Orbital floor fractures may cause long-term functional and esthetic impairments. Diplopia due to impaired function of the inferior rectus muscle is frequently an indication for surgical repair, but some cases, such as those where the diagnosis has been delayed or a previous attempt at repair has been made, may not always be amenable to surgical correction. It is advantageous for the surgeon to know whether the proper function of the inferior rectus muscle can be restored for the purposes of surgical planning and prognostication. The authors hypothesized that real-time MRI could be used to characterize the appearance of the inferior rectus muscle in a way that would facilitate future analysis of inferior rectus function in patients with diplopia due to orbital floor fractures. Real-time MRI was performed on 10 volunteer participants with normal ophthalmic function and orbital anatomy to assess inferior rectus appearance during vertical duction testing. ImageJ software was used to measure and record characteristics of the inferior rectus muscle, viewed in a quasi-sagittal plane. The ratios evaluated included inferior rectus muscle length in upgaze versus downgaze (UDR, mean 1.58) as well as inferior rectus muscle length versus distance from inferior rectus origin to inferior rectus inflection point in upgaze (LIR, mean 1.30) and downgaze (mean 1.20). These values were found to be conserved between orbits and individuals. This data offers quantitative insight regarding inferior rectus muscle appearance across the full arc of vertical gaze in healthy individuals. We plan to use this normative baseline dataset as a comparison for future phases of this project, using real-time MRI to evaluate traumatized orbits with diplopia and derangement of the inferior rectus muscle.
29 October 2026 in Journal of Imaging Informatics in Medicine
Late gadolinium enhancement (LGE) on cardiac magnetic resonance (CMR) imaging is a key diagnostic and prognostic biomarker in ischemic cardiomyopathy (ICM) and nonischemic cardiomyopathy (NICM). However, the completeness of LGE reporting in routine CMR impressions remains incompletely characterized. We aimed to evaluate the impact of large language model (LLM)-assisted reporting on the consistency of LGE pattern, distribution, and severity documentation in CMR impressions for ICM and NICM. We analyzed 828 unstructured (1990–2023) and 400 structured (2024–2025) CMR reports to identify documentation of LGE pattern, distribution, and severity. A Llama 3 language model fine-tuned on 27,000 structured CMR reports was used to generate 9650 CMR impressions; LGE attributes were extracted using an LLM judge. LLM reporting behavior was further assessed using quantized low-rank adaptation adapters trained on structured and unstructured impressions. Structured reporting captured LGE pattern more frequently than unstructured reporting overall (85.5% vs. 61.0%, p = 9.60 × 10 −7 ) and for ICM (90.0% vs. 69.4%, p = 0.005). LLM-generated impressions demonstrated more frequent reporting of LGE distribution (75.3%) and severity (70.9%) versus structured and unstructured reports ( p < 0.001). Fine-tuning with an adapter trained on unstructured impressions significantly improved reporting of all three LGE attributes. Structured reporting improved LGE pattern documentation, but gaps in LGE distribution and severity documentation persisted across human report formats. LLM-generated impressions achieved broader attribute coverage for LGE distribution and severity. Fine-tuning on unstructured data improved model performance, suggesting that passive augmentation with data capturing the documentation styles of multiple providers may help to address reporting gaps.