Prediction of MGMT Status for Glioblastoma Patients Using Radiomics Feature Extraction From 18F-DOPA-PET Imaging

Jing Qian, Michael G. Herman, Debra H. Brinkmann, Nadia N. Laack, Bradley J. Kemp, Christopher H. Hunt, Val Lowe, Deanna H. Pafundi

Research output: Contribution to journalArticlepeer-review

3 Scopus citations

Abstract

Purpose: Methylation of the O6-methylguanine methyltransferase (MGMT) gene promoter is associated with improved treatment response and survival in patients with glioblastoma (GB), but the necessary pathologic specimen can be nondiagnostic. In this study, we assessed whether radiomics features from pretreatment 18F-DOPA positron emission tomography (PET) imaging could be used to predict pathologic MGMT status. Methods and Materials: This study included 86 patients with newly diagnosed GB, split into 3 groups (training, validating, and predicting). We performed a radiomics analysis on 18F-DOPA PET images by extracting features from 2 tumor-based contours: a “Gold” contour of all abnormal uptake per expert nuclear medicine physician and a high-grade glioma (HGG) contour based on a tumor-to-normal hemispheric ratio >2.0, representing the most aggressive components. Feature selection was performed by comparing the weighted feature importance and filtering with bivariate analysis. Optimization of model parameters was explored using grid search with selected features. The stability of the model with increasing input features was also investigated for model robustness. The model predictions were then applied by comparing the overall survival probability of the patients with GB and unknown MGMT status versus those with known MGMT status. Results: A radiomics signature was constructed to predict MGMT methylation status. Using features extracted from HGG contour alone with a random forest model, we achieved 80% ± 10% accuracy for 95% confidence level in predicting MGMT status. The prediction accuracy was not improved with the addition of the Gold contour or with more input features. The model was applied to the patients with unknown MGMT methylation status. The prediction results are consistent with what is expected using overall survival as a surrogate. Conclusions: This study suggests that 3 features from radiomics modeling of 18F-DOPA PET imaging can predict MGMT methylation status with reasonable accuracy. These results could provide valuable therapeutic guidance for patients in whom MGMT testing is inconclusive or nondiagnostic.

Original languageEnglish (US)
Pages (from-to)1339-1346
Number of pages8
JournalInternational Journal of Radiation Oncology Biology Physics
Volume108
Issue number5
DOIs
StatePublished - Dec 1 2020

ASJC Scopus subject areas

  • Radiation
  • Oncology
  • Radiology Nuclear Medicine and imaging
  • Cancer Research

Fingerprint Dive into the research topics of 'Prediction of MGMT Status for Glioblastoma Patients Using Radiomics Feature Extraction From <sup>18</sup>F-DOPA-PET Imaging'. Together they form a unique fingerprint.

Cite this