Radiomics signature for temporal evolution and recurrence patterns of glioblastoma using multimodal magnetic resonance imaging.

Glioblastoma is a highly infiltrative neoplasm with a high propensity of recurrence. The location of recurrence usually cannot be anticipated and depends on various factors, including the surgical resection margins. Currently, radiation planning utilizes the hyperintense signal from T2-FLAIR MRI and is delivered to a limited area defined by standardized guidelines. To this end, noninvasive early prediction and delineation of recurrence can aid in tailored targeted therapy, which may potentially delay the relapse, consequently improving overall survival. In this work, we hypothesize that radiomics-based phenotypic quantifiers may support the detection of recurrence before it is visualized on multimodal MRI. We employ retrospective longitudinal data from 29 subjects with a varying number of time points (three to 13) that includes glioblastoma recurrence. Voxelwise textural and intensity features are computed from multimodal MRI (T1-contrast enhanced [T1CE], FLAIR, and apparent diffusion coefficient), primarily to gain insights into longitudinal radiomic changes from preoperative MRI to recurrence and subsequently to predict the region of relapse from 143 ± 42 days before recurrence using machine learning. T1CE MRI first-order and gray-level co-occurrence matrix features are crucial in detecting local recurrence, while multimodal gray-level difference matrix and first-order features are highly predictive of the distant relapse, with a voxelwise test accuracy of 80.1% for distant recurrence and 71.4% for local recurrence. In summary, our work exemplifies a step forward in predicting glioblastoma recurrence using radiomics-based phenotypic changes that may potentially serve as MR-based biomarkers for customized therapeutic intervention.

Microstructural abnormalities of substantia nigra in Parkinson's disease: A neuromelanin sensitive MRI atlas based study.

Microstructural changes associated with degeneration of dopaminergic neurons of the substantia nigra pars compacta (SNc) in Parkinson's disease (PD) have been studied using Diffusion Tensor Imaging (DTI). However, these studies show inconsistent results, mainly due to methodological variations in delineation of SNc. To mitigate this, our work aims to construct a probabilistic atlas of SNc based on a 3D Neuromelanin Sensitive MRI (NMS-MRI) sequence and demonstrate its applicability to investigate microstructural changes on a large dataset of PD. Using manual segmentation and deformable registration we created a novel SNc atlas in the MNI space using NMS-MRI sequences of 27 healthy controls (HC). We first quantitatively evaluated this atlas and then employed it to investigate the micro-structural abnormalities in SNc using diffusion MRI from 133 patients with PD and 99 HCs. Our results demonstrated significant increase in diffusivity with no changes in anisotropy. In addition, we also observed an asymmetry of the diffusion metrics with a higher diffusivity and lower anisotropy in the left SNc than the right. Finally, a multivariate classifier based on SNc diffusion features could delineate patients with PD with an average accuracy of 71.7%. Overall, from this work we establish a normative baseline for the SNc region of interest using NMS-MRI while the application on PD data emphasizes on the contribution of diffusivity measures rather than anisotropy of white matter in PD.

A Review of Radiomics and Deep Predictive Modeling in Glioma Characterization.

Recent developments in glioma categorization based on biological genotypes and application of computational machine learning or deep learning based predictive models using multi-modal MRI biomarkers to assess these genotypes provides potential assurance for optimal and personalized treatment plans and efficacy. Artificial intelligence based quantified assessment of glioma using MRI derived hand-crafted or auto-extracted features have become crucial as genomic alterations can be associated with MRI based phenotypes. This survey integrates all the recent work carried out in state-of-the-art radiomics, and Artificial Intelligence based learning solutions related to molecular diagnosis, prognosis, and treatment monitoring with the aim to create a structured resource on radiogenomic analysis of glioma. Challenges such as inter-scanner variability, requirement of benchmark datasets, prospective validations for clinical applicability are discussed with further scope for designing optimal solutions for glioma stratification with immediate recommendations for further diagnostic decisions and personalized treatment plans for glioma patients.

Predictive and discriminative localization of pathology using high resolution class activation maps with CNNs.

PURPOSE: Existing class activation mapping (CAM) techniques extract the feature maps only from a single layer of the convolutional neural net (CNN), generally from the final layer and then interpolate to upsample to the original image resolution to locate the discriminative regions. Consequently these provide a coarse localization that may not be able to capture subtle abnormalities in medical images. To alleviate this, our work proposes a technique called high resolution class activation mapping (HR-CAMs) that can provide enhanced visual explainability to the CNN models. METHODS: HR-CAMs fuse feature maps by training a network using the input from multiple layers of a trained CNN, thus gaining information from every layer that can localize abnormalities with greater details in original image resolution. The technique is validated qualitatively and quantitatively on a simulated dataset of 8,000 images followed by applications on multiple image analysis tasks that include (1) skin lesion classification (ISIC open dataset-25,331 cases) and (2) predicting bone fractures (MURA open dataset-40,561 images) (3) predicting Parkinson's disease (PD) from neuromelanin sensitive MRI (small cohort-80 subjects). RESULTS: We demonstrate that our model creates clinically interpretable subject specific high resolution discriminative localizations when compared to widely used CAMs and Gradient-CAMs. CONCLUSION: HR-CAMs provide finer delineation of abnormalities thus facilitating superior explainability to CNNs as has been demonstrated from its rigorous validation.

Developing a Radiomics Signature for Supratentorial Extra-Ventricular Ependymoma Using Multimodal MR Imaging.

Rationale and Objectives: To build a machine learning-based diagnostic model that can accurately distinguish adult supratentorial extraventricular ependymoma (STEE) from similarly appearing high-grade gliomas (HGG) using quantitative radiomic signatures from a multi-parametric MRI framework. Materials and Methods: We computed radiomic features on the preprocessed and segmented tumor masks from a pre-operative multimodal MRI dataset [contrast-enhanced T1 (T1ce), T2, fluid-attenuated inversion recovery (FLAIR), apparent diffusion coefficient (ADC)] from STEE (n = 15), HGG-Grade IV (HGG-G4) (n = 24), and HGG-Grade III (HGG-G3) (n = 36) patients, followed by an optimum two-stage feature selection and multiclass classification. Performance of multiple classifiers were evaluated on both unimodal and multimodal feature sets and most discriminative radiomic features involved in classification of STEE from HGG subtypes were obtained. Results: Multimodal features demonstrated higher classification performance over unimodal feature set in discriminating STEE and HGG subtypes with an accuracy of 68% on test data and above 80% on cross validation, along with an overall above 90% specificity. Among unimodal feature sets, those extracted from FLAIR demonstrated high classification performance in delineating all three tumor groups. Texture-based radiomic features particularly from FLAIR were most important in discriminating STEE from HGG-G4, whereas first-order features from T2 and ADC consistently ranked higher in differentiating multiple tumor groups. Conclusions: This study illustrates the utility of radiomics-based multimodal MRI framework in accurately discriminating similarly appearing adult STEE from HGG subtypes. Radiomic features from multiple MRI modalities could capture intricate and complementary information for a robust and highly accurate multiclass tumor classification.