A Predictive Nomogram for Atypical Meningioma Based On Preoperative Magnetic Resonance Imaging and Routine Blood Tests.

ABSTRACT

OBJECTIVE: The objective of the study was to establish a 5-year progression-free survival prediction nomogram using preoperative routine blood tests and magnetic resonance imaging to guide postoperative treatment. METHODS: Our study was a retrospective analysis of patients with atypical meningioma admitted into our facility from January 31, 2010, to January 31, 2016. We used single-factor logistic analysis to extract valuable indicators from preoperative blood test results and 3D Slicer software to extract radiomic features from magnetic resonance imaging. The radiomics score was calculated by least absolute shrinkage and selection operator logistic regression analysis. We then combined blood indicators and radiomic signatures to construct a radiomic nomogram image. The performance of the model was evaluated comprehensively using the following three aspects recognition ability, accuracy, and clinical value. RESULTS: Six significant radiological features were selected through least absolute shrinkage and selection operator logistic regression analysis. The radiometric label established by these six features has satisfactory predictive performance. The area under the curve in the training group was 0.885 (95% confidence interval, 0.8037-0.9659), and the area under the curve in the validation set was 0.789 (95% confidence interval, 0.6092-0.9686). We used the combined image tags and preoperative leukocyte and neutrophil count to construct a 5-year progression-free survival prediction nomogram. CONCLUSIONS: The analysis results of the calibration curve and the decision curve show that the nomogram constructed by combining radiomics and preoperative blood tests has a good predictive value for 5-year progression-free survival in atypical meningioma and can provide a reference for selecting postoperative treatment options.