A Comparative Study of Large Language Models, Human Experts, and Expert-Edited Large Language Models to Neuro-Ophthalmology Questions.

BACKGROUND: While large language models (LLMs) are increasingly used in medicine, their effectiveness compared with human experts remains unclear. This study evaluates the quality and empathy of Expert + AI, human experts, and LLM responses in neuro-ophthalmology. METHODS: This randomized, masked, multicenter cross-sectional study was conducted from June to July 2023. We randomly assigned 21 neuro-ophthalmology questions to 13 experts. Each expert provided an answer and then edited a ChatGPT-4-generated response, timing both tasks. In addition, 5 LLMs (ChatGPT-3.5, ChatGPT-4, Claude 2, Bing, Bard) generated responses. Anonymized and randomized responses from Expert + AI, human experts, and LLMs were evaluated by the remaining 12 experts. The main outcome was the mean score for quality and empathy, rated on a 1-5 scale. RESULTS: Significant differences existed between response types for both quality and empathy (P < 0.0001, P < 0.0001). For quality, Expert + AI (4.16 ± 0.81) performed the best, followed by GPT-4 (4.04 ± 0.92), GPT-3.5 (3.99 ± 0.87), Claude (3.6 ± 1.09), Expert (3.56 ± 1.01), Bard (3.5 ± 1.15), and Bing (3.04 ± 1.12). For empathy, Expert + AI (3.63 ± 0.87) had the highest score, followed by GPT-4 (3.6 ± 0.88), Bard (3.54 ± 0.89), GPT-3.5 (3.5 ± 0.83), Bing (3.27 ± 1.03), Expert (3.26 ± 1.08), and Claude (3.11 ± 0.78). For quality (P < 0.0001) and empathy (P = 0.002), Expert + AI performed better than Expert. Time taken for expert-created and expert-edited LLM responses was similar (P = 0.75). CONCLUSIONS: Expert-edited LLM responses had the highest expert-determined ratings of quality and empathy warranting further exploration of their potential benefits in clinical settings.

Ocular Neuromyotonia: Clinical Features, Diagnosis, and Outcomes.

PURPOSE: The purpose of this study was to describe the clinical features, management, outcomes, and diagnostic pitfalls in a large series of patients with ocular neuromyotonia. DESIGN: Retrospective cohort. METHODS: Patients diagnosed with ocular neuromyotonia from January 1, 2004, through January 1, 2023, seen at one of the 3 Mayo Clinic sites in Rochester, MN, Scottsdale, AZ, and Jacksonville, FL, comprised the study population. We ascertained patients with ocular neuromyotonia through a search using the medical records database. Only patients with an observed episode of ocular neuromyotonia were included and the medical records were reviewed. The main outcome measures were clinical features and outcomes of patients with ocular neuromyotonia. RESULTS: Forty-two patients who were diagnosed with ocular neuromyotonia were included. The median age was 58 years (range, 16-80 years). A history of cranial radiation therapy was present in 39 patients (93%). The sixth cranial nerve was involved in 31 patients (74%). Bilateral disease was found in 2 patients (5%). The median time from onset of diplopia to diagnosis was 8 months (range, 1 month-25 years), with a high rate of initial misdiagnosis in 52%. Twenty of 42 patients (48%) were treated with oral medication, of whom 95% had significant improvement or resolution of symptoms. CONCLUSION: Prior cranial irradiation is the most common cause for ocular neuromyotonia, affecting the sixth cranial nerve most often. Although delayed and initial misdiagnosis is common, most patients show improved symptoms on medical treatment.

Age-related macular degeneration is associated with probable cerebral amyloid angiopathy: A case-control study.

BACKGROUND: Age-related macular degeneration (AMD) is a common retinal degenerative disorder among older individuals. Amyloid deposits, a hallmark of cerebral amyloid angiopathy (CAA), may be involved in the pathogenesis of AMD. Since amyloid deposits may contribute to the development of both AMD and CAA, we hypothesized that patients with AMD have a higher prevalence of CAA. OBJECTIVE: To compare the prevalence of CAA in patients with or without AMD matched for age. METHODS: We conducted a cross-sectional, 1:1 age-matched, case-control study of patients ≥40 years of age at the Mayo Clinic who had undergone both retinal optical coherence tomography and brain MRI from 2011 to 2015. Primary dependent variables were probable CAA, superficial siderosis, and lobar and deep cerebral microbleeds (CMBs). The relationship between AMD and CAA was assessed using multivariable logistic regression and was compared across AMD severity (none vs early vs late AMD). RESULTS: Our analysis included 256 age-matched pairs (AMD 126, no AMD 130). Of those with AMD, 79 (30.9%) had early AMD and 47 (19.4%) had late AMD. The mean age was 75±9 years, and there was no significant difference in vascular risk factors between groups. Patients with AMD had a higher prevalence of CAA (16.7% vs 10.0%, p=0.116) and superficial siderosis (15.1% vs 6.2%, p=0.020), but not deep CMB (5.2% vs 6.2%, p=0.426), compared to those without AMD. After adjusting for covariates, having late AMD was associated with increased odds of CAA (OR 2.83, 95% CI 1.10-7.27, p=0.031) and superficial siderosis (OR 3.40, 95%CI 1.20-9.65, p=0.022), but not deep CMB (OR 0.7, 95%CI 0.14-3.51, p=0.669). CONCLUSIONS: AMD was associated with CAA and superficial siderosis but not deep CMB, consistent with the hypothesis that amyloid deposits play a role in the development of AMD. Prospective studies are needed to determine if features of AMD may serve as biomarkers for the early diagnosis of CAA.

Visual Outcomes Following Plasma Exchange for Optic Neuritis: An International Multicenter Retrospective Analysis of 395 Optic Neuritis Attacks.

PURPOSE: To evaluate the effectiveness of plasma exchange (PLEX) for optic neuritis (ON). METHODS: We conducted an international multicenter retrospective study evaluating the outcomes of ON following PLEX. Outcomes were compared to raw data from the Optic Neuritis Treatment Trial (ONTT) using a matched subset. RESULTS: A total of 395 ON attack treated with PLEX from 317 patients were evaluated. The median age was 37 years (range 9-75), and 71% were female. Causes of ON included multiple sclerosis (108), myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD) (92), aquaporin-4-IgG-positive neuromyelitis optica spectrum disorder (AQP4+NMOSD) (75), seronegative-NMOSD (34), idiopathic (83), and other (3). Median time from onset of vision loss to PLEX was 2.6 weeks (interquartile range [IQR], 1.4-4.0). Median visual acuity (VA) at the time of PLEX was count fingers (IQR, 20/200-hand motion), and median final VA was 20/25 (IQR, 20/20-20/60) with no differences among etiologies except MOGAD-ON, which had better outcomes. In 81 (20.5%) ON attacks, the final VA was 20/200 or worse. Patients with poor outcomes were older (P = .002), had worse VA at the time of PLEX (P < .001), and longer delay to PLEX (P < .001). In comparison with the ONTT subset with severe corticosteroid-unresponsive ON, a final VA of worse than 20/40 occurred in 6 of 50 (12%) PLEX-treated ON vs 7 of 19 (37%) from the ONTT treated with intravenous methylprednisolone without PLEX (P = .04). CONCLUSION: Most ON attacks improved with PLEX, and outcomes were better than attacks with similar severity in the ONTT. The presence of severe vision loss at nadir, older age, and longer delay to PLEX predicted a worse outcome whereas MOGAD-ON had a more favorable prognosis. NOTE: Publication of this article is sponsored by the American Ophthalmological Society.

Details and outcomes of a large cohort of MOG-IgG associated optic neuritis.

BACKGROUND: The goal of this study was to examine the temporal relationship of eye pain to visual loss and investigate whether timing of steroid treatment affects the rate and extent of visual recovery in optic neuritis (ON) from MOG-IgG associated disease (MOGAD) in a large cohort of MOGAD patients with ON. METHODS: This is a multicenter, retrospective cohort study of consecutive MOGAD patients with ON attacks seen from 2017 to 2021 fulfilling the following criteria: (1) clinical history of ON; (2) MOG-IgG seropositivity. ON attacks were evaluated for presence/duration of eye pain, nadir of vision loss, time to intravenous methylprednisolone (IVMP) treatment, time to recovery, and final visual outcomes. RESULTS: There were 107 patients with 140 attacks treated with IVMP and details on timing of treatment and outcomes. Eye pain was present in 125/140 (89%) attacks with pain onset a median of 3 days (range, 0 to 20) prior to vision loss. Among 46 ON attacks treated with IVMP within 2 days of onset of vision loss, median time to recovery was 4 days (range, 0 to 103) compared to 15 days (range, 0 to 365) in 94 ON attacks treated after 2 days (p = 0.004). Those treated within 2 days had less severe VA loss at time of treatment (median LogMAR VA 0.48, range, 0.1 to 3) compared to those treated after 2 days (median LogMAR VA 1.7, range, 0 to 3; p < 0.001), and were more likely to have a VA outcome of 20/40 or better (98% vs 83%, p = 0.01). After adjustment for the initial VA at time of treatment, the differences in final VA were no longer significantly different (p = 0.14). In addition, some patients were documented to recover without steroid treatment. CONCLUSION: This study suggests that pain precedes vision loss in the majority of ON attacks and early steroids may lead to better outcomes in MOG-IgG ON, but some patients can recover without steroid treatment. Prospective randomized clinical trials are required to confirm these findings.

Oculomotor nerve root split: incidental finding on MRI-case report and literature review.

BACKGROUND: The oculomotor nerve (OMN) innervates the pupil, ciliary body, upper eyelid, and extraocular muscles through two divisions: a superior division that innervates the levator palpebrae superioris (LPS) and superior rectus (SR), and an inferior division that supplies the medial rectus (MR), inferior rectus (IR), inferior oblique (IO), and parasympathetic fibers to the pupil and ciliary body. We present a case of complete splitting of the cisternal segment of bilateral OMNs that was discovered incidentally on magnetic resonance imaging (MRI) in a patient who had no ocular complaints. CASE REPORT: A 69-year-old patient was found to have bilateral splitting of the cisternal segments of OMNs during an MRI for trigeminal neuralgia workup. Both nerves sprang from the midbrain as distinct roots. They were symmetric on the right and minimally asymmetric on the left. On both sides, the medial root was slightly inferiorly situated. The patient had no visual problems and continued to function normally. A review of the literature for similar cases identified no such variants; however, it did identify eight examples of OMN fenestrations produced by aneurysms (AN), six of which had no OMN palsy symptoms. CONCLUSION: An anatomic variant of split bilateral OMN cisternal segments is described. The superior and inferior divisions may have different brainstem origins. Although this variant is an anatomic curiosity, it may have clinical significance and explain the various presentation of compressive OMN palsies.

Frequency of Asymptomatic Optic Nerve Enhancement in a Large Retrospective Cohort of Patients With Aquaporin-4+ NMOSD.

BACKGROUND AND OBJECTIVES: Asymptomatic or persistent optic nerve enhancement in aquaporin-4 (AQP4)-immunoglobulin G (IgG)-positive neuromyelitis optica spectrum disorder (NMOSD) is thought to be rare. Improved understanding may have important implications for assessment of treatment efficacy in clinical trials and in clinical practice. Our objective was to characterize NMOSD interattack optic nerve enhancement. METHODS: This was a retrospective cohort study performed between 2000 and 2019 (median follow-up 5.5 [range 1-35] years) of patients with AQP4-IgG-positive optic neuritis (ON) evaluated at Mayo Clinic. MRI orbits were reviewed by a neuroradiologist, neuro-ophthalmologist, and neuroimmunologist blinded to the clinical history. Interattack optic nerve enhancement (>30 days after attack) was measured. The correlation between interattack enhancement and Snellen visual acuity (VA), converted to logarithm of the minimum angle of resolution (logMAR), at attack and at follow-up were assessed. RESULTS: A total of 198 MRI scans in 100 patients with AQP4-IgG+ NMOSD were identified, with 107 interattack MRIs from 78 unique patients reviewed. Seven scans were performed before any ON (median 61 days before attack [range 21-271 days]) and 100 after ON (median 400 days after attack [33-4,623 days]). Optic nerve enhancement was present on 18/107 (16.8%) interattack scans (median 192.5 days from attack [33-2,943]) of patients with preceding ON. On 15 scans, enhancement occurred at the site of prior attacks; the lesion location was unchanged, but the lesion length was shorter. Two scans (1.8%) demonstrated new asymptomatic lesions (prior scan demonstrated no enhancement). In a third patient with subjective blurry vision, MRI showed enhancement preceding detectable eye abnormalities on examination noted 15 days later. There was no difference in VA at preceding attack nadir (logMAR VA 1.7 vs 2.1; p = 0.79) or long-term VA (logMAR VA 0.4 vs 0.2, p = 0.56) between those with and without interattack optic nerve enhancement. DISCUSSION: Asymptomatic optic nerve enhancement occurred in 17% of patients with NMOSD predominantly at the site of prior ON attacks and may represent intermittent blood-brain barrier breakdown or subclinical ON. New asymptomatic enhancement was seen only in 2% of patients. Therapeutic clinical trials for NMOSD require blinded relapse adjudication when assessing treatment efficacy, and it is important to recognize that asymptomatic optic nerve enhancement can occur in patients with ON.

Association of Maintenance Intravenous Immunoglobulin With Prevention of Relapse in Adult Myelin Oligodendrocyte Glycoprotein Antibody-Associated Disease.

Importance: Recent studies suggest that maintenance intravenous immunoglobulin (IVIG) may be an effective treatment to prevent relapses in myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD); however, most of these studies had pediatric cohorts, and few studies have evaluated IVIG in adult patients. Objective: To determine the association of maintenance IVIG with the prevention of disease relapse in a large adult cohort of patients with MOGAD. Design, Setting, and Participants: This was a retrospective cohort study conducted from January 1, 2010, to October 31, 2021. Patients were recruited from 14 hospitals in 9 countries and were included in the analysis if they (1) had a history of 1 or more central nervous system demyelinating attacks consistent with MOGAD, (2) had MOG-IgG seropositivity tested by cell-based assay, and (3) were age 18 years or older when starting IVIG treatment. These patients were retrospectively evaluated for a history of maintenance IVIG treatment. Exposures: Maintenance IVIG. Main Outcomes and Measures: Relapse rates while receiving maintenance IVIG compared with before initiation of therapy. Results: Of the 876 adult patients initially identified with MOGAD, 59 (median [range] age, 36 [18-69] years; 33 women [56%]) were treated with maintenance IVIG. IVIG was initiated as first-line immunotherapy in 15 patients (25%) and as second-line therapy in 37 patients (63%) owing to failure of prior immunotherapy and in 7 patients (12%) owing to intolerance to prior immunotherapy. The median (range) annualized relapse rate before IVIG treatment was 1.4 (0-6.1), compared with a median (range) annualized relapse rate while receiving IVIG of 0 (0-3) (t108 = 7.14; P < .001). Twenty patients (34%) had at least 1 relapse while receiving IVIG with a median (range) time to first relapse of 1 (0.03-4.8) years, and 17 patients (29%) were treated with concomitant maintenance immunotherapy. Only 5 of 29 patients (17%) who received 1 g/kg of IVIG every 4 weeks or more experienced disease relapse compared with 15 of 30 patients (50%) treated with lower or less frequent dosing (hazard ratio, 3.31; 95% CI, 1.19-9.09; P = .02). At final follow-up, 52 patients (88%) were still receiving maintenance IVIG with a median (range) duration of 1.7 (0.5-9.9) years of therapy. Seven of 59 patients (12%) discontinued IVIG therapy: 4 (57%) for inefficacy, 2 (29%) for adverse effects, and 1 (14%) for a trial not receiving therapy after a period of disease inactivity. Conclusions and Relevance: Results of this retrospective, multicenter, cohort study of adult patients with MOGAD suggest that maintenance IVIG was associated with a reduction in disease relapse. Less frequent and lower dosing of IVIG may be associated with treatment failure. Future prospective randomized clinical trials are warranted to confirm these findings.

OCT retinal nerve fiber layer thickness differentiates acute optic neuritis from MOG antibody-associated disease and Multiple Sclerosis: RNFL thickening in acute optic neuritis from MOGAD vs MS.

BACKGROUND: Optic neuritis (ON) is the most common manifestation of myelin oligodendrocyte glycoprotein antibody associated disorder (MOGAD) and multiple sclerosis (MS). Acute ON in MOGAD is thought to be associated with more severe optic disk edema than in other demyelinating diseases, but this has not been quantitatively confirmed. The goal of this study was to determine whether optical coherence tomography (OCT) can distinguish acute ON in MOGAD from MS, and establish the sensitivity of OCT as a confirmatory biomarker of ON in these entities. METHODS: This was a multicenter cross-sectional study of MOGAD and MS patients with peripapillary retinal nerve fiber layer (pRNFL) thickness measured with OCT within two weeks of acute ON symptom. Cirrus HD-OCT (Carl Zeiss Meditec, Inc. Dublin, CA, USA) was used to measure the pRNFL during acute ON. Eyes with prior ON or disk pallor were excluded. A receiver operating characteristic (ROC) curve analysis was performed to assess the ability of pRNFL thickness to distinguish MOGAD from MS. RESULTS: Sixty-four MOGAD and 50 MS patients met study inclusion criteria. Median age was 46.5 years (interquartile range [IQR]: 34.3-57.0) for the MOGAD group and 30.4 years (IQR: 25.7-38.4) for the MS group (p<0.001). Thirty-nine (61%) of MOGAD patients were female compared to 42 (84%) for MS (p = 0.007). The median pRNFL thickness was 164 µm (IQR: 116-212) in 96 acute MOGAD ON eyes compared to 103 µm (IQR: 93-113) in 51 acute MS ON eyes (p<0.001). The ROC area under the curve for pRNFL thickness was 0.81 (95% confidence interval 0.74-0.88) to discriminate MOGAD from MS. The pRNFL cutoff that maximized Youden's index was 118 µm, which provided a sensitivity of 74% and specificity of 82% for MOGAD. Among 31 MOGAD and 48 MS eyes with an unaffected contralateral eye or a prior baseline, the symptomatic eye had a median estimated pRNFL thickening of 45 µm (IQR: 17-105) and 7.5 µm (IQR: 1-18), respectively (p<0.001). All MOGAD affected eyes had a ≥ 5 µm pRNFL thickening, whereas 26 (54%) MS affected eyes had a ≥ 5 µm thickening. CONCLUSION: OCT-derived pRNFL thickness in acute ON can help differentiate MOGAD from MS. This can aid with early diagnosis and guide disease-specific therapy in the acute setting before antibody testing returns, and help differentiate borderline cases. In addition, pRNFL thickening is a sensitive biomarker for confirming acute ON in MOGAD, which is clinically helpful and could be used for adjudication of attacks in future MOGAD clinical trials.

Misdiagnosis of Third Nerve Palsy.

BACKGROUND: This study identifies the diagnostic errors leading to misdiagnosis of 3rd nerve palsy and to aid clinicians in making this diagnosis. The objective of this article is to determine the incidence of misdiagnosis of 3rd cranial nerve palsy (3rd nerve palsy) among providers referring to a tertiary care neuro-ophthalmology clinic and to characterize diagnostic errors that led to an incorrect diagnosis. METHODS: This was a retrospective clinic-based multicenter cross-sectional study of office encounters at 2 institutions from January 1, 2014, to January 1, 2017. All encounters with scheduling comments containing variations of "3rd nerve palsy" were reviewed. Patients with a documented referral diagnosis of new 3rd nerve palsy were included in the study. Examination findings, including extraocular movement examination, external lid examination, and pupil examination, were collected. The final diagnosis was determined by a neuro-ophthalmologist. The Diagnosis Error Evaluation and Research (DEER) taxonomy tool was used to categorize the causes of misdiagnosis. Seventy-eight patients referred were for a new diagnosis of 3rd nerve palsy. The main outcome measure was the type of diagnostic error that led to incorrect diagnoses using the DEER criteria as determined by 2 independent reviewers. Secondary outcomes were rates of misdiagnosis, misdiagnosis rate by referring specialty, and examination findings associated with incorrect diagnoses. RESULTS: Of 78 patients referred with a suspected diagnosis of 3rd nerve palsy, 21.8% were determined to have an alternate diagnosis. The most common error in misdiagnosed cases was failure to correctly interpret the physical examination. Ophthalmologists were the most common referring provider for 3rd nerve palsy, and optometrists had the highest overdiagnosis rate of 3rd nerve palsy. CONCLUSIONS: Misdiagnosis of 3rd nerve palsy was common. Performance and interpretation of the physical examination were the most common factors leading to misdiagnosis of 3rd nerve palsy.

OS reboot.

Cerebrospinal fluid (CSF) diversion or shunting procedures are the most commonly performed surgery for the treatment of hydrocephalus and are often employed in the management of elevated intracranial pressure due to a variety of diseases. Despite their popularity however, approximately 50% of shunts fail within the first two years, and several revisions are required within the first decade after placement. Ophthalmologists may encounter patients with a CSF shunt to evaluate for concerns of vision loss or diplopia and to determine if papilledema is present. We discuss the neuro-ophthalmic manifestations and evaluation of possible CSF shunt malfunction.

The Outpatient Approach to Dizziness.

The evaluation of the dizzy patient is complicated by many common pitfalls. The patient's description of symptoms and the standard neurologic examination are often nonspecific or unrevealing, and neuroimaging is most often normal. Over the past several years, research has demonstrated that a refocusing of history taking results in more reliable and diagnostically helpful information. This can guide a targeted expansion of the exam, often with an emphasis on eye movements.

Lumbar Puncture for Diagnosis of Idiopathic Intracranial Hypertension in Typical Patients.

BACKGROUND: Patients with typical features of pseudotumor cerebri syndrome (PTCS) must undergo lumbar puncture (LP) to demonstrate elevated opening pressure and cerebrospinal fluid (CSF) analysis to rule out alternative diagnoses. As LP may be associated with significant morbidity, this study aims to determine its necessity in diagnosing typical PTCS. METHODS: Retrospective chart review at 3 university-based neuro-ophthalmology practices included women aged 18-45 years with body mass index >25, papilledema, negative neuroimaging, and who met criteria for PTCS or probable PTCS. RESULTS: One hundred fifty-six patients were enrolled. Seven (4.5%) had clinically insignificant CSF abnormalities. No diagnoses or management changed based on LP/CSF results. CONCLUSION: LP may not be routinely required in the initial evaluation of typical patients with PTCS evaluated by experienced clinicians We caution, however, that further prospective study is required to determine potential risks and benefits of LP as a tool in the diagnosis of IIH before recommending general practice changes.

Neuro-Ophthalmological Findings in Early Fatal Familial Insomnia.

Fatal familial insomnia (FFI) is a rare inherited prion disease characterized by sleep, autonomic, and motor disturbances. Neuro-ophthalmological abnormalities have been reported at the onset of disease, although not further characterized. We analyzed video recordings of eye movements of 6 patients with FFI from 3 unrelated kindreds, seen within 6 months from the onset of illness. Excessive saccadic intrusions were the most prominent findings. In patients with severe insomnia, striking saccadic intrusions are an early diagnostic clue for FFI. The fact that the thalamus is the first structure affected in FFI also suggests its role in the control of steady fixation. ANN NEUROL 2021;89:823-827.