- Enhancing Lesion Detection in Inflammatory Myelopathies: A Deep Learning–Reconstructed Double Inversion Recovery MRI Approach
This observational study compared the acquisition time, image quality, diagnostic confidence, and lesion detection rates among sagittal T2WI, standard DIR, and DL-reconstructed DIR in patients with inflammatory myelopathies. It was found that DL-reconstructed DIR significantly reduced acquisition time and improved image quality. DL-reconstructed DIR also improved lesion detection, showing superior diagnostic performance in inflammatory myelopathies without compromising diagnostic confidence.
- A Review of Intracranial Aneurysm Imaging Modalities, from CT to State-of-the-Art MR
This review article covers the established imaging modalities (eg, CT, CTA, DSA, FLAIR, 3D TOF-MRA, contrast-enhanced MRA) and novel MR techniques (MR vessel wall imaging, dynamic contrast-enhanced MRI, computational fluid dynamics) of intracranial aneurysm evaluation. Artificial intelligence software and its integration into diagnostic and risk-stratification pipelines for intracranial aneurysms are also discussed.
- A Comparative Study of CT Perfusion Postprocessing Tools in Medium/Distal Vessel Occlusion Stroke
This retrospective single-center cohort study investigated the discrepancy of 2 widely used postprocessing tools (Syngo.via and RapidAI) for CTP in patients with medium/distal vessel occlusion (MDVO) stroke. The 2 postprocessing tools were discordant on the volume and the location of perfusion deficit in isolated MDVO as well as the volume of potentially salvageable penumbra. As such, this questions whether infarct core size or penumbra volume should routinely be used to identify candidates for MDVO thrombectomy.
- Development and Evaluation of Automated Artificial Intelligence–Based Brain Tumor Response Assessment in Patients with Glioblastoma
The goal of this study was to compare AI-based volumetric GBM MRI response assessment with standardized radiologist response assessments. The AI-based volumetric response assessment yielded overall moderate performance for recapitulating most human response assessment categories (BT-RADS 1, 2, and 4) but demonstrated the lowest performance for predicting BT-RADS 3, which is likely related to the high variability of this assessment. In comparison to radiologist assessment, the AI-based volumetric GBM MRI response assessment showed comparable performance for overall survival.
AJNR Awards, New Junior Editors, and more. Read the latest AJNR updates