospectively collected. The cohort was divided into training and testing sets using stratified random sampling. Radiomic features were extracted from MRI sequences, and feature selection was performed using the intraclass correlation coefficient (ICC) and least absolute shrinkage and selection operator (LASSO). A radiomics score was calculated, and univariate and multivariate logistic regression analyses were used to identify independent clinical risk factors. A combined nomogram model incorporating clinical and radiomic features was constructed. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), precision-recall (PR) curve, calibration curve, and decision curve analysis (DCA).

Four radiomic features and two clinical features were selected for model development. Age (odds ratio [OR] = 0.97, 95% confidence interval [CI]: 0.94-0.99, p = 0.01) and suprasellar invasion (OR = 0.47, 95% CI: 0.25-0.89, p = 0.02) were identified as independent predictors of p53 positivity. The combined clinical-radiomic model achieved good predictive performance with an AUC of 0.77 in the validation set, demonstrating favorable discrimination, calibration, and clinical utility.

The proposed MRI-based radiomics model, integrating clinical and imaging features, enables non-invasive preoperative prediction of p53 overexpression in PitNETs. This approach offers a promising tool for individualized risk stratification and personalized treatment planning in neurosurgical practice.