Purpose The purpose of this study was to summarize and evaluate artificial intelligence (AI) algorithms used in geographic atrophy (GA) diagnostic processes (e.g. isolating lesions or disease progression). Methods The search strategy and selection of publications were both conducted in accordance with the Preferred of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. PubMed and Web of Science were used to extract literary data. The algorithms were summarized by objective, performance, and scope of coverage of GA diagnosis (e.g. lesion automation and GA progression). Results Twenty-seven studies were identified for this review. A total of 18 publications focused on lesion segmentation only, 2 were designed to detect and classify GA, 2 were designed to predict future overall GA progression, 3 focused on prediction of future spatial GA progression, and 2 focused on prediction of visual function in GA. GA-related algorithms reported sensitivities from 0.47 to 0.98, specificities from 0.73 to 0.99, accuracies from 0.42 to 0.995, and Dice coefficients from 0.66 to 0.89. Conclusions Current GA-AI publications have a predominant focus on lesion segmentation and a minor focus on classification and progression analysis. AI could be applied to other facets of GA diagnoses, such as understanding the role of hyperfluorescent areas in GA. Using AI for GA has several advantages, including improved diagnostic accuracy and faster processing speeds. Translational Relevance AI can be used to quantify GA lesions and therefore allows one to impute visual function and quality-of-life. However, there is a need for the development of reliable and objective models and software to predict the rate of GA progression and to quantify improvements due to interventions.
Background: Identification and treatment of malnutrition are essential in upper gastrointestinal (UGI) cancer. However, there is limited understanding of the nutritional status of UGI cancer patients at the time of curative surgery. This prospective point prevalence study involving 27 Australian tertiary hospitals investigated nutritional status at the time of curative UGI cancer resection, as well as presence of preoperative nutrition impact symptoms, and associations with length of stay (LOS) and surgical complications. Methods: Subjective global assessment, hand grip strength (HGS) and weight were performed within 7 days of admission. Data on preoperative weight changes, nutrition impact symptoms, and dietary intake were collected using a purpose-built data collection tool. Surgical LOS and complications were also recorded. Multivariate regression models were developed for nutritional status, unintentional weight loss, LOS and complications. Results: This study included 200 patients undergoing oesophageal, gastric and pancreatic surgery. Malnutrition prevalence was 42% (95% confidence interval (CI) 35%, 49%), 49% lost ≥5% weight in 6 months, and 47% of those who completed HGS assessment had low muscle strength with no differences between surgical procedures (p = 0.864, p = 0.943, p = 0.075, respectively). The overall prevalence of reporting at least one preoperative nutrition impact symptom was 55%, with poor appetite (37%) and early satiety (23%) the most frequently reported. Age (odds ratio (OR) 4.1, 95% CI 1.5, 11.5, p = 0.008), unintentional weight loss of ≥5% in 6 months (OR 28.7, 95% CI 10.5, 78.6, p < 0.001), vomiting (OR 17.1, 95% CI 1.4, 207.8, 0.025), reduced food intake lasting 2–4 weeks (OR 7.4, 95% CI 1.3, 43.5, p = 0.026) and ≥1 month (OR 7.7, 95% CI 2.7, 22.0, p < 0.001) were independently associated with preoperative malnutrition. Factors independently associated with unintentional weight loss were poor appetite (OR 3.7, 95% CI 1.6, 8.4, p = 0.002) and degree of solid food reduction of <75% (OR 3.3, 95% CI 1.2, 9.2, p = 0.02) and <50% (OR 4.9, 95% CI 1.5, 15.6, p = 0.008) of usual intake. Malnutrition (regression coefficient 3.6, 95% CI 0.1, 7.2, p = 0.048) and unintentional weight loss (regression coefficient 4.1, 95% CI 0.5, 7.6, p = 0.026) were independently associated with LOS, but no associations were found for complications. Conclusions: Despite increasing recognition of the importance of preoperative nutritional intervention, a high proportion of patients present with malnutrition or clinically significant weight loss, which are associated with increased LOS. Factors associated with malnutrition and weight loss should be incorporated into routine preoperative screening. Further investigation is required of current practice for dietetics interventions received prior to UGI surgery and if this mitigates the impact on clinical outcomes.
The neutrophil–lymphocyte ratio (NLR) and platelet–lymphocyte ratio (PLR) are emerging haematological inflammatory biomarkers. However, their significance in retinal vein occlusion (RVO) and its subtypes, branch and central RVO (BRVO and CRVO, respectively), is uncertain. This systematic review and meta‐analysis aimed to clarify the association of NLR and PLR with RVO. We searched MEDLINE (Ovid), EMBASE (Ovid) and the Cochrane Library for studies investigating the association of NLR and PLR with RVO from inception to 2 December 2020. We used random‐effects inverse‐variance modelling to generate pooled effect measures. We used bivariate Bayesian modelling to meta‐analyse the ability of NLR and PLR to differ between individuals with and without RVO and performed meta‐regression and sensitivity analyses to explore inter‐study heterogeneity. Eight studies published encompassing 1059 patients were included for analysis. Both NLR and PLR were significantly elevated in RVO, with pooled mean differences of 0.63 (95% confidence interval (CI) 0.31–0.95) and 21.49 (95% CI 10.03–32.95), respectively. The pooled sensitivity, specificity and area under the Bayesian summary receiver operating characteristic curve were, respectively, 0.629 (95% credible interval (CrI) 0.284–0.872), 0.731 (95% CrI 0.373–0.934) and 0.688 (95% CrI 0.358–0.872) for NLR; and 0.645 (95% CrI 0.456–0.779), 0.616 (95% CrI 0.428–0.761) and 0.621 (95% CrI 0.452–0.741) for PLR. Mean and variability of age and diabetes mellitus prevalence partially explained between‐study heterogeneity. NLR and PLR are significantly elevated in RVO. Future research is needed to investigate the potential prognostic value and independence of these findings.
Background: Preoperative nutrition intervention is recommended prior to upper gastrointestinal (UGI) cancer resection; however, there is limited understanding of interventions received in current clinical practice. This study investigated type and frequency of preoperative dietetics intervention and nutrition support received and clinical and demographic factors associated with receipt of intervention. Associations between intervention and preoperative weight loss, surgical length of stay (LOS), and complications were also investigated. Methods: The NOURISH Point Prevalence Study was conducted between September 2019 and May 2020 across 27 Australian tertiary centres. Subjective global assessment and weight were performed within 7 days of admission. Patients reported on preoperative dietetics and nutrition intervention, and surgical LOS and complications were recorded. Results: Two-hundred patients participated (59% male, mean (standard deviation) age 67 (10)). Sixty percent had seen a dietitian preoperatively, whilst 50% were receiving nutrition support (92% oral nutrition support (ONS)). Patients undergoing pancreatic surgery were less likely to receive dietetics intervention and nutrition support than oesophageal or gastric surgeries (p < 0.001 and p = 0.029, respectively). Neoadjuvant therapy (p = 0.003) and malnutrition (p = 0.046) remained independently associated with receiving dietetics intervention; however, 31.3% of malnourished patients had not seen a dietitian. Patients who received ≥3 dietetics appointments had lower mean (SD) percentage weight loss at the 1-month preoperative timeframe compared with patients who received 0–2 appointments (1.2 (2.0) vs. 3.1 (3.3), p = 0.001). Patients who received ONS for >2 weeks had lower mean (SD) percentage weight loss than those who did not (1.2 (1.8) vs. 2.9 (3.4), p = 0.001). In malnourished patients, total dietetics appointments ≥3 was independently associated with reduced surgical complications (odds ratio 0.2, 95% confidence interval (CI) 0.1, 0.9, p = 0.04), and ONS >2 weeks was associated with reduced LOS (regression coefficient −7.3, 95% CI −14.3, −0.3, p = 0.04). Conclusions: Despite recommendations, there are low rates of preoperative dietetics consultation and nutrition support in this population, which are associated with increased preoperative weight loss and risk of increased LOS and complications in malnourished patients. The results of this study provide insights into evidence–practice gaps for improvement and data to support further research regarding optimal methods of preoperative nutrition support.
Purpose: Globally, an estimated 12.7 million people await a corneal transplant. Of these, 53% are without routine access to a domestic supply and are reliant on transnational activity (TNA) (importation) of corneal tissue (CT) for transplantation. Although CT TNA commenced in 1961, there has been no evaluation of its impact on import and export nations. Methods: We wished to examine the impact of clinical and nonclinical CT TNA on export and import nations, with nonclinical aspects our primary focus, to help guide future practice. We conducted a review of the academic literature through various search engines. We prefix and place our review in the relevant historical practice and global context. Results: Despite commencement in 1961, we only located 14 studies (11 clinical and 3 nonclinical) pertaining to CT TNA. These were published between 1991 and 2018. Clinical papers reported death-to-preservation time, preservation-to-transplantation time, logistics, donor and recipient selection, and quality as relevant. Nonclinical studies identified emerging themes pertaining to financial, ethical, and sustainability aspects of TNA. Conclusions: All aspects of CT TNA are grossly under-reported, resulting in our inability to effectively analyze the overall impact to export and import nations. The few clinical studies in our review concluded that despite endothelial cell loss and other risk factors, imported CT appears comparable with domestic CT and remains an option in the absence of domestic supply. Nonclinical aspects (eg, ethical, equitable, and economic) have also not been adequately addressed.
This study describes the development of a deep learning algorithm based on the U-Net architecture for automated segmentation of geographic atrophy (GA) lesions in fundus autofluorescence (FAF) images. Methods:Image preprocessing and normalization by modified adaptive histogram equalization were used for image standardization to improve effectiveness of deep learning. A U-Net-based deep learning algorithm was developed and trained and tested by fivefold cross-validation using FAF images from clinical datasets. The following metrics were used for evaluating the performance for lesion segmentation in GA: dice similarity coefficient (DSC), DSC loss, sensitivity, specificity, mean absolute error (MAE), accuracy, recall, and precision. Results:In total, 702 FAF images from 51 patients were analyzed. After fivefold crossvalidation for lesion segmentation, the average training and validation scores were found for the most important metric, DSC (0.9874 and 0.9779), for accuracy (0.9912 and 0.9815), for sensitivity (0.9955 and 0.9928), and for specificity (0.8686 and 0.7261). Scores for testing were all similar to the validation scores. The algorithm segmented GA lesions six times more quickly than human performance. Conclusions:The deep learning algorithm can be implemented using clinical data with a very high level of performance for lesion segmentation. Automation of diagnostics for GA assessment has the potential to provide savings with respect to patient visit duration, operational cost and measurement reliability in routine GA assessments. Translational Relevance: A deep learning algorithm based on the U-Net architecture and image preprocessing appears to be suitable for automated segmentation of GA lesions on clinical data, producing fast and accurate results.
Objective To describe the demographics and outcomes of sports‐related ocular injuries in an Australian tertiary eye hospital setting. Methods Retrospective descriptive study from the Royal Victorian Eye and Ear Hospital from 2015 to 2020. Patient demographics, diagnosis and injury causation were recorded from baseline and follow‐up. Outcomes included visual acuity (VA), intraocular pressure (IOP), ocular injury diagnosis, investigations and management performed. Results A total of 1793 individuals (mean age 28.67 ± 15.65 years; 80.42% males and 19.58% females) presented with sports‐related ocular trauma. The top three injury‐causing sports were soccer (n = 327, 18.24%), Australian rules football (AFL) (n = 306, 17.07%) and basketball (n = 215, 11.99%). The top injury mechanisms were projectile (n = 976, 54.43%) and incidental body contact (n = 506, 28.22%). The most frequent diagnosis was traumatic hyphaema (n = 725). Best documented VA was ≥6/12 at baseline in 84.8% and at follow‐up in 95.0% of cases. The greatest risk of globe rupture/penetration was associated with martial arts (odds ratio [OR] 16.22); orbital blow‐out fracture with skiing (OR 14.42); and hyphaema with squash (OR 4.18): P < 0.05 for all. Topical steroids were the most common treatment (n = 693, 38.7%). Computed tomography orbits/facial bones were the most common investigation (n = 184, 10.3%). The mean IOP was 16.1 mmHg; 103 (5.7%) cases required topical anti‐ocular hypertensives. Twenty‐six individuals (1.45%) required surgery with AFL contributing the most surgical cases (n = 5, 19.23%). Conclusion The top three ocular injury causing sports were soccer, AFL and basketball. The most frequent injury was traumatic hyphaema. Projectiles posed the greatest risk.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.