AUTHOR=Chen Hongyu , Lin Fuhua , Zhang Jinming , Lv Xiaofei , Zhou Jian , Li Zhi-Cheng , Chen Yinsheng 

TITLE=Deep Learning Radiomics to Predict PTEN Mutation Status From Magnetic Resonance Imaging in Patients With Glioma

JOURNAL=Frontiers in Oncology

VOLUME=11

YEAR=2021

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2021.734433

DOI=10.3389/fonc.2021.734433

ISSN=2234-943X

ABSTRACT=<sec><title>Objectives</title><p>Phosphatase and tensin homolog (<italic>PTEN</italic>) mutation is an indicator of poor prognosis of low-grade and high-grade glioma. This study built a reliable model from multi-parametric magnetic resonance imaging (MRI) for predicting the <italic>PTEN</italic> mutation status in patients with glioma.</p></sec><sec><title>Methods</title><p>In this study, a total of 244 patients with glioma were retrospectively collected from our center (<italic>n</italic> = 77) and The Cancer Imaging Archive (<italic>n</italic> = 167). All patients were randomly divided into a training set (<italic>n</italic> = 170) and a validation set (<italic>n</italic> = 74). Three models were built from preoperative MRI for predicting PTEN status, including a radiomics model, a convolutional neural network (CNN) model, and an integrated model based on both radiomics and CNN features. The performance of each model was evaluated by accuracy and the area under the receiver operating characteristic curve (AUC).</p></sec><sec><title>Results</title><p>The CNN model achieved an AUC of 0.84 and an accuracy of 0.81, which performed better than did the radiomics model, with an AUC of 0.83 and an accuracy of 0.66. Combining radiomics with CNN will further benefit the predictive performance (accuracy = 0.86, AUC = 0.91).</p></sec><sec><title>Conclusions</title><p>The combination of both the CNN and radiomics features achieved significantly higher performance in predicting the mutation status of <italic>PTEN</italic> in patients with glioma than did the radiomics or the CNN model alone.</p></sec>