Using Natural Language Processing to Identify Different Lens Pathology in Electronic Health Records.

PURPOSE: Nearly all published ophthalmology-related Big Data studies rely exclusively on International Classification of Diseases (ICD) billing codes to identify patients with particular ocular conditions. However, inaccurate or nonspecific codes may be used. We assessed whether natural language processing (NLP), as an alternative approach, could more accurately identify lens pathology. DESIGN: Database study comparing the accuracy of NLP versus ICD billing codes to properly identify lens pathology. METHODS: We developed an NLP algorithm capable of searching free-text lens exam data in the electronic health record (EHR) to identify the type(s) of cataract present, cataract density, presence of intraocular lenses, and other lens pathology. We applied our algorithm to 17.5 million lens exam records in the Sight Outcomes Research Collaborative (SOURCE) repository. We selected 4314 unique lens-exam entries and asked 11 clinicians to assess whether all pathology present in the entries had been correctly identified in the NLP algorithm output. The algorithm's sensitivity at accurately identifying lens pathology was compared with that of the ICD codes. RESULTS: The NLP algorithm correctly identified all lens pathology present in 4104 of the 4314 lens-exam entries (95.1%). For less common lens pathology, algorithm findings were corroborated by reviewing clinicians for 100% of mentions of pseudoexfoliation material and 99.7% for phimosis, subluxation, and synechia. Sensitivity at identifying lens pathology was better for NLP (0.98 [0.96-0.99] than for billing codes (0.49 [0.46-0.53]). CONCLUSIONS: Our NLP algorithm identifies and classifies lens abnormalities routinely documented by eye-care professionals with high accuracy. Such algorithms will help researchers to properly identify and classify ocular pathology, broadening the scope of feasible research using real-world data.

Weather Patterns and the Frequency of Ophthalmology Consultations in the Emergency Setting.

PURPOSE: To determine if weather patterns are associated with the frequency of ophthalmology consultations in the Emergency Room (ER) and trauma settings. METHODS: Hospital-based ophthalmology consultations between January 1, 2015 and December 31, 2020 at the University of Maryland Medical Center (UMMC) were included in the BALCITE (BALtimore Consultation, Inpatient, and Trauma of the Eye) database. Encounters were filtered to the general ER, the R. Crowley Adams Shock Trauma Center, and consultations within 24 h of admission where a delay was attributed to bed assignment. Weather data from the National Climatic Data Center (NCDC) for Baltimore, Maryland, was matched to the day of the initial encounter. RESULTS: A total of 3,877 patients were included. Overall, there was a significant association between increasing daily average temperature (R2 = .152, p = 0.0003) and daily maximum temperature (R2 = .243, p < 0.001) with incidence of ophthalmology consultations. There was no significant decrease in consultations with increasing precipitation (R2 = -.007, p = 0.7477). For trauma-related consultations, there was not a significant association between consultation incidence and average daily temperature (R2 = .011, p = 0.2013), maximum daily temperature (R2 = -0.012, p = 0.6529), or precipitation levels (R2 = .075, p = 0.24). The months with the highest consultation volume were September, August, and July, and the lowest numbers in April, March, and February. CONCLUSIONS: Emergency ophthalmology consultations increased with increasing daily temperature. Trauma-specific consultations did not show a significant association with weather patterns. This information may prove useful to ER providers and ophthalmologists in order to help predict frequency of consultations and better deploy personnel and resources.

Repeatability and reliability of semi-automated anterior segment-optical coherence tomography imaging compared to manual analysis in normal and keratoconus eyes.

PURPOSE: To assess the repeatability and reliability of semi-automated EyeMark Python program measurements compared to manual ImageJ image processing of anterior segment-optical coherence tomography (AS-OCT) structures in healthy and keratoconus eyes. METHODS: Heidelberg AS-OCT was used to image 25 eyes from 14 healthy subjects and 25 eyes from 15 subjects with keratoconus between the ages of 20 and 80 years, collected prospectively, in this observational case-control study. Visual axis scan containing vertical fixation light beam was selected from the 15-line AS-OCT scan raster. Central corneal thickness (CCT), anterior corneal radius of curvature (ACRC), posterior corneal radius of curvature (PCRC), and truncated anterior vault (TAV) were measured using ImageJ software and the EyeMark Python program. MedCalc and R were used to calculate the intraclass correlation coefficient (ICC) and generate Bland-Altman plots (BAP). RESULTS: When comparing the measurements of CCT, ACRC, PCRC, and TAV between manual ImageJ analysis and the EyeMark Python program, ICC values were consistently greater than 0.9, indicating excellent agreement. BAPs comparing the ImageJ and Python measurements of anterior segment structures show no systematic proportional bias and the average differences were near zero and within 95% of the limits of agreement. CONCLUSIONS: Semi-automated tools may provide the necessary efficiency for point-of-care quantitative corneal analysis of raw AS-OCT images. The semi-automated EyeMark Python program offers a repeatable and reliable tool compared to manual ImageJ analysis for measuring anterior segment structures from AS-OCT images among individuals with keratoconus.

Association Between Socioeconomic Deprivation and Orbital Trauma in a Hospital-Based Population.

PURPOSE: To examine the association between area-level socioeconomic deprivation and orbital trauma among emergency ophthalmology consults. METHODS: We conducted a cross-sectional study using 5-year Epic data for all hospital-based ophthalmology consults at the University of Maryland Medical System, and the Distressed Communities Index (DCI) data for area-level socioeconomic deprivation. We ran multivariable logistic regression models adjusting for age to compute odds ratios (OR) and 95% confidence intervals (CI) for the association between the DCI quintile 5 distressed score and orbital trauma. RESULTS: A total of 3,811 cases of acute emergency consults were identified, of whom 750 (19.7%) had orbital trauma and 2,386 (62.6%) had other traumatic ocular emergencies. The odds of orbital trauma among people living in a distressed community were 0.59 (95% CI 0.46-0.76) times the odds for those living in a prosperous community. Among White subjects, the odds of orbital trauma for people living in a distressed community were 1.71 (95% CI 1.12-2.62) times the odds for those living in a prosperous community; among Black subjects, the OR was 0.47 (95% CI 0.30-0.75; p-interaction = 0.0001). Among women, the OR for orbital trauma among those living in a distressed community was 0.46 (95% CI 0.29-0.71); among men, the OR was 0.70 (95% CI 0.52-0.97; p-interaction = 0.03). CONCLUSION: Overall, we found an inverse association between higher area-level socioeconomic deprivation and orbital trauma among both men and women. The association varied by race, such that there was an inverse association with higher deprivation among Black subjects in contrast to a positive association among White subjects.

Characteristics of Semiautomated Endothelial Cell-Density Measurements Among Corneal Donor Eyes.

Importance: Endothelial cell density in corneal donor eyes remains a critical component in selecting tissue suitable for corneal transplant. Objective: To describe an anomaly in endothelial cell density discovered in a large cohort of corneal donor eyes. Design, Setting, and Participants: Donor information for this cross-sectional study was obtained from the CorneaGen eye bank for donors from June 1, 2012, to June 30, 2016. Endothelial cell density was reported by eye bank technicians for each eye. The retrospective donor data set included donor demographics, endothelial cell count, time of death, medical and surgical history, and suitability for transplant. The donor data set contained information on donated eyes over the 4-year period. Main Outcomes and Measures: Endothelial cell-density heat map, 2-dimensional binned plots, and bayesian change point analysis. Results: A total of 48 207 donated eyes were evaluated. Mean (SD) cell count was 2717 (448) cells/mm2, and mean (SD) donor age was 58 (13) years. Heat maps using 2-dimensional binned plots demonstrated accentuated endothelial cell density results particularly at and immediately above 2500 cells/mm2 with a small gap right under this figure. A bayesian change point analysis revealed a change point at exactly this location as well. Conclusions and Relevance: Data from this cross-sectional study suggest a discontinuity in endothelial cell-density measurements just below 2500 cells/mm2. This anomaly suggests that the semiautomated counting technique used in this cohort preferentially skipped over measurements just below this cutoff. Whether this represents an automation error, or more likely, a subjective human bias, requires further study.

Estimated Prevalence of Keratoconus in the United States From a Large Vision Insurance Database.

PURPOSE: To estimate the prevalence of keratoconus in the United States using a large vision insurance database. METHODS: We used data from the 2016 vision service plan (VSP) Global claims database to estimate the prevalence of keratoconus among enrollees by state. The VSP Global database includes cross-sectional de-identified claims data summarized by state of residence in predefined age and gender groups. Crude values of prevalence of keratoconus are provided in the VSP Global database. The open-source software package qGIS version 3.12 was used as the geographic information system for geospatial analysis. RESULTS: The prevalence of keratoconus among VSP enrollees in the United States was estimated to be 0.15% with 95% confidence interval (CI) (0.15%, 0.15%). The prevalence of keratoconus was highest in Western states and among male population. The highest prevalence of keratoconus was in Colorado (0.51%; 95% CI 0.49%-0.52%), with particularly high prevalence for men in the age groups 18 to 39 years (0.90%; 95% CI 0.84%-0.95%) and 40 to 64 years (0.90%; 95% CI 0.85%-0.95%). CONCLUSION: The average prevalence of keratoconus in the United States in 2016 may be higher than previously reported. Living in Western states and male sex were associated with high prevalence.

Inter-rater Reliability and Repeatability of Manual Anterior Segment Optical Coherence Tomography Image Grading in Keratoconus.

OBJECTIVES: To determine the repeatability of corneal measurements from anterior segment optical coherence tomography (AS-OCT) images using ImageJ software in healthy eyes compared with eyes with keratoconus. METHODS: Anterior segment OCT images of 25 eyes from 14 healthy subjects and 25 eyes from 15 subjects with keratoconus between the ages of 20 and 80 years were evaluated. Two trained observers used ImageJ to measure the central corneal cross-sectional area and anterior and posterior corneal arc lengths. MedCalc statistical software was used to generate the intraclass correlation coefficient (ICC) and Bland-Altman plots (BAPs) for observer measurements. RESULTS: Observer measurements of the central corneal cross-sectional area and anterior and posterior corneal arc lengths yielded an ICC >0.7. The ICC comparing the 3 parameters ranged from 0.75 to 0.84 for the control and 0.96 to 0.98 for the keratoconus group. No systematic proportional bias was detected by the BAPs. There were minimal differences between the 2 observer's measurements, with a mean of the difference of 0.3 mm2, 0 mm, and 0 mm, for the 3 measurements, respectively. CONCLUSIONS: This study suggests that ImageJ software is a repeatable and reliable tool in the analysis of corneal parameters from AS-OCT images among patients with keratoconus and may be applicable to AS-OCT imaging protocol development, an area of active keratoconus research.

Effect of Body Refrigeration Time on Cornea Donor Tissue.

PURPOSE: To determine how early body refrigeration affects corneal donor transplant suitability and endothelial cell density. METHODS: Donor information was obtained from the CorneaGen Eye Bank including demographics, time of death to preservation, and body refrigeration status, for donors between 2012 and 2016. The death to preservation interval was classified into 3 categories: 0 to 10, 10 to 20, and 20+ hours. Two primary logistic method models were fit using a main effects model and an interaction model to determine the association of body refrigeration on unsuitability of transplantation and endothelial cell density. RESULTS: Analysis was from 42,929 donor eyes, with a mean (standard deviation) endothelial cell count of 2743 (415) cells/mm2. Fifty-nine percent of donor eyes were from male donors in the eye bank data set, and the mean death to preservation interval was 11.0 (5.6) hours for all eyes. Unsuitability for transplantation demonstrated a reduced adjusted odds ratio by 22% (OR = 0.78, P = 0.009) when the body was refrigerated during the death to preservation interval versus when the body was not refrigerated. Eyes that were refrigerated, however, exhibited no statistically significant difference in endothelial cell count from eyes that were not refrigerated (P = 0.12). CONCLUSIONS: We demonstrate an appreciable effect of early body refrigeration on transplant suitability in this large cohort of eye bank eyes. There was no beneficial effect of body refrigeration on endothelial cell count.

A review of imaging modalities for detecting early keratoconus.

OBJECTIVES: Early identification of keratoconus is imperative for preventing iatrogenic corneal ectasia and allowing for early corneal collagen cross-linking treatments to potentially halt progression and decrease transplant burden. However, early diagnosis of keratoconus is currently a diagnostic challenge as there is no uniform screening criteria. We performed a review of the current literature to assess imaging modalities that can be used to help identify subclinical keratoconus. METHODS: A Pubmed database search was conducted. We included primary and empirical studies for evaluating different modalities of screening for subclinical keratoconus. RESULTS: A combination of multiple imaging tools, including corneal topography, tomography, Scheimpflug imaging, anterior segment optical coherence tomography, and in vivo confocal microscopy will allow for enhanced determination of subclinical keratoconus. In patients who are diagnostically borderline using a single screening criteria, use of additional imaging techniques can assist in diagnosis. Modalities that show promise but need further research include polarization-sensitive optical coherence tomography, Brillouin microscopy, and atomic force microscopy. CONCLUSIONS: Recognition of early keratoconus can reduce risk of post-refractive ectasia and reduce transplantation burden. Though there are no current uniform screening criterion, multiple imaging modalities have shown promise in assisting with the early detection of keratoconus.

A Review of Ocular Graft-Versus-Host Disease.

: Graft-versus-host disease (GVHD) is a major complication that occurs following allogeneic hematopoietic stem cell transplantation, which is a potential curative therapy used in a variety of malignant or benign hematological diseases. Graft-versus-host disease primarily occurs in many organs, but most notably in the skin, lungs, gastrointestinal tract, liver, eyes, mucosa, and musculoskeletal system. Ocular manifestations of GVHD may precede other systemic GVHD findings, and it may be a poor prognosis for mortality. While all parts of the eye may be affected, ocular GVHD occurs primarily in the ocular surface. Dry eye disease or keratoconjunctivitis sicca is the most common presenting manifestation of chronic ocular GVHD. Dry eye disease in ocular GVHD is a multifactorial process, which involves destruction and fibrosis of lacrimal glands and conjunctiva, leading to tear film deficiency and instability. Depending on the severity of ocular involvement and response to treatment, ocular GVHD may cause decreased quality of life. Management of GVHD begins with prevention by understanding risk factors and by implementing prophylactic treatment after allogeneic hematopoietic stem cell transplantation. A multidisciplinary approach to the prevention and treatment of GVHD is important, and there are currently no preventive therapies available for ocular GVHD. Once diagnosed, ocular GVHD treatment strategies target ocular surface lubrication and support, tear film stabilization, inflammation reduction, and surgical intervention. The goal of this review is to define ocular GVHD and its categorical manifestations, as well as to describe the importance of comprehensive assessment, diagnosis, and ophthalmologic treatment and management of ocular GVHD with a multidisciplinary approach.