Large language models approach expert-level clinical knowledge and reasoning in ophthalmology: A head-to-head cross-sectional study.

Large language models (LLMs) underlie remarkable recent advanced in natural language processing, and they are beginning to be applied in clinical contexts. We aimed to evaluate the clinical potential of state-of-the-art LLMs in ophthalmology using a more robust benchmark than raw examination scores. We trialled GPT-3.5 and GPT-4 on 347 ophthalmology questions before GPT-3.5, GPT-4, PaLM 2, LLaMA, expert ophthalmologists, and doctors in training were trialled on a mock examination of 87 questions. Performance was analysed with respect to question subject and type (first order recall and higher order reasoning). Masked ophthalmologists graded the accuracy, relevance, and overall preference of GPT-3.5 and GPT-4 responses to the same questions. The performance of GPT-4 (69%) was superior to GPT-3.5 (48%), LLaMA (32%), and PaLM 2 (56%). GPT-4 compared favourably with expert ophthalmologists (median 76%, range 64-90%), ophthalmology trainees (median 59%, range 57-63%), and unspecialised junior doctors (median 43%, range 41-44%). Low agreement between LLMs and doctors reflected idiosyncratic differences in knowledge and reasoning with overall consistency across subjects and types (p>0.05). All ophthalmologists preferred GPT-4 responses over GPT-3.5 and rated the accuracy and relevance of GPT-4 as higher (p<0.05). LLMs are approaching expert-level knowledge and reasoning skills in ophthalmology. In view of the comparable or superior performance to trainee-grade ophthalmologists and unspecialised junior doctors, state-of-the-art LLMs such as GPT-4 may provide useful medical advice and assistance where access to expert ophthalmologists is limited. Clinical benchmarks provide useful assays of LLM capabilities in healthcare before clinical trials can be designed and conducted.

Superior oblique tuck-effectiveness in reducing vertical deviations when performed as a primary versus secondary procedure in superior oblique paresis.

PURPOSE: To quantify the effectiveness of superior oblique tuck (SOT) surgery in patients with a hyperdeviation secondary to superior oblique paresis (SOP). Surgical outcomes were compared in patients undergoing SOT surgery as a primary procedure with those who had previously undergone ipsilateral inferior oblique weakening surgery. METHODS: This retrospective study assessed surgical outcomes from all patients undergoing SOT surgery for SOP between 2012 and 2021 across 2 hospitals. The effectiveness of SOT surgery in reducing the hyperdeviation was assessed in the primary position (PP) and in contralateral elevation and depression. Results were compared between those undergoing primary SOT surgery with those who had previously undergone ipsilateral inferior oblique weakening surgery. RESULTS: A total of 60 SOT procedures were performed between 2012 and 2021. 7 were removed due to incomplete data. The remaining 53 cases experienced a mean reduction in hyperdeviation of 6.5 prism dioptres (PD), 6.7PD and 12.0PD in the PP, contralateral elevation and contralateral depression respectively. In eyes with previous IO weakening, the reduction of hyperdeviation was larger than in those eyes with no previous IO weakening surgery, with mean reductions of 8.0PD vs 5.2PD, 7.4PD vs 6.2PD and 12.4PD vs 11.6PD in the PP, contralateral elevation and contralateral depression respectively. CONCLUSION: SOT surgery is a safe and effective procedure with high patient satisfaction and resolution of symptoms in those patients with troublesome diplopia in downgaze secondary to SOP. This is true in both unoperated eyes and in those who have previously undergone inferior oblique weakening surgery.

Combined Phacoemulsification and Intravitreal Dexamethasone Is an Effective Option for High-Risk Diabetic Macula Oedema Patients.

Objective Cataract surgery in diabetic patients carries an increased risk of post-operative macula oedema, particularly in those with a history of diabetic macula oedema (DMO) treatment or DMO at the time of surgery. We investigated whether simultaneous phacoemulsification with intravitreal Ozurdex® reduces the risk of developing new, or deteriorating current, DMO. Methods We conducted a retrospective review of 79 consecutive 'high-risk' diabetic patients who underwent phacoemulsification with intraocular lens insertion and intravitreal Ozurdex® implantation immediately subsequently. 'High risk' was defined as diabetic patients with prior treatment history for DMO or current DMO. Central macula thickness (CMT), best-corrected visual acuity and intraocular pressure were recorded pre-operatively, at two to four weeks and at three months post-operatively. A significant change in CMT was defined as a change of ≥0.1 LogOCT units. Results The mean age was 72.6 years; 52% were males. The mean pre-operative CMT was 365um. Thirty-seven per cent (37%) patients had prior DMO history that had resolved; 63% had confirmed DMO in surgery. Two to four weeks post-operatively, 82% of patients had stable CMT and 18% showed improvement. No patients deteriorated. Three months post-operatively, 48% of patients had stable CMT relative to pre-operative measurements, 38% improved, and 14% deteriorated. Analysis of variance (ANOVA) indicated no significant differences in response with demographical or pathological factors, including diabetic retinopathy grade and treatment history. Conclusion Phacoemulsification surgery combined with Ozurdex® insertion at the end of the procedure is a highly effective strategy for protecting against the formation of new, or the deterioration of current DMO, in the highest risk diabetic patients undergoing cataract surgery.

Paediatric Idiopathic Intracranial Hypertension (IIH)-A Review.

Paediatric idiopathic intracranial hypertension (IIH), is a rare but important differential diagnosis in children presenting with papilloedema. It is characterised by raised intracranial pressure in the absence of an identifiable secondary structural or systemic cause and is, therefore, a diagnosis of exclusion. In the adult population, there is a strong predilection for the disease to occur in female patients who are obese. This association is also seen in paediatric patients with IIH but primarily in the post-pubertal cohort. In younger pre-pubertal children, this is not the case, possibly reflecting a different underlying disease aetiology and pathogenesis. Untreated IIH in children can cause significant morbidity from sight loss, chronic headaches, and the psychological effects of ongoing regular hospital monitoring, interventions, and medication. The ultimate goal in the management of paediatric IIH is to protect the optic nerve from papilloedema-induced optic neuropathy and thus preserve vision, whilst reducing the morbidity from other symptoms of IIH, in particular chronic headaches. In this review, we will outline the typical work-up and diagnostic process for paediatric patients with suspected IIH and how we manage these patients.

Use of MultiColor imaging in the assessment of suspected papilledema in 20 consecutive children.

PURPOSE: To investigate the use of MultiColor Scanning Laser Imaging as an adjunct to traditional imaging modalities in the assessment of children with suspected disk swelling. METHODS: In this study of 20 consecutive children with suspected papilledema, MultiColor imaging was performed on all patients in addition to fundus examination, fundus autofluorescence, disk spectral domain optical coherence tomography (SD-OCT), and disk retinal nerve fiber layer thickness (RNFL) measurement. Diagnoses were determined in all 20 patients, and all images were evaluated by a medical retina specialist. RESULTS: Of the 20 cases (average age 11 years [range, 5-16 years]; 10 males), papilledema was confirmed in 11 (55%) and pseudopapilledema diagnosed in 9 (45%). Of pseudopapilledema cases, there were 4 cases of optic disk drusen, 2 hypermetropic disks, 2 "crowded disks," and 1 anomalous disk. In patients with true papilledema, on the combined MultiColor image, a green shift in the form of an elevated green ring was seen consistently. This ring was also hyperreflective on the blue and green images and surrounded a central "shadow," which was seen best on near infrared reflectance (NIR). The disk margins and vasculature were obscured on the combined, green, blue, and NIR images. These changes were not present in the pseudopapilledema cases. CONCLUSIONS: MultiColor imaging of disks with papilledema shows characteristic changes that may prove useful in differentiating true papilledema from pseudopapilledema. It is therefore a useful adjunct to traditional disk imaging modalities in assessing children with suspected disk swelling.