Clinical radiology
Audio Summaries

Comparative reliability of internal and vendor-derived computed tomography angiography (CTA) measurements in pretranscatheter aortic valve replacement (TAVR) assessment

May 9, 2026

The authors aimed to evaluate the reliability of internal versus vendor-derived computed tomography angiography (CTA) measurements in the preprocedural assessment for transcatheter aortic valve replacement (TAVR). Their findings indicated that while cardiac CTA metrics showed excellent agreement between internal and external measurements, iliofemoral metrics displayed moderate agreement, which was not clinically significant. This suggests that internally generated CTA measurements can be a reliable and efficient alternative for TAVR planning.

Diffusion-weighted imaging and dynamic susceptibility contrast-enhanced perfusion-weighted imaging of the non-enhancing peritumoural region predict overall survival in glioblastoma

May 9, 2026

The authors aimed to determine the predictive value of diffusion-weighted imaging (DWI) and dynamic susceptibility contrast-enhanced perfusion-weighted imaging (DSC-PWI) parameters in the non-enhancing peritumoural region (NEPR) for overall survival in patients with IDH-wildtype glioblastoma. They found that lower relative minimum apparent diffusion coefficient (rADCmin-p) and higher relative maximum cerebral blood volume (rCBVmax-p) in the NEPR were significantly associated with shorter survival, highlighting their potential as important imaging biomarkers for prognosis in this patient population.

Imaging practices in neurofibromatosis type 1 (NF1) across Romanian centers: a national cross-sectional survey

May 6, 2026

This study aimed to evaluate the imaging practices for neurofibromatosis type 1 (NF1) across Romanian centers and identify gaps in standardization compared to international guidelines. The survey revealed significant variability in imaging practices, with limited use of structured reporting and multidisciplinary collaboration, highlighting the need for improved protocols. The findings serve as a baseline for developing standardized national imaging practices in alignment with European guidelines.

Individualised prediction of HER2 status in colorectal cancer: development and validation of a radiomics prediction model using ¹⁸F-fluorodeoxyglucose positron emission tomography/magnetic resonance (¹⁸F-FDG PET/MR)

Apr 30, 2026

The authors aimed to develop and validate a radiomics prediction model using preoperative ^18F-FDG PET/MR imaging to accurately identify HER2 status in colorectal cancer (CRC) patients. By analyzing radiomic features from PET and MR images of 98 patients, they found that a logistic regression model combining multiple imaging features achieved high performance in predicting HER2 status. This study suggests that PET/MR-based radiomics can serve as a noninvasive and interpretable method for preoperative HER2 assessment in CRC.

Ablation-first versus biopsy-first strategy in computed tomography (CT)-guided biopsy-cryoablation for pulmonary nodules: a retrospective cohort study

Apr 28, 2026

This study aimed to compare the safety and diagnostic performance of ablation-first versus biopsy-first strategies in CT-guided biopsy-cryoablation for pulmonary nodules. The findings indicated that while both approaches had similar positive biopsy rates, the ablation-first strategy resulted in significantly lower rates of procedure-related complications and intraoperative bleeding. Thus, the authors suggest that the ablation-first approach may be a safer option for patients undergoing this procedure.

Review of agentic artificial intelligence (AI) in radiology: from current clinical integration to future innovations

Apr 28, 2026

This review aims to synthesize the current literature on agentic artificial intelligence (AI) in radiology, focusing on its clinical integration and potential future innovations. The authors identify five main themes related to agentic AI's role in clinical decision support, workflow orchestration, multimodal image analysis, reporting, and ethical considerations. They conclude that while agentic AI shows promise for enhancing efficiency in radiology, further prospective studies are needed to evaluate its real-world application.

Pictorial review of chronic ketamine abuse on urinary and biliary systems: what a radiologist needs to know

Apr 25, 2026

This pictorial review aims to educate radiologists on the imaging characteristics and pathophysiological effects of chronic ketamine abuse, particularly its impact on the urinary and biliary systems. By presenting various imaging modalities and case studies, the authors highlight the common presentations of uropathy and cholangiopathy associated with ketamine use, emphasizing the importance of timely identification for appropriate patient management.

Subtle ground enhancement a new definition for breast magnetic resonance imaging (MRI) after neoadjuvant chemotherapy: is it a marker for microscopic residual disease?

Apr 23, 2026

This study investigates the diagnostic value of a newly defined MRI pattern called 'subtle ground enhancement' (SGE) in breast cancer patients following neoadjuvant chemotherapy (NACT) and its correlation with histopathological outcomes. The authors found that SGE was present in 27 out of 332 patients, with a significant proportion showing residual invasive tumors, suggesting that SGE may enhance MRI accuracy in detecting microscopic residual disease post-NACT.

Double reading mammograms in the symptomatic setting

Apr 22, 2026

The authors investigate the effectiveness of double reading mammograms compared to single reading in symptomatic breast clinics, specifically assessing its impact on cancer detection rates and arbitration delays. Their findings indicate that double reading, particularly in surveillance cases, enhances cancer detection, resulting in the identification of additional malignancies despite a 12.7-day delay in arbitration. This study suggests that implementing double reading with arbitration could improve diagnostic outcomes in symptomatic settings.

Prompt engineering enables open-source large language models to match proprietary models in diagnostic accuracy for annotation of radiology reports

Apr 20, 2026

This study investigates whether open-source large language models (LLMs) can achieve diagnostic accuracy comparable to proprietary models in annotating trauma radiology reports in a low-resource language. The findings demonstrate that with effective prompt engineering, small open-source LLMs can accurately identify clinical findings, achieving high accuracy rates that rival those of proprietary models, thus providing a viable and privacy-conscious alternative for clinical applications.