Radiological angle assessment of Haglund's deformity: validation on Magnetic Resonance Imaging.

PURPOSE: Haglund's deformity, an abnormality at the postero-superior corner of the calcaneus is a common cause of posterior heel pain. To date numerous measurements of radiological angles related to the calcaneus have been proposed to differentiate between symptomatic and asymptomatic patients with the deformity. Traditionally, these measurements have been assessed on plain radiographs. The aim of this study was to identify measurements which can be applied to Magnetic Resonance Imaging (MRI) studies of the ankle. METHODS: A retrospective cohort analysis of 30 MRI ankle studies from patients with symptomatic Haglund's deformity and 32 normal studies as controls was undertaken. The angle of BRINK, the Achilles angle, Calcaneal pitch, Achilles-plantar fascia angle and soleus calcaneal distance were measured on optimal T2 fat-saturated sagittal slices. RESULTS: There was a statistically significant difference (p < 0.0001) in the angle of BRINK between the Haglund's and control group. The Area-Under-the-Curve (AUC) was 0.7783 in keeping with good discrimination between the two groups. The angle of BRINK measurement is reproducible, with an intra-observer ICC of 0.837 and an inter-observer ICC of 0.824. There was no statistically significant difference between the two groups for the other measurements. In the Haglund's group the Achilles tendon was more likely to attach to the mid 1/3 of the posterior calcaneus as opposed to the superior 1/3 (p = 0.02), calcaneal oedema was more likely to be present (p < 0.001) and non-insertional tendinopathy was more likely to be present (p < 0.001). The presence of a retrocalcaneal bursa is non-specific (p = 0.602). CONCLUSION: The angle of BRINK demonstrates good discrimination between normal and Haglund's cases on MRI studies and may improve patient management by supporting surgical decision-making. Future work should correlate the angle of BRINK to long-term outcomes.

Validity of ChatGPT-generated musculoskeletal images.

OBJECTIVE: In the evolving landscape of medical research and radiology, effective communication of intricate ideas is imperative, with visualizations playing a crucial role. This study explores the transformative potential of ChatGPT4, a powerful Large Language Model (LLM), in automating the creation of schematics and figures for radiology research papers, specifically focusing on its implications for musculoskeletal studies. MATERIALS AND METHODS: Deploying ChatGPT4, the study aimed to assess the model's ability to generate anatomical images of six large joints-shoulder, elbow, wrist, hip, knee, and ankle. Four variations of a text prompt were utilized, to generate a coronal illustration with annotations for each joint. Evaluation parameters included anatomical correctness, correctness of annotations, aesthetic nature of illustrations, usability of figures in research papers, and cost-effectiveness. Four panellists performed the assessment using a 5-point Likert Scale. RESULTS: Overall analysis of the 24 illustrations encompassing the six joints of interest (4 of each) revealed significant limitations in ChatGPT4's performance. The anatomical design ranged from poor to good, all of the illustrations received a below-average rating for annotation, with the majority assessed as poor. All of them ranked below average for usability in research papers. There was good agreement between raters across all domains (ICC = 0.61). CONCLUSION: While LLMs like ChatGPT4 present promising prospects for rapid figure generation, their current capabilities fall short of meeting the rigorous standards demanded by musculoskeletal radiology research. Future developments should focus on iterative refinement processes to enhance the realism of LLM-generated musculoskeletal schematics.

An evaluation of AI generated literature reviews in musculoskeletal radiology.

PURPOSE: The use of artificial intelligence (AI) tools to aid in summarizing information in medicine and research has recently garnered a huge amount of interest. While tools such as ChatGPT produce convincing and naturally sounding output, the answers are sometimes incorrect. Some of these drawbacks, it is hoped, can be avoided by using programmes trained for a more specific scope. In this study we compared the performance of a new AI tool (the-literature.com) to the latest version OpenAI's ChatGPT (GPT-4) in summarizing topics that the authors have significantly contributed to. METHODS: The AI tools were asked to produce a literature review on 7 topics. These were selected based on the research topics that the authors were intimately familiar with and have contributed to through their own publications. The output produced by the AI tools were graded on a 1-5 Likert scale for accuracy, comprehensiveness, and relevance by two fellowship trained consultant radiologists. RESULTS: The-literature.com produced 3 excellent summaries, 3 very poor summaries not relevant to the prompt, and one summary, which was relevant but did not include all relevant papers. All of the summaries produced by GPT-4 were relevant, but fewer relevant papers were identified. The average Likert rating was for the-literature was 2.88 and 3.86 for GPT-4. There was good agreement between the ratings of both radiologists (ICC = 0.883). CONCLUSION: Summaries produced by AI in its current state require careful human validation. GPT-4 on average provides higher quality summaries. Neither tool can reliably identify all relevant publications.

Assessing bone density on MRI: comparison between routine MRI sequences and DEXA scans.

PURPOSE: Magnetic Resonance Imaging (MRI) is frequently utilised to aid in the comprehensive assessment of back pain, while dual-energy x-ray absorptiometry (DEXA) is the gold standard test for the assessment of bone density. Assessing bone density on MRI could reduce costs and avoid exposing patients to ionising radiation. The aim of this paper is to investigate whether the relative signal intensity of vertebral bodies compared to other structures can detect osteoporosis on MRI. METHODS: 100 patients that had undergone both a lumbar spine MRI and a DEXA scan were identified. The T1 and T2 signal intensity of L1-L4 vertebral bodies (VB), cerebro-spinal fluid (CSF), and psoas muscle were measured within a 1-cm2 region of interest (ROI), and the signal intensity ratios were calculated. The ratios were stratified as normal, osteopenic, or osteoporotic based on DEXA T-scores. RESULTS: The T1 VB /T1 CSF ratio was significantly higher in the osteoporotic group than the normal and osteopenic groups (p < 0.001). The T1 VB /T1 CSF ratio had excellent discrimination (AUC = 0.841) for the presence of osteoporosis. The Pearson correlation coefficient between the DEXA T-score and the T1 VB/T1 CSF ratio was -0.474 (p < 0.001). The intra-observer (ICC = 0.910, 95% CI = 0.757-0.966) and inter-observer reliability (ICC = 0.927, 95% CI = 0.824-0.970) were excellent. In our cohort, a T1 VB / T1 CSF ratio of greater than 4 is 66.7% sensitive but 90.0% specific for the presence of osteoporosis. CONCLUSION: A high T1 VB/T1 CSF ratio suggests osteoporosis on MRI. Prospective validation is needed to confirm these findings.