MRI-Based Radiomics Combined with Deep Learning for Distinguishing IDH-Mutant WHO Grade 4 Astrocytomas from IDH-Wild-Type Glioblastomas

This study aimed to investigate the potential of quantitative radiomic data extracted from conventional MR images in discriminating IDH-mutant grade 4 astrocytomas from IDH-wild-type glioblastomas (GBMs). A cohort of 57 treatment-naïve patients with IDH-mutant grade 4 astrocytomas (n = 23) and IDH-wild-type GBMs (n = 34) underwent anatomical imaging on a 3T MR system with standard parameters. Post-contrast T1-weighted and T2-FLAIR images were co-registered. A semi-automatic segmentation approach was used to generate regions of interest (ROIs) from different tissue components of neoplasms. A total of 1050 radiomic features were extracted from each image. The data were split randomly into training and testing sets. A deep learning-based data augmentation method (CTGAN) was implemented to synthesize 200 datasets from the training sets. A total of 18 classifiers were used to distinguish two genotypes of grade 4 astrocytomas. From generated data using 80% training set, the best discriminatory power was obtained from core tumor regions overlaid on post-contrast T1 using the K-best feature selection algorithm and a Gaussian naïve Bayes classifier (AUC = 0.93, accuracy = 0.92, sensitivity = 1, specificity = 0.86, PR_AUC = 0.92). Similarly, high diagnostic performances were obtained from original and generated data using 50% and 30% training sets. Our findings suggest that conventional MR imaging-based radiomic features combined with machine/deep learning methods may be valuable in discriminating IDH-mutant grade 4 astrocytomas from IDH-wild-type GBMs.

Radiomics for Discriminating Benign and Malignant Salivary Gland Tumors; Which Radiomic Feature Categories and MRI Sequences Should Be Used?

The lack of a consistent MRI radiomic signature, partly due to the multitude of initial feature analyses, limits the widespread clinical application of radiomics for the discrimination of salivary gland tumors (SGTs). This study aimed to identify the optimal radiomics feature category and MRI sequence for characterizing SGTs, which could serve as a step towards obtaining a consensus on a radiomics signature. Preliminary radiomics models were built to discriminate malignant SGTs (n = 34) from benign SGTs (n = 57) on T1-weighted (T1WI), fat-suppressed (FS)-T2WI and contrast-enhanced (CE)-T1WI images using six feature categories. The discrimination performances of these preliminary models were evaluated using 5-fold-cross-validation with 100 repetitions and the area under the receiver operating characteristic curve (AUC). The differences between models’ performances were identified using one-way ANOVA. Results show that the best feature categories were logarithm for T1WI and CE-T1WI and exponential for FS-T2WI, with AUCs of 0.828, 0.754 and 0.819, respectively. These AUCs were higher than the AUCs obtained using all feature categories combined, which were 0.750, 0.707 and 0.774, respectively (p < 0.001). The highest AUC (0.846) was obtained using a combination of T1WI + logarithm and FS-T2WI + exponential features, which reduced the initial features by 94.0% (from 1015 × 3 to 91 × 2). CE-T1WI did not improve performance. Using one feature category rather than all feature categories combined reduced the number of initial features without compromising radiomic performance.

Differentiation between supratentorial pilocytic astrocytoma and extraventricular ependymoma using multiparametric MRI

BACKGROUND

The differentiation of supratentorial pilocytic astrocytomas (STPAs) and supratentorial extraventricular ependymomas (STEEs) is clinically pivotal because of distinct therapeutic management and prognosis, which is sometimes challenging to both neuroradiologists and pathologists.

PURPOSE

To explore and compare the conventional and advanced magnetic resonance imaging (MRI) features between STPA and STEE.

MATERIAL AND METHODS

A total of 23 patients with STPAs and 23 patients with STEEs were reviewed in this study. All patients performed conventional MRI, susceptibility-weighted imaging (SWI), and diffusion-weighted imaging (DWI), and 34 patients (17 with STPAs and 17 with STEEs) examined dynamic susceptibility contrast-enhanced perfusion-weighted imaging (DSC-PWI) in addition. Clinical data, conventional MRI features, minimum relative apparent diffusion coefficient ratio (rADC_min), and maximum relative cerebral blood volume ratio (rCBV_max) were compared between the two groups and subgroups. The optimal cutoff values of rADC_min and rCBV_max with sensitivity and specificity were calculated.

RESULTS

STPA manifested similar to STEE as a solid-cystic mass but more frequently presented with a marked enhancing deep nodule (P = 0.02), no peritumoral edema (P = 0.036), higher rADC_min value (2.0 ± 0.5 vs. 0.9 ± 0.2; P < 0.001), and lower rCBV_max value (2.1 ± 0.4 vs. 14.4 ± 5.5; P < 0.001). The cutoff value of >1.39 for rADC_min and ≤ 2.81 for rCBV_max produced a high sensitivity of 95.65% and 100.0%, respectively, and all produced a specificity of 100.0% in differentiating STPAs from STEEs.

CONCLUSION

Multiparametric MRI techniques including conventional MRI, DWI, and DSC-PWI contribute to the differential diagnosis of STPA and STEE.

IDH genotypes differentiation in glioblastomas using DWI and DSC-PWI in the enhancing and peri-enhancing region

BACKGROUND

Isocitrate dehydrogenase (IDH) mutation has diagnostic and prognostic values in glioblastomas. Peritumoral invasion of glioma cells is a cardinal feature of glioblastomas.

PURPOSE

To evaluate the contribution of DWI and DSC-PWI in the enhancing and peri-enhancing region for discriminating glioblastomas IDH genotypes, and the diagnostic values of combining two techniques in the peri-enhancing region compared with those in the enhancing region.

MATERIAL AND METHODS

We retrospectively reviewed the conventional MRI (cMRI), DWI and DSC-PWI obtained from 10 patients with IDH-mutated (IDH-m) glioblastomas and 65 patients with IDH wild-type (IDH-w) glioblastomas. Features of cMRI, relative minimum ADC in the enhancing region (rADC_min-t) and peri-enhancing area (rADC_min-p), and relative maximum CBV values in the enhancing region (rCBV_max-t) and peri-enhancing region (rCBV_max-p) were compared between two groups. Receiver operating characteristic curves and logistic regression analysis were used to assess diagnostic performance.

RESULTS

IDH-m glioblastomas tended to present in frontal lobes and younger patients. The rADC_min-t (P = 0.042) were significantly lower in IDH-w than IDH-m. Both rCBV_max-t and rCBV_max-p showed significant differences between two subgroups (all P < 0.001). The optimal cutoff values in prediction of IDH-m were >0.98 for rADC_min-t, <7.27 for rCBV_max-t, and < 0.97 for rCBV_max-p. Multivariate logistic regression revealed that the combination of rADC_min-t and rCBV_max-t yielded the highest sensitivity and specificity.

CONCLUSION

The rCBV_max-t or rCBV_max-p may serve as preferable and comparable imaging biomarkers for evaluation of glioblastomas IDH status. The combination of rADC_min-t and rCBV_max-t may yield the maximum predictive power for differentiating IDH status.

Magnetic resonance imaging markers for early diagnosis of Parkinson's disease

Parkinson's disease (PD) is a neurodegenerative disorder characterized by selective and progressive degeneration, as well as loss of dopaminergic neurons in the substantia nigra. In PD, approximately 60-70% of nigrostriatal neurons are degenerated and 80% of content of the striatal dopamine is reduced before the diagnosis can be established according to widely accepted clinical diagnostic criteria. This condition describes a stage of disease called “prodromal”, where non-motor symptoms, such as olfactory dysfunction, constipation, rapid eye movement behaviour disorder, depression, precede motor sign of PD. Detection of prodromal phase of PD is becoming an important goal for determining the prognosis and choosing a suitable treatment strategy. In this review, we present some non-invasive instrumental approaches that could be useful to identify patients in the prodromal phase of PD or in an early clinical phase, when the first motor symptoms begin to be apparent. Conventional magnetic resonance imaging (MRI) and advanced MRI techniques, such as magnetic resonance spectroscopy imaging, diffusion-weighted and diffusion tensor imaging and functional MRI, are useful to differentiate early PD with initial motor symptoms from atypical parkinsonian disorders, thus, making easier early diagnosis. Functional MRI and diffusion tensor imaging techniques can show abnormalities in the olfactory system in prodromal PD.