New-onset keratitis associated with epidermal growth factor receptor-based targeted therapies in Han Chinese patients with lung cancer: A multi-center cohort study.

PURPOSE: To determine the risk and incidence of keratitis following treatment with epidermal growth factor receptor inhibitors (EGFRi) and subtypes of EGFRi-associated keratitis. METHODS: This multi-center cohort study included EGFRi-treated patients and non-users with lung cancer between 2010 and 2023. EGFRi included first-generation agent gefitinib and erlotinib, second-generation agent afatinib, and third-generation agent osimertinib. The primary outcome was new-onset keratitis. Cox proportional hazard models with multivariable adjustment were applied to determine the effect of EGFRi on keratitis over time. Subgroup analyses were conducted, stratified by agents of EGFRi. Sub-outcome analyses were performed to identify the subtypes of EGFRi-associated keratitis. RESULTS: A total of 1549 EGFRi-treated patients and 6146 non-users were included. 38 (2.5%) EGFRi-treated patients developed keratitis. The incidence of keratitis in EGFRi-treated patients was significantly higher than that in controls (incidence rate, IR, per 1000 person-years = 14.7 vs 4.49, p < 0.0001). EGFRi-treated patients presented with an increased risk for keratitis (adjusted hazard ratio, aHR = 3.14, 95% CI = 1.85-5.35, p < 0.001). Erlotinib (aHR = 2.64, 95% CI = 1.35-5.15, p = 0.004), afatinib (aHR = 4.42, 95% CI = 2.17-9.02, p < 0.001), and osimertinib (aHR = 4.67, 95% CI = 1.60-13.64, p = 0.005), but not gefitinib (aHR = 2.30, 95% CI = 0.96-5.55, p = 0.063), significantly contributed to the risk of keratitis. Subtypes of EGFRi-associated keratitis included corneal ulcer (IR = 2.31 vs 0.166, p < 0.0001) and keratoconjunctivitis (IR = 9.27 vs 2.91, p < 0.0001). None of the EGFRi-treated patients developed perforated corneal ulcer, interstitial and deep keratitis, or corneal neovascularization. CONCLUSION: Treatment with EGFRi was associated with an increased risk of keratitis. Ocular toxicity of EGFRi was highest for third-generation agents, followed by second-generation agents, and then first-generation agents.

Allergic rhinitis and keratoconus: a systematic review and meta-analysis.

BACKGROUND: The relationship between keratoconus and various allergic diseases has been a subject of controversy. OBJECTIVE: In the present study, a systematic review and meta-analysis was conducted to investigate the association between allergic rhinitis (AR) and keratoconus. METHODS: Relevant and eligible studies from PubMed, Web of Science, and Cochrane Library were systematically reviewed to evaluate the association between AR and keratoconus. Observational studies containing the number of patients with and without keratoconus and the number of patients with keratoconus diagnosed with or without AR were included. Two reviewers independently screened for eligible studies and extracted data from the included studies. A bivariate meta-analysis was conducted to compare the odds of keratoconus occurrence in patients with and without AR. The main outcome was the odds ratio of keratoconus occurrence in patients with AR. A sensitivity test was performed using the adjusted odds ratio reported in the included studies to validate the findings. RESULTS: Seven studies involving 775,574 participants were included in this meta-analysis. Among them, 29,082 patients had keratoconus. The pooled odds ratio of keratoconus occurrence in patients with AR was 1.71 (95% confidence interval [CI]: 1.36-2.15; P < 0.001; I2 = 96%), and the pooled adjusted odds ratio was 1.72 (95% CI: 1.23-2.40; P = 0.001; I2 = 97%). CONCLUSION: Patients with AR showed significantly higher odds of keratoconus occurrence than those without AR. Future studies are warranted to investigate the causal relationship and evaluate the cost-effectiveness of early screening using methods such as corneal topography and referral for keratoconus in patients with AR.

Impact of different corticosteroids on severe community-acquired pneumonia: a systematic review and meta-analysis.

OBJECTIVES: Randomised controlled trials (RCTs) have demonstrated conflicting results regarding the effects of corticosteroids on the treatment of severe community-acquired pneumonia (CAP). We aimed to investigate the efficacy and safety of different corticosteroids on patients who were hospitalised for severe CAP. METHODS: We performed a systematic search through PubMed, Embase, Cochrane Central Register of Controlled Trials, Web of Science, and Scopus from inception to May 2023. The primary outcome was all-cause mortality. Data analysis was performed using a random-effects model. RESULTS: A total of 10 RCTs comprising 1962 patients were included. Corticosteroids were associated with a lower rate of all-cause mortality (risk ratio (RR), 0.70 (95% CI 0.54 to 0.90); I2=0.00%). When stratified into different corticosteroid types, hydrocortisone was associated with an approximately 50% lower mortality risk (RR, 0.48 (95% CI 0.32 to 0.72); I2=0.00%). However, dexamethasone, methylprednisolone or prednisolone were not associated with an improvement in mortality. Furthermore, hydrocortisone was associated with a reduction in the rate of mechanical ventilation, acute respiratory distress syndrome, shock and duration of intensive care unit stay. These trends were not observed for dexamethasone, methylprednisolone or prednisolone. Corticosteroids were not associated with an increased risk of adverse events including gastrointestinal bleeding, secondary infection or hyperglycaemia. CONCLUSIONS: The use of hydrocortisone, but not other types of corticosteroids, was associated with a reduction in mortality and improvement in pneumonia outcomes among patients hospitalised with severe CAP.PROSPERO registration numberCRD42023431360.

Genotypes Influence Clinical Progression in EYS-Associated Retinitis Pigmentosa.

Purpose: The purpose of this study was to investigate the genetic and clinical characteristics of eyes shut homolog (EYS)-associated retinitis pigmentosa (RP). Methods: This was a retrospective cross-sectional observational study of 36 patients with EYS-associated autosomal recessive RP (arRP). Results: The gene sequencing results revealed that c.6416G>A (p.Cys2139Tyr) and c.7228+1G>A were the two most predominant variants in our cohort and that variants near the C-terminus, which contains alternating laminin and epidermal growth factor (EGF) domains, accounted for the majority of the allele counts (58 of a total of 72) and relative allele frequencies (81%). Over half of the patients presented with pericentral-type RP (n = 19, 60%), which frequently occurred in combination with macular lesions (n = 10, 52%). Patients having both variants within the alternating laminin and EGF domains near the C-terminus had a more severe disease progression (average 0.045 logMAR increase per year) than those having one variant in the N-terminus and the other in the C-terminus (average 0.001 logMAR increase per year). Conclusions: Pericentral RP was the major phenotype in patients with EYS-associated arRP. There was also a statistically significant relationship between the location of the variants and the severity of the disease. Translational Relevance: This study may aid patients with EYS-associated arRP to predict future vision acuity based on their genetic and clinical features.

Data Homogeneity Effect in Deep Learning-Based Prediction of Type 1 Diabetic Retinopathy.

This study is aimed at evaluating a deep transfer learning-based model for identifying diabetic retinopathy (DR) that was trained using a dataset with high variability and predominant type 2 diabetes (T2D) and comparing model performance with that in patients with type 1 diabetes (T1D). The Kaggle dataset, which is a publicly available dataset, was divided into training and testing Kaggle datasets. In the comparison dataset, we collected retinal fundus images of T1D patients at Chang Gung Memorial Hospital in Taiwan from 2013 to 2020, and the images were divided into training and testing T1D datasets. The model was developed using 4 different convolutional neural networks (Inception-V3, DenseNet-121, VGG1, and Xception). The model performance in predicting DR was evaluated using testing images from each dataset, and area under the curve (AUC), sensitivity, and specificity were calculated. The model trained using the Kaggle dataset had an average (range) AUC of 0.74 (0.03) and 0.87 (0.01) in the testing Kaggle and T1D datasets, respectively. The model trained using the T1D dataset had an AUC of 0.88 (0.03), which decreased to 0.57 (0.02) in the testing Kaggle dataset. Heatmaps showed that the model focused on retinal hemorrhage, vessels, and exudation to predict DR. In wrong prediction images, artifacts and low-image quality affected model performance. The model developed with the high variability and T2D predominant dataset could be applied to T1D patients. Dataset homogeneity could affect the performance, trainability, and generalization of the model.