Understanding the Strengths and Limitations of Online Oocyte Cryopreservation Calculators.

Between 2010 and 2016, elective oocyte cryopreservation (OC) increased in use by 880% in the United States; however, there have been increasing reports of regret among patients after elective OC. There is a growing need for individualized counseling on the timing and number of oocytes to cryopreserve for patients to make informed choices and set realistic expectations, but currently available tools seem to be insufficient. The purpose of this review is to describe the OC calculators currently available online, identify sources of regret, and illustrate the need for unified counseling tools for improved patient care and education. OC calculators were identified via Google search. Only calculators that cite scientific literature were included in the review. Calculators for in vitro fertilization or embryo transfer were excluded. Thirteen OC calculators were found; however, only six cited literature supporting the calculator's design. When entering the same hypothetical patient parameters for age and number of oocytes cryopreserved, the calculators provided drastically different probabilities of live births. The lack of cohesive online educational materials creates confusion and stress for patients considering OC, leading to unrealistic expectations and increased feelings of regret thereafter. Physicians need tools to provide comprehensive guidance to patients seeking to cryopreserve oocytes.

Predicting polycystic ovary syndrome with machine learning algorithms from electronic health records.

Introduction: Predictive models have been used to aid early diagnosis of PCOS, though existing models are based on small sample sizes and limited to fertility clinic populations. We built a predictive model using machine learning algorithms based on an outpatient population at risk for PCOS to predict risk and facilitate earlier diagnosis, particularly among those who meet diagnostic criteria but have not received a diagnosis. Methods: This is a retrospective cohort study from a SafetyNet hospital's electronic health records (EHR) from 2003-2016. The study population included 30,601 women aged 18-45 years without concurrent endocrinopathy who had any visit to Boston Medical Center for primary care, obstetrics and gynecology, endocrinology, family medicine, or general internal medicine. Four prediction outcomes were assessed for PCOS. The first outcome was PCOS ICD-9 diagnosis with additional model outcomes of algorithm-defined PCOS. The latter was based on Rotterdam criteria and merging laboratory values, radiographic imaging, and ICD data from the EHR to define irregular menstruation, hyperandrogenism, and polycystic ovarian morphology on ultrasound. Results: We developed predictive models using four machine learning methods: logistic regression, supported vector machine, gradient boosted trees, and random forests. Hormone values (follicle-stimulating hormone, luteinizing hormone, estradiol, and sex hormone binding globulin) were combined to create a multilayer perceptron score using a neural network classifier. Prediction of PCOS prior to clinical diagnosis in an out-of-sample test set of patients achieved an average AUC of 85%, 81%, 80%, and 82%, respectively in Models I, II, III and IV. Significant positive predictors of PCOS diagnosis across models included hormone levels and obesity; negative predictors included gravidity and positive bHCG. Conclusion: Machine learning algorithms were used to predict PCOS based on a large at-risk population. This approach may guide early detection of PCOS within EHR-interfaced populations to facilitate counseling and interventions that may reduce long-term health consequences. Our model illustrates the potential benefits of an artificial intelligence-enabled provider assistance tool that can be integrated into the EHR to reduce delays in diagnosis. However, model validation in other hospital-based populations is necessary.

Predicting polycystic ovary syndrome (PCOS) with machine learning algorithms from electronic health records.

Introduction: Predictive models have been used to aid early diagnosis of PCOS, though existing models are based on small sample sizes and limited to fertility clinic populations. We built a predictive model using machine learning algorithms based on an outpatient population at risk for PCOS to predict risk and facilitate earlier diagnosis, particularly among those who meet diagnostic criteria but have not received a diagnosis. Methods: This is a retrospective cohort study from a SafetyNet hospital's electronic health records (EHR) from 2003-2016. The study population included 30,601 women aged 18-45 years without concurrent endocrinopathy who had any visit to Boston Medical Center for primary care, obstetrics and gynecology, endocrinology, family medicine, or general internal medicine. Four prediction outcomes were assessed for PCOS. The first outcome was PCOS ICD-9 diagnosis with additional model outcomes of algorithm-defined PCOS. The latter was based on Rotterdam criteria and merging laboratory values, radiographic imaging, and ICD data from the EHR to define irregular menstruation, hyperandrogenism, and polycystic ovarian morphology on ultrasound. Results: We developed predictive models using four machine learning methods: logistic regression, supported vector machine, gradient boosted trees, and random forests. Hormone values (follicle-stimulating hormone, luteinizing hormone, estradiol, and sex hormone binding globulin) were combined to create a multilayer perceptron score using a neural network classifier. Prediction of PCOS prior to clinical diagnosis in an out-of-sample test set of patients achieved AUC of 85%, 81%, 80%, and 82%, respectively in Models I, II, III and IV. Significant positive predictors of PCOS diagnosis across models included hormone levels and obesity; negative predictors included gravidity and positive bHCG. Conclusions: Machine learning algorithms were used to predict PCOS based on a large at-risk population. This approach may guide early detection of PCOS within EHR-interfaced populations to facilitate counseling and interventions that may reduce long-term health consequences. Our model illustrates the potential benefits of an artificial intelligence-enabled provider assistance tool that can be integrated into the EHR to reduce delays in diagnosis. However, model validation in other hospital-based populations is necessary.

Association Between Serum Vitamin D Levels and Myopia in the National Health and Nutrition Examination Survey (2001-2006).

PURPOSE: To assess the relationship between serum vitamin D levels and myopia in people aged 12-50 years using the National Health and Nutrition Examination Survey (NHANES) database. METHODS: Demographics, vision, and serum vitamin D levels from NHANES (2001-2006) were analyzed. Multivariate analyses were performed to examine the relationship between serum vitamin D levels and myopia while controlling for sex, age, ethnicity, education level, serum vitamin A, and poverty status. The main outcome was presence or absence of myopia, defined as a spherical equivalent of -1 diopters or more. RESULTS: Of the 11669 participants, 5,310 (45.5%) had myopia. The average serum vitamin D concentration was 61.6 ± 0.9 nmol/L for the myopic group and 63.1 ± 0.8 nmol/L for the non-myopic group (p = .01). After adjusting for all covariates, having higher serum vitamin D was associated with lower odds of having myopia (odds ratio 0.82 [0.74-0.92], p = .0007). In linear regression modeling that excluded hyperopes (spherical equivalent > +1 diopters), there was a positive relationship between spherical equivalent and serum vitamin D levels. Specifically, as serum vitamin D doubled, spherical equivalent increased by 0.17 (p = .02) indicating a positive dose-response relationship between vitamin D and myopia. CONCLUSIONS: Participants with myopia, on average, had lower serum concentrations of vitamin D compared to those without myopia. While further studies are needed to determine the mechanism, this study suggests that higher vitamin D levels are associated with lower incidence of myopia.

Disease progression pathways of wet AMD: opportunities for new target discovery.

INTRODUCTION: Age-related macular degeneration (AMD) is the leading cause of irreversible blindness among people age 60 years or older in developed countries. Current standard-of-care anti-vascular endothelial growth factor (VEGF) therapy, which inhibits angiogenesis and vascular permeability, has been shown to stabilize choroidal neovascularization and increase visual acuity in neovascular AMD. However, therapeutic limitations of anti-VEGF therapy include limited durability with consequent need for frequent intravitreal injections, and a ceiling of efficacy. Current strategies under investigation include targeting VEGF-C and VEGF-D, integrins, tyrosine kinase receptors, and the Tie2/angiopoietin-2 pathway. A literature search was conducted through November 30, 2021 on PubMed, Medline, Google Scholar, and associated digital platforms with the following keywords: wet macular degeneration, age-related macular degeneration, therapy, VEGF-A, VEGF-C, VEGF-D, integrins, Tie2/Ang2, and tyrosine kinase inhibitors. AREAS COVERED: The authors provide a comprehensive review of AMD disease pathways and mechanisms involved in wet AMD as well as novel targets for future therapies. EXPERT OPINION: With novel targets and advancements in drug delivery, there is potential to address treatment burden and to improve outcomes for patients afflicted with neovascular AMD.

Vascular biomarkers from optical coherence tomography angiography and glaucoma: where do we stand in 2021?

Biomarkers of ocular blood flow originating from a wide variety of imaging modalities have been associated with glaucoma onset and progression for many decades. Advancements in imaging platforms including optical coherence tomography angiography (OCTA) have provided the ability to quantify vascular changes in glaucoma patients, alongside traditional measures such as retinal nerve fibre layer thickness and optic nerve head structure. Current literature on vascular biomarkers, as measured by OCTA, indicates significant relationships between glaucoma and blood flow and capillary density in the retina and ONH. The data currently available, however, are highly diverse and lack robust longitudinal data on OCTA vascular outcomes and glaucoma progression. Herein we discuss and summarize the relevant current literature on OCTA vascular biomarkers and glaucoma reviewed through March 1, 2021. Associations between OCTA vascular biomarkers and clinical structural and functional glaucoma outcomes as well as differences between glaucoma patients and healthy controls are reviewed and summarized. The available data identify significantly decreased flow density, flow index and vessel density in the ONH, peripapillary vascular layer and macula of glaucoma patients compared with controls. Whole image vessel density is also significantly decreased in glaucoma patients compared with controls, and this outcome has been found to correspond to severity of visual field loss. OCTA vascular biomarkers alongside clinical structural outcomes may aid in assessing overall risk for glaucoma in patients.