Large Language Models as Tools to Generate Radiology Board-Style Multiple-Choice Questions.

RATIONALE AND OBJECTIVES: To determine the potential of large language models (LLMs) to be used as tools by radiology educators to create radiology board-style multiple choice questions (MCQs), answers, and rationales. METHODS: Two LLMs (Llama 2 and GPT-4) were used to develop 104 MCQs based on the American Board of Radiology exam blueprint. Two board-certified radiologists assessed each MCQ using a 10-point Likert scale across five criteria-clarity, relevance, suitability for a board exam based on level of difficulty, quality of distractors, and adequacy of rationale. For comparison, MCQs from prior American College of Radiology (ACR) Diagnostic Radiology In-Training (DXIT) exams were also assessed using these criteria, with radiologists blinded to the question source. RESULTS: Mean scores (±standard deviation) for clarity, relevance, suitability, quality of distractors, and adequacy of rationale were 8.7 (±1.4), 9.2 (±1.3), 9.0 (±1.2), 8.4 (±1.9), and 7.2 (±2.2), respectively, for Llama 2; 9.9 (±0.4), 9.9 (±0.5), 9.9 (±0.4), 9.8 (±0.5), and 9.9 (±0.3), respectively, for GPT-4; and 9.9 (±0.3), 9.9 (±0.2), 9.9 (±0.2), 9.9 (±0.4), and 9.8 (±0.6), respectively, for ACR DXIT items (p < 0.001 for Llama 2 vs. ACR DXIT across all criteria; no statistically significant difference for GPT-4 vs. ACR DXIT). The accuracy of model-generated answers was 69% for Llama 2 and 100% for GPT-4. CONCLUSION: A state-of-the art LLM such as GPT-4 may be used to develop radiology board-style MCQs and rationales to enhance exam preparation materials and expand exam banks, and may allow radiology educators to further use MCQs as teaching and learning tools.

Automated dynamic motion correction improves repeatability and reproducibility of myocardial blood flow quantification with rubidium-82 PET imaging.

BACKGROUND: Patient motion reduces the accuracy of PET myocardial blood flow (MBF) measurements. This study evaluated the effect of automatic motion correction on test-retest repeatability and inter-observer variability in a clinically relevant population. METHODS: Patients with known or suspected CAD underwent repeat rest 82Rb PET scans within minutes as part of their scheduled rest-stress perfusion study. Two trained observers evaluated the presence of heart motion in each scan. Global LV and per-vessel MBF were computed from the dynamic rest images before and after automatic motion correction. Test-retest and inter-observer variability were assessed using intra-class correlation and Bland-Altman analysis. RESULTS: 140 pairs of test-retest scans were included, with visual motion noted in 18%. Motion correction decreased the global MBF values by 3.5% (0.80 ± 0.24 vs 0.82 ± 0.25 mL⋅min-1⋅g-1; P < 0.001) suggesting that the blood input function was underestimated in cases with patient motion. Test-retest repeatability of global MBF improved by 9.7% (0.25 vs 0.28 mL⋅min-1⋅g-1; P < 0.001) and inter-observer repeatability was improved by 7.1% (0.073 vs 0.079 mL⋅min-1⋅g-1; P = 0.012). There was a marked impact on both test-retest repeatability as well as inter-observer repeatability in the LCX territory, with improvements of 16.5% (0.30 vs 0.36 mL⋅min-1⋅g-1; P < 0.0000) and 18.4% (0.13 vs 0.16 mL⋅min-1⋅g-1; P < 0.001), respectively. CONCLUSION: Automatic motion correction improved test-retest repeatability and reduced differences between observers.

Intense FDG Uptake in the Common Bile Duct Post-ERCP Mimics Acute Infectious Cholangitis.

In patients with obstructive pancreatitis due to choledocholithiasis, endoscopic retrograde cholangiopancreatography (ERCP) is the standard of care. ERCP-induced inflammation or infection of the common bile duct (i.e., cholangitis) is a rare complication with a high mortality rate in severe cases. We report an unusual case of incidental findings of intense FDG uptake in the common bile duct one month post-ERCP without clinical features of acute cholangitis, indicative of inflammation of CBD associated with or exaggerated by ERCP.