Evaluation of FET PET Radiomics Feature Repeatability in Glioma Patients

Abstract

Amino acid PET using the tracer O-(2-[18F]fluoroethyl)-L-tyrosine (FET) has attracted considerable interest in neurooncology. Furthermore, initial studies suggested the additional diagnostic value of FET PET radiomics in brain tumor patient management. However, the conclusiveness of radiomics models strongly depends on feature generalizability. We here evaluated the repeatability of feature-based FET PET radiomics. A test–retest analysis based on equivalent but statistically independent subsamples of FET PET images was performed in 50 newly diagnosed and histomolecularly characterized glioma patients. A total of 1,302 radiomics features were calculated from semi-automatically segmented tumor volumes-of-interest (VOIs). Furthermore, to investigate the influence of the spatial resolution of PET on repeatability, spherical VOIs of different sizes were positioned in the tumor and healthy brain tissue. Feature repeatability was assessed by calculating the intraclass correlation coefficient (ICC). To further investigate the influence of the isocitrate dehydrogenase (IDH) genotype on feature repeatability, a hierarchical cluster analysis was performed. For tumor VOIs, 73% of first-order features and 71% of features extracted from the gray level co-occurrence matrix showed high repeatability (ICC 95% confidence interval, 0.91–1.00). In the largest spherical tumor VOIs, 67% of features showed high repeatability, significantly decreasing towards smaller VOIs. The IDH genotype did not affect feature repeatability. Based on 297 repeatable features, two clusters were identified separating patients with IDH-wildtype glioma from those with an IDH mutation. Our results suggest that robust features can be obtained from routinely acquired FET PET scans, which are valuable for further standardization of radiomics analyses in neurooncology.

TERT-Promoter Mutational Status in Glioblastoma – Is There an Association With Amino Acid Uptake on Dynamic 18F-FET PET?

“…Moreover, a vast body of literature exists dealing with radiomics, deep learning and machine learning with special emphasis on ( 18 F-FET) PET and hybrid imaging in neurooncology (43)(44)(45)(46)(47)(48)(49), not just for the differentiation of treatmentrelated changes from real progression (44,50,51), but also for the predication of prognostically relevant mutations such as the IDH-mutation (52). Hence, it needs to be evaluated, if further PET-based analyses with the extraction of radiomic features may add value to the conventional image analysis in order to noninvasively identify the TERTp-mutational status.…”

Current Advances and Challenges in Radiomics of Brain Tumors

“…Early diagnosis is the key to appropriate therapies since the strategies for these two tumors are distinct with different local control rates and intervention prognosis: the prior treatment for GBM is maximum-safe surgery resection, following adjuvant radiotherapy and chemotherapy (100), while regarding BM the more effective and less invasive treatment is stereotactic radiosurgery (101). Qian et al assessed high-dimensional radiomics features from T1-WI, T2-WI, and CE-T1 to distinguish GBM from solitary BM (47). In the retrospective study, patient's population, including 242 GBM and 170 solitary BM, was randomly grouped (training cohort: 227, test cohorts: 185).…”

“…Qian et al. assessed high-dimensional radiomics features from T1-WI, T2-WI, and CE-T1 to distinguish GBM from solitary BM ( 47 ). In the retrospective study, patient’s population, including 242 GBM and 170 solitary BM, was randomly grouped (training cohort: 227, test cohorts: 185).…”

Molecular Biology in Treatment Decision Processes—Neuro-Oncology Edition

“…Radiogenomics, at the intersection of these two approaches allows statistical correlations of radiomic features with genetic aberrations obtained from mutational analyses or next-generation sequencing data and has the potential to provide "virtual biopsy" maps [3] and the creation of radiogenomics pipelines can help define tumor biology [2]. Machine learning (ML) (Table 1) techniques, including deep learning, can thus be employed to identify a pattern that can predict and or refine tumor classification from MRI or MRI/PET [51][52][53][54] such as using MRI and deep learning to define 1p19q co-deletion in gliomas [52]. Existing approaches do suffer from lack of standardisation with respect to image acquisition, processing, segmentation, feature extraction, machine learning algorithm and validation and large ground truth data sets (Table 1 and Figure 3) [3,7,15,24,55].…”