This Week in Radiology — Jul 25, 2026

Welcome to This Week in Radiology. This week we're covering 10 notable papers spanning neuroimaging biomarkers, advanced cardiovascular and abdominal imaging, and metabolic liver disease assessment. Let's dive in.

In the realm of neuroimaging, several studies highlight how advanced post-processing and machine learning can refine tissue characterization and risk stratification. For evaluating diffuse gliomas, partial volume effects from cerebrospinal fluid or edema can bias standard apparent diffusion coefficient measurements and obscure areas of reduced diffusion. Säll and colleagues, publishing in the American Journal of Neuroradiology, evaluated high b-value spherical b-tensor encoding to map fluid-suppressed mean diffusivity in tissue across 60 patients [9]. In both simulations and patient data, this technique successfully visualized areas with low diffusivity better than conventional apparent diffusion coefficient maps, with lower tissue mean diffusivity significantly associated with isocitrate dehydrogenase wildtype glioblastomas [9]. Moving to brain tumors with variable clinical courses, preoperative risk stratification in pleomorphic xanthoastrocytoma has historically been challenging. Li and colleagues, also in the American Journal of Neuroradiology, developed a recurrence prediction model using 27 Visually Accessible Rembrandt Images features in a dual-center cohort of 116 patients [10]. Multivariable Cox regression demonstrated that diffusion restriction, midline crossing, and mixed cystic-solid non-classic morphology were independently associated with recurrence, yielding an optimism-corrected concordance index of 0.79 [10]. For neurodegenerative disorders, differentiating frontotemporal dementia from Alzheimer disease can be clinically difficult due to overlapping presentations. Lee and colleagues, writing in the American Journal of Neuroradiology, developed an XGBoost classifier using automated brain volumetry and asymmetry indices from 758 subjects, with external validation in an 89-subject tertiary cohort [4]. The automated model achieved an area under the curve of 92.94% in external validation, and when integrated into a reader study, experienced readers experienced a significant reduction in reading time while maintaining accuracy, whereas less experienced readers showed improved accuracy without a change in reading time [4].

Turning to cardiovascular imaging, two investigations explore advanced computed tomography applications in coronary artery disease. Tang and colleagues, publishing in the Journal of the American College of Radiology, evaluated whether functional Coronary Artery Disease Reporting and Data System categories integrating computed tomography-derived fractional flow reserve provide superior 5-year prognostic value compared to anatomical categories alone in 1096 stable patients [2]. Functional Coronary Artery Disease Reporting and Data System showed a higher C-index of 0.780 compared to 0.723 for anatomical categories, along with a lower proportion of inappropriate management decisions at 9.9% versus 11.0% [2]. Complementing this focus on coronary computed tomography angiography, Vecsey-Nagy and colleagues, writing in European Radiology, assessed the reproducibility of quantitative plaque characterization between energy-integrating detector and photon-counting detector computed tomography in 38 patients [6]. Using optimized reconstruction pairs, inter-scanner standard deviations were low, and power modeling demonstrated that detecting a 10% change in total plaque volume would require 92 patients per group at 80% power [6].

In abdominal, pelvic, and pediatric imaging, three studies address diagnostic workflows across different organ systems. Pavlides and colleagues, reporting in Nature Medicine, conducted a prospective multicenter study of 357 participants with metabolic dysfunction-associated steatotic liver disease to assess noninvasive imaging and serum biomarkers [1]. For cirrhosis and advanced fibrosis, magnetic resonance elastography and vibration-controlled transient elastography met minimum acceptable performance criteria, with magnetic resonance elastography achieving an area under the curve of 0.91 for both advanced fibrosis and cirrhosis [1]. Conversely, serum biomarkers tended to outperform imaging for identifying at-risk metabolic dysfunction-associated steatohepatitis [1]. In pediatric emergency care, Epstein and colleagues in Radiographics reviewed the implementation and technique of rapid magnetic resonance imaging for pediatric acute abdominopelvic pain [5]. This approach covers the entire abdomen and pelvis without breath-holds, intravenous contrast, or the need to fill the urinary bladder, potentially replacing multiple ultrasound examinations for conditions like appendicitis and ovarian torsion [5]. Meanwhile, in pelvic imaging of adult populations, Beyer and colleagues, publishing in the European Journal of Radiology, evaluated the background prevalence of sacroiliitis-like and degenerative findings on whole-body magnetic resonance imaging in 2063 participants from the SHIP-Trend cohort [8]. Bone marrow edema-like signal was present in 10.9% of participants, predominantly in superior sacral quadrants, emphasizing that small, focal signal abnormalities on non-dedicated protocols require careful morphologic and clinical context rather than defining axial spondyloarthritis prevalence [8].

Finally, addressing systemic disease tracking, Bjornstad and colleagues, writing in Science Translational Medicine, investigated the intrarenal effects of the sodium-glucose cotransporter-2 inhibitor dapagliflozin versus placebo over 16 weeks in 98 youth with type 1 diabetes and hyperfiltration [7]. Single-cell RNA sequencing from sequential kidney biopsies demonstrated coordinated transcriptional shifts, including down-regulated glycolysis and oxidative stress in proximal tubules, reduced profibrotic signaling in endothelial cells, and enhanced cytoskeletal reinforcement in podocytes, with more than 55% of dapagliflozin-responsive genes shifting toward healthy control expression patterns [7].

If you only have time for one paper this week, make it the prospective validation of imaging biomarkers in metabolic dysfunction-associated steatotic liver disease published in Nature Medicine [1]. It provides definitive, multicenter comparative data confirming that elastography techniques like magnetic resonance elastography and vibration-controlled transient elastography excel at staging advanced fibrosis and cirrhosis, directly guiding noninvasive clinical pathways.

Here are the key takeaways from this week in Radiology. Fluid-suppressed mean diffusivity mapping improves the visualization of restricted diffusion in diffuse gliomas by removing free water contamination. In stable coronary artery disease, functional Coronary Artery Disease Reporting and Data System integrating computed tomography-derived fractional flow reserve outperforms anatomical categorization alone for predicting major adverse cardiac events and reducing inappropriate management decisions. Automated brain volumetry combined with machine learning classifiers can significantly reduce reading times for experienced radiologists while assisting less experienced readers in differentiating frontotemporal dementia from Alzheimer disease. Magnetic resonance elastography and vibration-controlled transient elastography demonstrate excellent diagnostic accuracy for staging advanced fibrosis and cirrhosis in metabolic dysfunction-associated steatotic liver disease, whereas serum markers show higher accuracy for at-risk metabolic dysfunction-associated steatohepatitis. Rapid magnetic resonance imaging offers a robust, non-contrast alternative for evaluating pediatric acute abdominopelvic pain without requiring bladder filling.

That's your roundup for This Week in Radiology. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And for audio briefings on your own clinical questions and papers, visit audioscholar dot C C.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.