Purpose: This study aimed to evaluate the diagnostic performance of GPT-4 Omni (GPT-4o) in emergency abdominal computed tomography (CT) cases compared to radiology residents with varying levels of experience, under conditions that closely mimic real clinical scenarios. Material and Methods: A total of 45 emergency cases were categorized into three levels of difficulty (easy, moderate, and difficult) and evaluated by six radiology residents with varying levels of experience (limited: R1–R2; intermediate: R3–R4; advanced: R5–R6) and GPT-4o. Cases were presented sequentially to both groups with consistent clinical data and CT images. Each case included 4 to 7 CT slice images, resulting in a total of 243 images. The participants were asked to provide the single most likely diagnosis for each case. GPT-4o’s CT image interpretation performance without clinical data and hallucination rate were evaluated. Results: Overall diagnostic accuracy rates were 76% for R1–R2, 89% for R3, 82% for R4–R5, 84% for R6, and 82% for GPT-4o. Case difficulty significantly affected the diagnostic accuracy for both the residents and GPT-4o, with accuracy decreasing as case complexity increased (p < 0.001). No statistically significant differences in diagnostic accuracy were found between GPT-4o and the residents, regardless of the experience level or case difficulty (p > 0.05). GPT-4o demonstrated a hallucination rate of 75%. Conclusions: GPT-4o demonstrated a diagnostic accuracy comparable to that of radiology residents in emergency abdominal CT cases. However, its dependence on structured prompts and high hallucination rate indicates the need for further optimization before clinical integration.

Correlation of 3T Diffusion-weighted MRI and 18F-FDG-PET/CT in Liver Metastases: SUV Versus ADC - Molecular Imaging and Radionuclide Therapy

Purpose: Ureteral obstructions are clinically significant conditions that may develop due to benign or malignant causes and threaten renal function. Although the retrograde stenting method is generally preferred as the first-line approach, its success rate markedly decreases in cases with malignant or severe strictures. While antegrade stenting offers an alternative in such cases, it may fail when the guidewire cannot cross the stenosis. Case Presentation: 67 y/o male patient with left ureteral obstruction secondary to bladder cancer, both retrograde and conventional antegrade stenting attempts had failed. Subsequently, we applied a novel technique developed by our team and previously undescribed in the literature, termed the “Hydroguiding” technique. In this approach, a Y-hemostasis valve was connected to the tip of the guiding catheter, allowing simultaneous contrast injection and guidewire manipulation. Under fluoroscopic guidance, the hydrostatic pressure generated by the contrast injection enabled the guidewire to traverse the stenotic segment with minimal trauma and reach the bladder. Conclusion: The Hydroguiding technique may represent a feasible and promising adjunct to facilitate antegrade ureteral stent placement in selected cases with severe ureteral strictures where conventional methods fail. Further clinical studies are required to establish its safety, reproducibility, and long-term clinical effectiveness.

Purpose: Complications developing after interventional procedures performed via the femoral artery are usually limited to a single pseudoaneurysm (PA) sac; however, the coexistence of a multilobular complex PA with arteriovenous fistula (AVF) is an extremely rare clinical entity. In this case report, we present the endovascular management of this complex iatrogenic vascular injury that developed following atrial fibrillation (AF) ablation therapy. Case presentation: The patient, with a 25 year history of AF and medically refractory recurrent palpitations, underwent cardioversion; however, sustained rhythm control was not achieved. AF ablation was performed. Within several days after the procedure, the patient developed groin pain, swelling, hematoma, and an audible thrill, prompting re-evaluation and advanced imaging. Doppler ultrasonography (US) and Computed Tomography angiography (CTA) revealed a large-neck multilobular PA originated from superficial femoral artery (SFA) composed of multiple interconnected lobes, accompanied by a dynamic high-velocity early arteriovenous shunt at the same level. The patient underwent endovascular treatment and after confirmation of a suitable landing zone, a covered stent–graft was successfully deployed to exclude both the PA neck and AVF tract. At six-month follow-up, Doppler US demonstrated sustained stent patency, complete PA resolution, and no residual AVF. Conclusion: This case demonstrates the effectiveness of covered stent–graft repair in the management of rare but clinically significant vascular complications following femoral artery vascular access. The combination of a multilobular PA and AVF presents unique therapeutic challanges; however, endovascular treatment offers a minimally invasive and effective solution. Successful exclusion of both lesions in a single session highlights the value of contemporary endovascular techniques in managing complex vascular injuries.

Introduction: This study aimed to evaluate the diagnostic performance of GPT-4V, a vision-enabled large language model, in classifying liver lesions on MRI according to LI-RADS v2018 criteria. Methods: Seventy contrast-enhanced liver MRI examinations were retrospectively selected, comprising 10 cases from each LI-RADS category. Each case was presented to GPT-4V as a standardised set of seven anonymised axial MRI slices, accompanied only by lesion size. The model was prompted to assign a single LI-RADS category based solely on visual input. The model’s performance was assessed using accuracy metrics, Cohen’s kappa, ROC analysis, and correlation with lesion size. Results: GPT-4V achieved an overall classification accuracy of 37.1%. While the accuracy for LR-5 was high (90%), it was notably poor in LR-3 (0%) and LR-4 (20%). More than half of the lesions were misclassified as LR-5 (54.2%). Binary classification into benign (LR-1 and LR-2) versus malignant (LR-4, LR-5, LR-M, LR-TIV) yielded an accuracy of 84.3%, with an AUC of 0.72, sensitivity of 100%, and specificity of 45%. Cohen’s kappa values were 0.27 for detailed classification and 0.54 for benign–malignant grouping. Lesion size positively correlated with classification accuracy (ρ = 0.26, p = 0.031). The model demonstrated a tendency to favour high-certainty categories, often defaulting to LR-5 when diagnostic ambiguity was present. Conclusions: GPT-4V demonstrated limited performance in detailed LI-RADS classification, with a strong bias toward LR-5 and poor accuracy in intermediate categories. While the model demonstrated relatively better performance in distinguishing benign from malignant lesions, its current form remains inadequate for precise image-based categorisation. Further development with structured visual training and clinically contextual prompting is warranted.

Splenic artery aneurysms (SAAs) are the most common visceral artery aneurysms, often detected incidentally due to the widespread use of cross-sectional imaging. While many remain asymptomatic, rupture carries a high mortality risk. Despite established size-based treatment guidelines, the impact of morphological features, vascular anatomical factors, and comorbidities on SAA growth remains unclear. This study investigates the long-term growth dynamics of SAAs and their associated risk factors. This retrospective cohort study included 49 patients with true SAAs who were assessed between 2014 and 2024. The annual change in aneurysm size was calculated by measuring the difference in diameter between the initial and final computed tomography scans. Aneurysm growth was analysed in relation to morphological characteristics (mural thrombus, wall calcification), vascular anatomical factors (coeliac artery stenosis, splenic artery, superior mesenteric artery, and aortic diameters), and comorbidities (smoking, hypertension, atherosclerosis, etc.). Data were analysed using the Mann-Whitney U test, Kruskal-Wallis test, and chi-square test. A multivariable regression analysis was performed to assess the impact of comorbidities on aneurysm growth, while receiver operating characteristic (ROC) analysis was conducted to determine the predictive threshold value. A p < 0.05 was considered statistically significant. The mean age of the study cohort was 71 ± 16 years, with 43% of the patients being male. Aneurysm growth was observed in 45% of patients, while stability and regression were noted in 18% and 37%, respectively. The median initial maximum aneurysm diameter was 16.4 mm (8.4–36). The median growth rate was 0.55 mm/year, whereas the median regression rate was 0.86 mm/year. Aneurysm growth was significantly associated with smoking (p = 0.0001), mural thrombus presence (p = 0.021), absence of wall calcification (p = 0.041), younger age (p = 0.013), and a larger abdominal aortic diameter (p = 0.023). ROC analysis identified an initial aneurysm diameter of ≥ 17 mm as a predictive threshold for growth. The growth dynamics of splenic artery aneurysms are heterogeneous, with smoking, mural thrombus presence, younger age, and a larger abdominal aortic diameter identified as key factors associated with aneurysm growth. Guidelines should be refined and strengthened with patient-specific follow-up and treatment algorithms based on updated clinical data. Not applicable.