Enhancing predictability of IDH mutation status in glioma patients at initial diagnosis: a comparative analysis of radiomics from MRI, [F]FET PET, and TSPO PET.

in European journal of nuclear medicine and molecular imaging by Lena Kaiser, S Quach, A J Zounek, B Wiestler, A Zatcepin, A Holzgreve, A Bollenbacher, L M Bartos, V C Ruf, G Böning, N Thon, J Herms, M J Riemenschneider, S Stöcklein, M Brendel, R Rupprecht, J C Tonn, P Bartenstein, L von Baumgarten, S Ziegler, N L Albert

TLDR

  • The study looked at how well different types of imaging tests could predict whether a patient has a specific type of brain tumor called a glioma. The study used a combination of two types of imaging tests: positron emission tomography (PET) and magnetic resonance imaging (MRI). The study found that certain features from these imaging tests, such as the tumor-to-background ratio (TBR) and dynamic [F]FET PET, were the best predictors of whether a patient had an IDH mutation. The study also found that age and kurtosis from TBR were important factors to consider. The findings suggest that these predictors can help doctors make more accurate diagnoses and potentially improve patient outcomes.

Abstract

According to the World Health Organization classification for tumors of the central nervous system, mutation status of the isocitrate dehydrogenase (IDH) genes has become a major diagnostic discriminator for gliomas. Therefore, imaging-based prediction of IDH mutation status is of high interest for individual patient management. We compared and evaluated the diagnostic value of radiomics derived from dual positron emission tomography (PET) and magnetic resonance imaging (MRI) data to predict the IDH mutation status non-invasively. Eighty-seven glioma patients at initial diagnosis who underwent PET targeting the translocator protein (TSPO) using [F]GE-180, dynamic amino acid PET using [F]FET, and T1-/T2-weighted MRI scans were examined. In addition to calculating tumor-to-background ratio (TBR) images for all modalities, parametric images quantifying dynamic [F]FET PET information were generated. Radiomic features were extracted from TBR and parametric images. The area under the receiver operating characteristic curve (AUC) was employed to assess the performance of logistic regression (LR) classifiers. To report robust estimates, nested cross-validation with five folds and 50 repeats was applied. TBRfeatures extracted from TSPO-positive volumes had the highest predictive power among TBR images (AUC 0.88, with age as co-factor 0.94). Dynamic [F]FET PET reached a similarly high performance (0.94, with age 0.96). The highest LR coefficients in multimodal analyses included TBRfeatures, parameters from kinetic and early static [F]FET PET images, age, and the features from TBRimages such as the kurtosis (0.97). The findings suggest that incorporating TBRfeatures along with kinetic information from dynamic [F]FET PET, kurtosis from TBR, and age can yield very high predictability of IDH mutation status, thus potentially improving early patient management.

Overview

  • The study aimed to compare and evaluate the diagnostic value of radiomics derived from dual positron emission tomography (PET) and magnetic resonance imaging (MRI) data to predict the IDH mutation status non-invasively in glioma patients at initial diagnosis. The study used [F]GE-180, dynamic amino acid PET using [F]FET, and T1-/T2-weighted MRI scans. Radiomic features were extracted from TBR and parametric images, and the area under the receiver operating characteristic curve (AUC) was employed to assess the performance of logistic regression (LR) classifiers. Nested cross-validation with five folds and 50 repeats was applied to report robust estimates. The primary objective of the study was to determine the predictive power of radiomics features for IDH mutation status in glioma patients.

Comparative Analysis & Findings

  • The study compared the outcomes observed under different experimental conditions or interventions, including TBRfeatures extracted from TSPO-positive volumes, dynamic [F]FET PET, kinetic information from dynamic [F]FET PET, kurtosis from TBR, and age. The findings suggest that incorporating TBRfeatures along with kinetic information from dynamic [F]FET PET, kurtosis from TBR, and age can yield very high predictability of IDH mutation status, potentially improving early patient management.

Implications and Future Directions

  • The study's findings have significant implications for the field of research and clinical practice, as they suggest that radiomics derived from dual PET and MRI data can be used to predict IDH mutation status non-invasively in glioma patients. The study identifies TBRfeatures extracted from TSPO-positive volumes, dynamic [F]FET PET, kinetic information from dynamic [F]FET PET, kurtosis from TBR, and age as key predictors of IDH mutation status. Future research directions could include further validation of these predictors in larger cohorts, exploration of other imaging modalities, and development of clinical decision support tools based on these findings.
Read Full Article