AUTHOR=Zhao Zhen , Xiao Dongdong , Nie Chuansheng , Zhang Hao , Jiang Xiaobing , Jecha Ali Rajab , Yan Pengfei , Zhao Hongyang 

TITLE=Development of a Nomogram Based on Preoperative Bi-Parametric MRI and Blood Indices for the Differentiation Between Cystic-Solid Pituitary Adenoma and Craniopharyngioma

JOURNAL=Frontiers in Oncology

VOLUME=Volume 11 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2021.709321

DOI=10.3389/fonc.2021.709321

ISSN=2234-943X

ABSTRACT=Abstract
Background: Given the similarity of the clinical symptoms between cystic-solid pituitary adenoma (CS-PA) and craniopharyngioma (CP), this study aims to establish and validate a nomogram based on preoperative images and blood indices for identifying CS-PA and CP.
Methods: This retrospective study included 201 patients with CS-PA and 71 patients with CP who were pathologically confirmed from January 2012 to December 2020 (training set: n = 182; test set: n = 90). We collected and analyzed the magnetic resonance imaging (MRI) features and blood indices of patients. Radiological features were extracted from the tumor regions on the contrast enhancement T1-weighted (CE-T1) and T2-weighted MRI. The two-sample t-test and principal component analysis (PCA) were used for feature selection, data dimension reduction, and radiomics signature building. Then, multivariate Logistic regression analysis was used to establish a radiomic-clinical model containing radiomic and hematological features, and was presented with a nomogram. The performance of the radiomic-clinical model was evaluated through the calibration, clinical effectiveness and assessed by internal validation.
Results: The radiomics signature, which consisted of 18 features after dimensionality reduction, has been shown the superior discrimination performance in 5 different classification models. The area under the curve (AUC) value of the training set and the test set obtained by the radiomics signature are 0.92 and 0.88 in Logistic regression model, 0.90 and 0.85 in Ridge Classifier, 0.88 and 0.82 in the stochastic gradient descent classifier (SGD Classifier), 0.78 and 0.85 in the linear support vector classification (Linear SVC), 0.93 and 0.86 in the multi-layers perceptron classifier (MLP Classifier) respectively. The predictive factors of the personalized nomogram include radiomic signature, age, WBC count, and FIB. The nomogram showed good discrimination, with an AUC of 0.93 in the training set and 0.90 in the test set, and good calibration. Furthermore, decision curve analysis (DCA) represented the clinical usefulness of the radiomic-clinical nomogram.
Conclusion: A personalized nomogram containing radiomic signature and blood indices was proposed in this study, which can easily and accurately distinguish between CS-PA and CP.