PurposeTo quantitatively analyze swept-source OCT angiography (SS-OCTA) features of choroidal melanoma and nevi, focusing on their distinguishing angiographic features.DesignCase-series study.ParticipantsThis retrospective study included 102 consecutive patients with choroidal melanoma or nevi who were referred to Beijing Tongren Hospital between July 2019 and June 2024.Main Outcome MeasuresThe choriocapillaris and choroid slabs were en face reconstructed. A 5-grade system was developed based on the density and morphology of disorganized vasculature in the choroid slab. Choriocapillaris flow deficit (CCFD%) void was calculated within circular regions defined by the tumor's radius and diameter.MethodsThis retrospective study included 102 consecutive patients with choroidal melanoma or nevi referred to Beijing Tongren Hospital between July 2019 and June 2024. The choriocapillaris and choroid slabs were en face reconstructed. A 5-grade system was developed based on the density and morphology of disorganized vasculature in the choroid slab. Choriocapillaris flow deficit (CCFD%) void was calculated within circular regions defined by the tumor's radius and diameter.ResultsOf all the 102 patients, 87 (85%) patients with qualified SS-OCTA images, comprising 61 cases of choroidal melanoma and 26 cases of nevi. Sixty melanoma cases (98%) exhibited disorganized tumor vasculature forming twisted vessel loops and networks on the choroid slab (grade ≥2). Nevi showed 2 types of intrinsic tumor vasculature patterns in the choroidal slab: well-organized (grade 0 or 1) and melanoma-like disorganized networks (grade ≥2). Classified based on multimodal imaging risk score, nevi with a high risk for growth (≥2 risk factors) were more likely to exhibit melanoma-like disorganized networks tumor vasculature (82% vs. 27%, P = 0.015) than low-risk nevi. Analysis of the choriocapillaris slab revealed a higher prevalence of patchy avascular areas in melanomas compared to nevi. The CCFD% was significantly higher in choroidal melanomas than in nevi within circular regions defined by the tumor's radius (56.4% vs. 34.4%, P < 0.001) and diameter (44.9% vs. 23.4%, P < 0.001).ConclusionsSwept-source OCT angiography enables detailed noninvasive assessment of intrinsic vasculature in melanotic choroidal tumors.Financial Disclosure(s)Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

PurposeTo assess the relationship between foveal cone topography, foveal outer nuclear layer (ONL) thickness, foveal morphology, and foveal avascular zone (FAZ) area in individuals with normal vision.DesignRetrospective cross-sectional study.ParticipantsA total of 68 participants with normal vision were included (49 female; 19 male).MethodsDirectional OCT images were used to derive ONL thickness measurements. Images of the foveal cone mosaic were obtained using adaptive optics scanning light ophthalmoscopy, from which peak cone density (PCD) was measured. Foveal avascular zone area and foveal pit morphology were estimated using OCT angiography images and OCT macular thickness maps, respectively.Main Outcome MeasuresFoveal cone density metrics, foveal ONL thickness, foveal pit diameter and volume, and FAZ area.ResultsThere was a weak positive correlation between maximum ONL thickness and PCD in individuals with normal vision (r = 0.23; P = 0.06), and PCD was significantly negatively correlated with both foveal pit diameter (r = –0.54; P < 0.0001) and foveal pit volume (r = –0.39; P = 0.0011).ConclusionsFindings suggest that foveal ONL thickness should be used with caution as a clinical biomarker of foveal cone density, at least when measured using current OCT technology. The relationship between foveal pit size and foveal cone density supports possible mechanistic links between the processes that establish these important features of foveal specialization.Financial Disclosure(s)Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

PurposeTo investigate the association of caffeine citrate therapy with the onset and exacerbation of retinopathy of prematurity (ROP).DesignA retrospective observational cohort study.ParticipantsPreterm neonates born at Hyogo Prefectural Amagasaki General Medical Center between January 2018 and December 2023 were included.MethodsData on maternal (age; comorbid hypertensive disorders of pregnancy and gestational diabetes mellitus; fetal growth restriction; and cesarean section) and neonatal (sex; gestational age; birth weight; zone at initial examination; Apgar scores; respiratory distress syndrome; chronic lung disease; tracheal intubation; duration of tracheal intubation; duration of oxygen therapy; and caffeine therapy) risk factors for ROP were collected from medical records.Main Outcome MeasuresRetinopathy of prematurity onset and severe ROP were defined as stage ≥1 and stage ≥3 during the course of follow-up, respectively. The association between caffeine therapy and the onset of any ROP and severe ROP was investigated.ResultsAmong 202 neonates included in the current study, 94 (46.5%) developed any ROP (stage ≥1) during the observational period. Only the zone at initial examination affected the onset of any ROP (P < 0.01). However, in 115 neonates born at <32 weeks gestational age and with <1500 g birth weight, larger zones at baseline and caffeine treatment were associated with decreased severe ROP (P = 0.017 and P < 0.01, respectively). Additionally, in 54 neonates with gestational age between 23 and 28 weeks, only caffeine therapy was associated with decreased development of severe ROP (P < 0.01). All 4 neonates with gestational age <23 weeks progressed to severe ROP despite caffeine therapy.ConclusionsCaffeine therapy may be a potential treatment strategy to prevent the progression of severe ROP in neonates born between 23 and 28 weeks' gestational age.Financial Disclosure(s)The authors have no proprietary or commercial interest in any materials discussed in this article.

ObjectiveTo examine whether there is an association between age-related macular degeneration (AMD) and dementia using a large, multi-institutional clinical data.DesignA retrospective cohort study.ParticipantsPatients with AMD, including both neovascular AMD (nvAMD) and non-neovascular AMD (non-nvAMD) types, along with matched controls who had a record of eye examination but no diagnosis of AMD.MethodsData were extracted from the 3 largest TriNetX networks (Global, United States, and Europe, Middle East, and Africa [EMEA]). Dementia diagnoses (Alzheimer disease [AD], vascular dementia, and unspecified dementia) and dementia-related medication use were compared between AMD and non-AMD groups. Diagnoses were identified using International Classification of Diseases, 10th Revision codes. Survival analyses were performed using Kaplan–Meier curves, and hazard ratios (HRs) were estimated using Cox proportional hazards models across 5-, 7-, and 10-year follow-up periods.Main Outcome MeasuresHazard ratios for dementia subtypes in AMD patients versus non-AMD patients and in nvAMD patients versus non-nvAMD patients.ResultsNo significant association between AMD and AD was observed in the global network. In the United States, AMD was associated with a reduced risk of AD at 5 and 7 years (HR = 0.79 and 0.75; P < 0.0001), but not at 10 years. In the EMEA cohort, a protective association emerged at 7 and 10 years (HR = 0.16 and 0.33; P = 0.0063 and 0.0265). Vascular dementia showed no global association but was reduced in the United States at 5 and 7 years (HR = 0.73 and 0.79; P < 0.0001 and 0.0003). Unspecified dementia was associated with increased risk globally (HR = 1.17–1.27; P ≤ 0.0016), while the US data indicated a protective trend at 5 and 7 years. Dementia medication use was elevated in the global network at 7 years (HR = 1.27, P < 0.0001), but lower in the United States across all time points (P < 0.01). The negative control outcome (appendicitis) showed no difference between groups. No significant differences in AD or vascular dementia were found between nvAMD and non-AMD patients, although unspecified dementia and dementia medication use were lower in the nvAMD group.ConclusionsAge-related macular degeneration was not consistently positively associated with AD or vascular dementia across networks. Regional variation, particularly negative associations in the United States, suggests underlying health care or diagnostic differences.Financial Disclosure(s)Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

ObjectiveAssess retinal perfusion in sepsis, compared with uncomplicated postoperative care and healthy controls, and assess the effects of reduced perfusion on retinal structure and visual function.DesignWe conducted a prospective observational cohort study between March 2018 and December 2022, with follow-up measures collected 3 to 6 months after discharge.SubjectsTwenty-four patients with sepsis were assessed in the intensive care unit (ICU) and 3 to 6 months later, 45 ICU control patients assessed during elective ICU admission after upper gastrointestinal cancer surgery, preoperatively, and 3 to 6 months later, and 15 healthy controls.TestingAssessments included retinal layer thickness using OCT, retinal perfusion using OCT angiography, and visual function using Humphrey visual field analysis. Organ dysfunction was assessed by Sequential Organ Failure Assessment (SOFA) scoring.Main Outcome MeasuresSuperficial vascular plexus (SVP) retinal perfusion, OCT retinal ganglion cell layer (GCL) thickness, and mean deviation (MD) on Humphrey visual field testing were evaluated.ResultsSuperficial vascular plexus retinal perfusion was 37.4% lower in patients with sepsis compared with ICU control patients (P < 0.001) and 59.7% lower than in healthy controls, which returned to normal by final follow-up. Retinal perfusion correlated with the SOFA score (Pearson r = -0.57, P < 0.001) and weakly correlated with C-reactive protein (r = -0.337, P = 0.01) and mean arterial pressure (r = 0.354, P = 0.006). In patients with sepsis and ICU controls, retinal perfusion in the ICU predicted subsequent GCL thickening, with every 1-unit decrease in SVP sum predicting a 1.88 μm increase in GCL thickness at follow-up (P = 0.003), and worsening visual field MD, with every 1-unit decrease in SVP sum predicting a 0.078 decibel lower MD (P = 0.023).ConclusionsRetinal perfusion was impaired in patients with sepsis compared with both healthy controls and patients after major surgery. It was moderately associated with other measures of organ dysfunction assessed by SOFA. Reduced retinal perfusion in both patients with sepsis and patients after major surgery is strongly associated with subsequent GCL thickening and less strongly associated with decreased visual field MD, suggesting reduced retinal perfusion is associated with retinal damage, with consequent visual dysfunction.Financial Disclosure(s)Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

PurposeTo validate a custom FIJI (ImageJ) program for more reproducible, faster curvilinear periorbital measurements, as compared with 2 custom artificial intelligence–based tools.DesignCombined technical validation and method comparison study.SubjectsFront-facing photographs of 45 cleft palate syndromic patients.MethodsA FIJI (ImageJ)-based macro script tool, OrbitJ (semiautomated), was developed, requiring 15 user input steps to generate 38 measurements per photo. The user outlines the irises to set image scale and then selects points along the lid margins and brow lines. Linear interpolation and fourth-degree polynomial fit lines were used to generate periorbital measurements. This tool was compared against our previously developed deep learning algorithm for periorbital measurements, OrbitMap (automated), another open-source algorithm PeriOrbitAI (automated), and against manual measurements. Four human graders measured 45 photos once both manually and with OrbitJ. Intrarater and interrater measurements were performed with 10 photos manually, 3 in triplicate manually and 5 in triplicate with OrbitJ. Fourteen manual measurements were performed: time per image, iris diameter, margin reflex distance (MRD) 1 and 2, inferior scleral show (ISS), medial, central, and lateral brow height, canthal tilt, vertical dystopia, interpupillary distance, and inner and outer canthal distance (OCD).Main Outcome MeasuresThe mean absolute error, reliability, bias, and Pearson correlation of periorbital measurements.ResultsAnalysis was successful in all 45 images for all methods except PeriOrbitAI, which failed on 6 images. For manual and semiautomated intra- and interrater measurements, reliability was considered moderate or better (intraclass correlation coefficient [ICC] >0.5) for all measurements excluding elapsed time. Manual interrater mean absolute error was <1 mm all measures except OCD. Reliability and correlation were high (ICC, Pearson correlation coefficient >0.8) between all OrbitJ and manual measurements. Comparing OrbitMap to manual measurements, ICC and Pearson correlation coefficient were >0.5 except for ISS, borderline for OCD (ICC = 0.51). Reliability between PeriOrbitAI and manual measurements was low (ICC <0.5) except for MRD2 and OCD, and correlation was moderate (Pearson correlation coefficient = 0.49–0.75). The mean analysis time per image was 13.4 ± 4 minutes (manual measurements), 5.4 ± 1.9 minutes (semiautomated OrbitJ), 10.71 ± 1.65 seconds (automated PeriOrbitAI), and 1.45 ± 0.15 seconds (automated OrbitMap) (P < 0.001).ConclusionsCompared with manual measurements, all semiautomated OrbitJ measurements and most automated OrbitMap measurements were reliable. Notably, only 2 PeriOrbitAI measurements were reliable. All methods demonstrated significant time savings over manual measurements.Financial Disclosure(s)Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.