Locality is the strongest predictor of expert performance in image-based differentiation of bacterial and fungal corneal ulcers from India.

PURPOSE: This study sought to identify the sources of differential performance and misclassification error among local (Indian) and external (non-Indian) corneal specialists in identifying bacterial and fungal keratitis based on corneal photography. METHODS: This study is a secondary analysis of survey data assessing the ability of corneal specialists to identify acute bacterial versus fungal keratitis by using corneal photography. One-hundred images of 100 eyes from 100 patients with acute bacterial or fungal keratitis in South India were previously presented to an international cohort of cornea specialists for interpretation over the span of April to July 2021. Each expert provided a predicted probability that the ulcer was either bacterial or fungal. Using these data, we performed multivariable linear regression to identify factors predictive of expert performance, accounting for primary practice location and surrogate measures to infer local fungal ulcer prevalence, including locality, latitude, and dew point. In addition, Brier score decomposition was used to determine experts' reliability ("calibration") and resolution ("boldness") and were compared between local (Indian) and external (non-Indian) experts. RESULTS: Sixty-six experts from 16 countries participated. Indian practice location was the only independently significant predictor of performance in multivariable linear regression. Resolution among Indian experts was significantly better (0.08) than among non-Indian experts (0.01; P < 0.001), indicating greater confidence in their predictions. There was no significant difference in reliability between the two groups ( P = 0.40). CONCLUSION: Local cornea experts outperformed their international counterparts independent of regional variability in tropical risk factors for fungal keratitis. This may be explained by regional characteristics of infectious ulcers with which local corneal specialists are familiar.

Validation of the C-DU(KE) Calculator as a Predictor of Outcomes in Patients Enrolled in Steroids for Corneal Ulcer and Mycotic Ulcer Treatment Trials.

PURPOSE: The aim of this study was to validate the C-DU(KE) calculator as a predictor of treatment outcomes on a data set derived from patients with culture-positive ulcers. METHODS: C-DU(KE) criteria were compiled from a data set consisting of 1063 cases of infectious keratitis from the Steroids for Corneal Ulcer Trial (SCUT) and Mycotic Ulcer Treatment Trial (MUTT) studies. These criteria include corticosteroid use after symptoms, visual acuity, ulcer area, fungal etiology, and elapsed time to organism-sensitive therapy. Univariate analysis was performed followed by multivariable logistic regressions on culture-exclusive and culture-inclusive models to assess for associations between the variables and outcome. The predictive probability of treatment failure, defined as the need for surgical intervention, was calculated for each study participant. Discrimination was assessed using the area under the curve for each model. RESULTS: Overall, 17.9% of SCUT/MUTT participants required surgical intervention. Univariate analysis showed that decreased visual acuity, larger ulcer area, and fungal etiology had a significant association with failed medical management. The other 2 criteria did not. In the culture-exclusive model, 2 of 3 criteria, decreased vision [odds ratio (OR) = 3.13, P < 0.001] and increased ulcer area (OR = 1.03, P < 0.001), affected outcomes. In the culture-inclusive model, 3 of 5 criteria, decreased vision (OR = 4.9, P < 0.001), ulcer area (OR = 1.02, P < 0.001), and fungal etiology (OR = 9.8, P < 0.001), affected results. The area under the curves were 0.784 for the culture-exclusive model and 0.846 for the culture-inclusive model which were comparable to the original study. CONCLUSIONS: The C-DU(KE) calculator is generalizable to a study population from large international studies primarily taking place in India. These results support its use as a risk stratification tool assisting ophthalmologists in patient management.

Assessing the Impact of Image Quality on Deep Learning Classification of Infectious Keratitis.

Objective: To investigate the impact of corneal photograph quality on convolutional neural network (CNN) predictions. Design: A CNN trained to classify bacterial and fungal keratitis was evaluated using photographs of ulcers labeled according to 5 corneal image quality parameters: eccentric gaze direction, abnormal eyelid position, over/under-exposure, inadequate focus, and malpositioned light reflection. Participants: All eligible subjects with culture and stain-proven bacterial and/or fungal ulcers presenting to Aravind Eye Hospital in Madurai, India, between January 1, 2021 and December 31, 2021. Methods: Convolutional neural network classification performance was compared for each quality parameter, and gradient class activation heatmaps were generated to visualize regions of highest influence on CNN predictions. Main Outcome Measures: Area under the receiver operating characteristic and precision recall curves were calculated to quantify model performance. Bootstrapped confidence intervals were used for statistical comparisons. Logistic loss was calculated to measure individual prediction accuracy. Results: Individual presence of either light reflection or eyelids obscuring the corneal surface was associated with significantly higher CNN performance. No other quality parameter significantly influenced CNN performance. Qualitative review of gradient class activation heatmaps generally revealed the infiltrate as having the highest diagnostic relevance. Conclusions: The CNN demonstrated expert-level performance regardless of image quality. Future studies may investigate use of smartphone cameras and image sets with greater variance in image quality to further explore the influence of these parameters on model performance. Financial Disclosures: Proprietary or commercial disclosure may be found after the references.

Detection of microRNAs expression signatures in vitreous humor of intraocular tuberculosis.

BACKGROUND: MicroRNA (miRNA) expression analysis has been shown to provide them as biomarkers in several eye diseases and has a regulatory role in pathogenesis. However, miRNA expression analysis in the vitreous humor (VH) of intraocular tuberculosis (IOTB) is not studied. Thus, we aim to find miRNA expression signatures in the VH of IOTB patients to identify their regulatory role in disease pathogenesis and to find them as potential biomarkers for IOTB. METHODS AND RESULTS: First, we profiled miRNAs in VH of three IOTB and three Macular hole (MH) samples as controls through small-RNA deep sequencing using Illumina Platform. In-house bioinformatics analysis identified 81 dysregulated miRNAs in IOTB. Further validation in VH of IOTB (n = 15) compared to MH (n = 15) using Real-Time quantitative PCR (RT-qPCR) identified three significantly upregulated miRNAs, hsa-miR-150-5p, hsa-miR-26b-5p, and hsa-miR-21-5p. Based on the miRNA target prediction, functional network analysis, and RT-qPCR analysis of target genes, the three miRNAs downregulating WNT5A, PRKCA, MAP3K7, IL7, TGFB2, IL1A, PRKCB, TNFA, and TP53 genes involving MAPK signaling pathway, PI3K-AKT signaling pathway, WNT signaling pathway, Cell cycle, TGF-beta signaling pathway, Long-term potentiation, and Sphingolipid signaling pathways, have a potential role in disease pathogenesis. The ROC analysis of RT-qPCR data showed that hsa-miR-150-5p with AUC = 0.715, hsa-miR-21-5p with AUC = 0.789, and hsa-miR-26b-5p with AUC = 0.738; however, the combination of hsa-miR-21-5p and hsa-miR-26b-5p with AUC = 0.796 could serve as a potential biomarker for IOTB. CONCLUSIONS: This study provides the first report on miRNA expression signatures detected in VH for IOTB pathogenesis and also provides a potential biomarker for IOTB.

Host cell-type and pathogen-specific immunomodulatory functions of macrophage migration inhibitory factor (MIF) in infectious keratitis.

Therapeutic management of inflammation in infectious keratitis (IK) requires new strategy and targets for selective immunomodulation. Targeting host cell-type specific inflammatory responses might be a viable strategy to curtail unnecessary inflammation and reduce tissue damage without affecting pathogen clearance. This study explores the possibility of pathogen and host cell-type dependent differences in the inflammatory pathways relevant in the pathogenesis of IK. Human corneal epithelial cell line (HCEC) and phorbol 12-myristate-13 acetate (PMA) differentiated THP-1 macrophage line were infected with either Aspergillus flavus conidia or Acanthamoeba castellanii trophozoites and the elicited inflammatory responses were studied in terms of gene expression and secretion of proinflammatory factors interleukin-8 (IL-8) and tumor necrosis factor-alpha (TNF-α) and an upstream inflammatory regulator and mediator protein-the Macrophage Migration Inhibitory Factor (MIF). Given the pleotropic mode of MIF function in diverse cell types relevant in many human diseases, we tested if MIF driven responses to infection is different in HCECs and THP-1 macrophages by studying its expression, secretion and involvement in inflammation by siRNA mediated knockdown. We also examined IK patient tear samples for MIF levels. Infection with A. flavus or A. castellanii induced IL-8 and TNF-α responses in HCECs and THP-1 macrophages but to different levels. Our preliminary human data showed that the level of secreted MIF protein was elevated in IK patient tear, however, MIF secretion by the two cell types were strikingly different in-vitro, under both normal and infected conditions. We found that HCECs released MIF constitutively, which was significantly inhibited with infection, whereas THP-1 macrophages were stimulated to release MIF during infection. MIF gene expression remained largely unaffected by infection in both the cell lines. Although MIF in HCECs appeared to be intracellularly captured during infection, MIF knockdown in HCECs associated with a partial reduction of the IL-8 and TNF-α expression produced by either of the pathogens, suggesting a pro-inflammatory role for MIF in HCECs, independent of its canonical cytokine like function. In contrast, MIF knockdown in THP-1 macrophages accompanied a dramatic increase in IL-8 and TNF-α expression during A. castellanii infection, while the responses to A. flavus infection remained unchanged. These data imply a host cell-type and pathogen specific distinction in the MIF- related inflammatory signaling and MIF as a potential selective immunomodulatory target in infectious keratitis.

Rapid Point-of-Care Identification of Aspergillus Species in Microbial Keratitis.

Importance: Microbial keratitis (MK) is a common cause of unilateral visual impairment, blindness, and eye loss in low-income and middle-income countries. There is an urgent need to develop and implement rapid and simple point-of-care diagnostics for MK to increase the likelihood of good outcomes. Objective: To evaluate the diagnostic performance of the Aspergillus-specific lateral-flow device (AspLFD) to identify Aspergillus species causing MK in corneal scrape and corneal swab samples of patients presenting with microbial keratitis. Design, Setting, and Participants: This diagnostic study was conducted between May 2022 and January 2023 at the corneal clinic of Aravind Eye Hospital in Madurai, Tamil Nadu, India. All study participants were recruited during their first presentation to the clinic. Patients aged 15 years or older met the eligibility criteria if they were attending their first appointment, had a corneal ulcer that was suggestive of a bacterial or fungal infection, and were about to undergo diagnostic scrape and culture. Main Outcomes and Measures: Sensitivity and specificity of the AspLFD with corneal samples collected from patients with MK. During routine diagnostic scraping, a minimally invasive corneal swab and an additional corneal scrape were collected and transferred to aliquots of sample buffer and analyzed by lateral-flow device (LFD) if the patient met the inclusion criteria. Photographs of devices were taken with a smartphone and analyzed using a ratiometric approach, which was developed for this study. The AspLFD results were compared with culture reports. Results: The 198 participants who met the inclusion criteria had a mean (range) age of 51 (15-85) years and included 126 males (63.6%). Overall, 35 of 198 participants with corneal scrape (17.7%) and 17 of 40 participants with swab samples (42.5%) had positive culture results for Aspergillus species. Ratiometric analysis results for the scrape samples found that the AspLFD achieved high sensitivity (0.89; 95% CI, 0.74-0.95), high negative predictive value (0.97; 95% CI, 0.94-0.99), low negative likelihood ratio (0.12; 95% CI, 0.05-0.30), and an accuracy of 0.94 (95% CI, 0.90-0.97). Ratiometric analysis results for the swab samples showed that the AspLFD had high sensitivity (0.94; 95% CI, 0.73-1.00), high negative predictive value (0.95; 95% CI, 0.76-1.00), low negative likelihood ratio (0.07; 95% CI, 0.01-0.48), and an accuracy of 0.88 (95% CI, 0.73-0.96). Conclusions and Relevance: Results of this diagnostic study suggest that AspLFD along with the ratiometric analysis of LFDs developed for this study has high diagnostic accuracy in identifying Aspergillus species from corneal scrapes and swabs. This technology is an important step toward the provision of point-of-care diagnostics for MK and could inform the clinical management strategy.

Microbiological Profile of Post-cataract Surgery Endophthalmitis with Usage of Prophylactic Intracameral Moxifloxacin.

PURPOSE: To study the effect of prophylactic intracameral moxifloxacin (ICM) on microbiological profile and antimicrobial sensitivity of culture-proven postoperative endophthalmitis (POE). METHODS: This study evaluated culture-proven POE diagnosed within 6 weeks of cataract surgery over two periods, period-1 (January 2010-March 2015) and period-2 (April 2015-December 2019), before and after introduction of prophylactic ICM, respectively. RESULTS: In period-1, 100 cases of culture-positive POE were reported (1 in 4879, 0.02%), while 20 cases (1 in 24635, 0.004%) were reported in period-2 (5-fold reduction, p < .001). The cumulative culture positivity rate decreased from 27.6% to 17.1% (1.6-fold) . Coagulase-negative Staphylococci (CoNS) were significantly reduced (p = .005). CONCLUSIONS: ICM reduced the incidence of culture-proven POE, CoNS in particular. In future, POE caused by rarer pathogens may become more prevalent due to reduction in the rate of commoner and more virulent pathogens with use of intracameral antibiotics. The moxifloxacin sensitivity of CONS did not show change with the use of ICM. Studies with larger population of POE will be more helpful to understand the trend better.

Deep Convolutional Neural Networks Detect no Morphological Differences Between Culture-Positive and Culture-Negative Infectious Keratitis Images.

Purpose: To determine whether convolutional neural networks can detect morphological differences between images of microbiologically positive and negative corneal ulcers. Methods: A cross-sectional comparison of prospectively collected data consisting of bacterial and fungal cultures and smears from eyes with acute infectious keratitis at Aravind Eye Hospital. Two convolutional neural network architectures (DenseNet and MobileNet) were trained using images obtained from handheld cameras collected from culture-positive and negative images and smear-positive and -negative images. Each architecture was trained on two image sets: (1) one with labels assigned using only culture results and (2) one using culture and smear results. The outcome measure was area under the receiver operating characteristic curve for predicting whether an ulcer would be microbiologically positive or negative. Results: There were 1970 images from 886 patients were included. None of the models were better than random chance at predicting positive microbiologic results (area under the receiver operating characteristic curve ranged from 0.49 to 0.56; all confidence intervals included 0.5). Conclusions: These two state-of-the-art deep convolutional neural network architectures could not reliably predict whether a corneal ulcer would be microbiologically positive or negative based on clinical photographs. This absence of detectable morphological differences informs the future development of computer vision models trained to predict the causative agent in infectious keratitis using corneal photography. Translational Relevance: These deep learning models were not able to identify morphological differences between microbiologically positive and negative corneal ulcers. This finding suggests that similar artificial intelligence models trained to identify the causative pathogen using only microbiologically positive cases may have potential to generalize well, including to cases with falsely negative microbiologic testing.

Image-Based Differentiation of Bacterial and Fungal Keratitis Using Deep Convolutional Neural Networks.

Purpose: Develop computer vision models for image-based differentiation of bacterial and fungal corneal ulcers and compare their performance against human experts. Design: Cross-sectional comparison of diagnostic performance. Participants: Patients with acute, culture-proven bacterial or fungal keratitis from 4 centers in South India. Methods: Five convolutional neural networks (CNNs) were trained using images from handheld cameras collected from patients with culture-proven corneal ulcers in South India recruited as part of clinical trials conducted between 2006 and 2015. Their performance was evaluated on 2 hold-out test sets (1 single center and 1 multicenter) from South India. Twelve local expert cornea specialists performed remote interpretation of the images in the multicenter test set to enable direct comparison against CNN performance. Main Outcome Measures: Area under the receiver operating characteristic curve (AUC) individually and for each group collectively (i.e., CNN ensemble and human ensemble). Results: The best-performing CNN architecture was MobileNet, which attained an AUC of 0.86 on the single-center test set (other CNNs range, 0.68-0.84) and 0.83 on the multicenter test set (other CNNs range, 0.75-0.83). Expert human AUCs on the multicenter test set ranged from 0.42 to 0.79. The CNN ensemble achieved a statistically significantly higher AUC (0.84) than the human ensemble (0.76; P < 0.01). CNNs showed relatively higher accuracy for fungal (81%) versus bacterial (75%) ulcers, whereas humans showed relatively higher accuracy for bacterial (88%) versus fungal (56%) ulcers. An ensemble of the best-performing CNN and best-performing human achieved the highest AUC of 0.87, although this was not statistically significantly higher than the best CNN (0.83; P = 0.17) or best human (0.79; P = 0.09). Conclusions: Computer vision models achieved superhuman performance in identifying the underlying infectious cause of corneal ulcers compared with cornea specialists. The best-performing model, MobileNet, attained an AUC of 0.83 to 0.86 without any additional clinical or historical information. These findings suggest the potential for future implementation of these models to enable earlier directed antimicrobial therapy in the management of infectious keratitis, which may improve visual outcomes. Additional studies are ongoing to incorporate clinical history and expert opinion into predictive models.

Outpatient human coronavirus associated conjunctivitis in India.

BACKGROUND: Viral conjunctivitis (pink eye) can be highly contagious and is of public health importance. There remains significant debate whether SARS-CoV-2 can present as a primary conjunctivitis. The aim of this study was to identify pathogens associated with outpatient infectious conjunctivitis during the COVID-19 Delta surge. METHODS: This prospective study was conducted in the spring and summer months of 2021. 106 patients with acute conjunctivitis who presented to the Aravind Eye Center in Madurai, India were included. One anterior nasal swab and one conjunctival swab of each eye were obtained for each enrolled patient. Samples were subsequently processed for unbiased metagenomic RNA deep sequencing (RNA-seq). Outcomes included clinical findings and codetection of other pathogens with SARS-CoV-2 in patients with conjunctivitis. RESULTS: Among the 13 patients identified with human coronavirus RNA fragments in their swabs, 6 patients had SARS-CoV-2 infection, 5 patients had coinfections of SARS-CoV-2 and human adenovirus (HAdV), 1 patient had a coinfection with human coronavirus OC43 and HAdV, and 1 patient had a coinfection of Vittaforma corneae and SARS-CoV-2. 30% had bilateral disease and symptoms on presentation. Petechial hemorrhage was noted in 33% of patients with SARS-CoV-2 infection. No patients with SARS-CoV-2 or SARS-CoV-2 and HAdV infections had subepithelial infiltrates on presentation. All patients denied systemic symptoms. CONCLUSIONS: Among the patients presented with conjunctivitis associated with human coronavirus infection, over 50% of the patients had co-infections with other circulating pathogens, suggesting the public-health importance of broad pathogen testing and surveillance in the outpatient conjunctivitis population.

Deep sequencing analysis of clinical samples from patients with acute infectious conjunctivitis during the COVID-19 delta surge in Madurai, India.

BACKGROUND: Seasonal outbreaks of infectious conjunctivitis remain a public health issue. Determination of outbreak etiologies in the context of a worldwide pandemic may provide useful information to guide public health strategies. The aim of this study was to identify pathogens associated with outpatient infectious conjunctivitis during the COVID-19 Delta surge. METHODS: This prospective study was conducted from April 2021 to September 2021. All outpatients presenting to the Aravind Eye Center (Madurai, India) with signs and symptoms consistent with acute infectious conjunctivitis were eligible. Three swabs were obtained from each participant: one from each conjunctiva and one from the anterior nares. Samples were processed for metagenomic RNA deep sequencing (RNA-seq). RESULTS: Samples from 106 study participants were sequenced. The most common presenting symptoms were tearing (86%) and itching (71%). Preauricular lymphadenopathy was present in 38% of participants. 20% of participants had close contacts with similar symptoms. Systemic symptoms such as coughing, runny nose, vomiting or diarrhea were uncommonly reported. 60% of all participants used some medicated eye drops upon enrollment. 75% of study participants demonstrated infection with human adenovirus D (HAdV-D). 11% of conjunctivitis was associated with SARS-CoV-2. 15% had no definitive pathogen detected. 8% of all participants had codetection of more than one pathogen on RNA-seq. CONCLUSIONS: During the COVID-19 Delta surge in India, HAdV-D was the most common pathogen associated with infectious conjunctivitis. SARS-CoV-2 was the second most common associated pathogen. Seasonal surveillance may be necessary for the determination of emerging and reemerging pathogens responsible for infectious conjunctivitis.

Prevalence and Molecular Characterization of Metallo ß-Lactamase Producing Gram-Negative Pathogens Causing Eye Infections.

Purpose: Metallo ß-lactamases (MßL) production is a worldwide problem, particularly in gram-negative bacteria. As scanty data is available on the prevalence of MBL, the present study is being undertaken to determine the prevalence, antibacterial sensitivity patterns, and molecular characterization of MßL associated resistant genes in gram-negative bacteria isolated from ocular infections. Material and Methods: At a tertiary eye care center in south India, 359 gram-negative pathogens, 200 isolates from eye infections, and 159 isolates from normal flora of the eye were studied. A gold standard microbiology method was used to identify the isolates. An antibiotic double disc synergy test and a combination disc test were used to detect MßL production. Multiplex PCR was used to investigate the molecular characteristics of the MßL encoding genes blaVIM, blaIMP, and blaNDM. Results: Of the 359 gram-negative bacterial pathogens, Pseudomonas aeruginosa 108 (30.1%) and Enterobacter agglomerans 46 (12.8%) were commonly isolated. High prevalence of P. aeruginosa 81% (17 strains) was detected as an MßL producer and it shows 100% resistance to 2nd and 3rd generation cephalosporins and meropenem. Multiplex PCR detected only the blaVIM gene in 56 (28%) of various eye infections and 27 (17%) of normal flora of the gram-negative bacteria (GNB). The blaVIM gene is detected predominantly in 51.8% of keratitis and 21.4% of postoperative endophthalmitis. High prevalence of the gene was detected in P. aeruginosa 42.9% (24 of 56) and Alcaligens denitrificans 10.7% (6 of 56) from eye infections. Whereas, in the control group, P. aeruginosa and E. coli each had 14.8% (4 of 27) that were shown positive. Conclusion: The emerging MßLs mediated resistance among P. aeruginosa is a challenging task for ophthalmologists, especially in patients with endophthalmitis and bacterial keratitis. This local knowledge will aid in advising appropriate antibiotic use and avoiding unnecessary antibiotic prescriptions, which are highly warranted.

Outcomes of amoebic, fungal, and bacterial keratitis: A retrospective cohort study.

BACKGROUND: Acanthamoeba keratitis is challenging to treat and thought to result in poor outcomes, but very few comparative studies exist to assess whether ulcers caused by Acanthamoeba are worse than those caused by bacteria or fungus. METHODS: In a retrospective cohort study, all cases of smear- or culture-proven Acanthamoeba keratitis diagnosed from January 2006 to June 2011 at an eye hospital in South India were identified from the microbiology database. Random samples of the same number of cases of bacterial and fungal keratitis, matched by year, were identified from the same database in order to compare outcomes between the three types of organism. The main outcomes were the time until the following events: re-epithelialization, discontinuation of antimicrobials, perforation/keratoplasty, elevated intraocular pressure, and new cataract. RESULTS: The median time until re-epithelialization was 113 days for Acanthamoeba keratitis, 30 days for fungal keratitis, and 25 days for bacterial keratitis, and the median time until discontinuation of antimicrobial therapy was 100 days for Acanthamoeba keratitis, 49 days for fungal keratitis, and 40 days for bacterial keratitis. Compared to the other two organisms, Acanthamoeba ulcers took significantly longer to re-epithelialize (adjusted HR 0.4, 95% CI 0.3 to 0.6 relative to bacterial ulcers and HR 0.3, 95% CI 0.2 to 0.5 relative to fungal ulcers; overall p<0.001) and had significantly longer courses of antimicrobials (adjusted HR 0.3, 95% CI 0.2 to 0.6 relative to bacterial ulcers and HR 0.5, 95%CI 0.3 to 0.8 relative to fungal ulcers; overall p<0.001). No statistically significant difference was observed between the three organisms for the other time-to-event outcomes. CONCLUSIONS: Acanthamoeba keratitis was more difficult to treat and had worse clinical outcomes than bacterial or fungal ulcers, highlighting the lack of adequate treatment regimens for this infection.

Patterns of Antifungal Resistance in Adult Patients With Fungal Keratitis in South India: A Post Hoc Analysis of 3 Randomized Clinical Trials.

IMPORTANCE: Antifungal resistance has been shown to impact treatment success, but research analyzing antifungal resistance is scarce. OBJECTIVE: To evaluate changes in antifungal resistance over time. DESIGN, SETTING, AND PARTICIPANTS: Ad hoc analysis of 3 randomized clinical trials including consecutive patients 18 years and older presenting with smear-positive fungal ulcers to Aravind Eye Hospitals in Madurai, Coimbatore, Pondicherry, and Tirunelveli in South India who participated in 1 of 3 clinical trials: the Mycotic Ulcer Treatment Trials (MUTT) I (2010 to 2011) or II (2010 to 2015) or the Cross-Linking Assisted Infection Reduction (CLAIR) trial (2016 to 2018). This post hoc analysis was designed in March 2021 and data were analyzed in May and November 2021. INTERVENTIONS: Minimum inhibitory concentration (MIC) of natamycin and voriconazole was determined from corneal cultures obtained using standardized methods outlined in the Clinical and Laboratory Standards Institute. MAIN OUTCOMES AND MEASURES: The primary outcome of this post hoc analysis was MIC of natamycin and voriconazole. RESULTS: A total of 890 fungal isolates were obtained from 651 patients (mean [SD] age, 49.6 [13.0]; 191 [43.3%] female) from 2010 to 2018. MICs were available for 522 samples in 446 patients. Fungal isolates overall demonstrated a 1.02-fold increase per year in voriconazole resistance as measured by MICs (95% CI, 1.00-1.04; P = .06). In subgroup analyses, Fusarium species demonstrated a 1.04-fold increase in voriconazole resistance per year (95% CI, 1.00-1.06; P = .01). Fungal isolates showed a 1.06-fold increase in natamycin resistance per year overall (95% CI, 1.03-1.09; P < .001). Fusarium species had a 1.06-fold increase in natamycin resistance (95% CI, 1.05-1.08; P < .001), Aspergillus had a 1.09-fold increase in resistance (95% CI, 1.05-1.15; P < .001), and other filamentous fungi had a 1.07-fold increase in resistance to natamycin per year (95% CI, 1.04-1.10; P < .001). CONCLUSIONS AND RELEVANCE: This post hoc analysis suggests that susceptibility to both natamycin and voriconazole may be decreasing over the last decade in South India. While a trend of increasing resistance could impact treatment of mycoses in general and infectious fungal keratitis in particular, further study is needed to confirm these findings and determine their generalizability to other regions of the world. TRIAL REGISTRATION: ClinicalTrials.gov Identifiers: NCT00996736 and NCT02570321.

Mediators of the Effect of Corneal Cross-Linking on Visual Acuity for Fungal Ulcers: A Prespecified Secondary Analysis From the Cross-Linking-Assisted Infection Reduction Trial.

PURPOSE: The purpose of this study was to determine whether astigmatism or corneal scarring is mediating the reduced visual acuity among patients with fungal keratitis randomized to corneal cross-linking (CXL). DESIGN: This was a prespecified exploratory outcome from an outcome-masked, 2 × 2 factorial design, randomized controlled clinical trial. STUDY PARTICIPANTS: Consecutive patients presented with moderate vision loss from a smear-positive fungal ulcer at Aravind Eye Hospital, Madurai, India. INTERVENTION: Study eyes were randomized to one of 4 treatment combinations using an adaptive randomization protocol. The treatment arms included (1) topical natamycin 5% alone, (2) topical natamycin 5% plus CXL, (3) topical amphotericin B 0.15% alone, and (4) topical amphotericin 0.15% plus CXL. Best spectacle-corrected visual acuity (BSCVA), contact lens over-refraction, and scar size and depth as measured by a masked study ophthalmologist using a standardized protocol were recorded at 3 months. Pentacam Scheimpflug imaging was also obtained at 3 months. MAIN OUTCOME MEASURES: BSCVA and contact lens over-refraction, infiltrate and/or scar size and depth, total astigmatism of the front and back of the cornea, total lower-order and higher-order aberrations of the anterior and posterior cornea, and total densitometry of the anterior, central, and posterior stroma were recorded. We performed a mediation analysis looking at the proportion of the effect of CXL on BSCVA that was mediated through scar size, scar depth, astigmatism and density. RESULTS: BSCVA at 3 months was available for 99 of 111 patients (89%) who had a mean of 0.82-LogMAR (SD 0.68). Three-month infiltrate and/or scar size ( P < 0.001), depth ( P < 0.001), and densitometry ( P = 0.001) were statistically significant predictors of 3-month BSCVA. Astigmatism seemed to mediate 23% of the effect of CXL on BSCVA, whereas scar size mediated 23%, scar depth 17%, and densitometry 7%. CONCLUSIONS: Corneal scarring and astigmatism are mediators of worse visual acuity after cross-linking in fungal keratitis. Corneal densitometry may be a helpful cornea-specific variable for clinicians and researchers in determining the effect of corneal scarring on visual acuity in specific patients and as an objective study outcome. TRIAL REGISTRATION: NCT02570321.

Outcomes of Pythium keratitis: a meta-analysis of individual patient data.

Background: Pythium keratitis is a difficult-to-treat corneal infection. Methods: A meta-analysis of individual patient data from observational studies of Pythium keratitis was performed. The outcomes of interest were therapeutic penetrating keratoplasty (TPK) and globe removal (evisceration, enucleation, or exenteration); the main exposures were linezolid and azithromycin use. Findings: Of 46 eligible articles, individual patient data were available for 306 eyes (34 studies). Pythium keratitis was associated with high rates of TPK (80%, 95%CI 70-87%) and globe removal (25%, 95%CI 13-43). In multivariable models adjusting for age and country, fewer TPKs were performed in patients treated with azithromycin (RR=0.80, 95%CI 0.67-0.96; P=0.04) and linezolid (RR=0.82, 95%CI 0.67-0.99; P=0.02). Conclusions: Studies of Pythium keratitis reported high rates of TPK and globe removal. Use of azithromycin and linezolid was associated with a lower rate of TPK. While promising, these results should be interpreted with caution given the biases inherent to observational studies.

Impact of Sample Collection Order on the Diagnostic Performance of Metagenomic Deep Sequencing for Infectious Keratitis.

PURPOSE: The purpose of this article was to evaluate the impact of sample collection order on the diagnostic yield of metagenomic deep sequencing (MDS) for determining the causative pathogen in infectious keratitis. METHODS: We performed a cross-sectional diagnostic test evaluation among subjects with infectious keratitis at Aravind Eye Hospital in Madurai, India. All subjects underwent corneal scrapings of the affected eye to obtain potassium hydroxide smear, Gram stain, bacterial culture, and fungal culture, in this order. The order of MDS specimen collection relative to smear and culture samples was randomized and served as the primary predictor. Outcomes included the normalized copy number of pathogenic RNA detected by MDS, the proportion of MDS samples that were diagnostic, and the agreement of MDS results with cultures. RESULTS: MDS samples from 46 subjects with corneal ulcers were evaluated. MDS was positive in 33 subjects (76%) and had 74% overall agreement with culture results. There was no association between order of MDS sample collection and normalized copy number of genetic material detected (P = 0.62) or the likelihood of MDS positivity (P = 0.46). However, the likelihood of agreement between MDS and cultures decreased when MDS corneal swabs were collected after other diagnostic corneal scrapings (P = 0.05). CONCLUSIONS: The overall yield of MDS for detecting the cause of infectious keratitis was not affected by sample collection order. However, diagnostic agreement between MDS and cultures decreased when MDS samples were collected after other specimens. Additional investigation is warranted to determine whether this represents increased sensitivity of MDS compared with cultures or higher susceptibility to contaminants.

Dysregulated expression of microRNAs in aqueous humor from intraocular tuberculosis patients.

BACKGROUND: Systemic Mycobacterium tuberculosis (Mtb) infection alters microRNA's expression that controls cellular processes and modulates host defense mechanisms. However, the role of miRNAs in intraocular tuberculosis (IOTB) remains unknown. Therefore, this study aims to identify dysregulated miRNAs in the aqueous humor (AH) of patients with IOTB. METHODS: AH from intraocular tuberculosis patients (n = 2) and cataract controls (n = 2) were used for small RNA deep sequencing using HiSeq Illumina sequencing platform. Differentially expressed miRNAs and their targets were identified by the bioinformatics approach, and their regulatory functions were predicted by pathway enrichment analysis. The expression of selected miRNAs and their binding targets were further validated by real-time quantitative PCR (RT-qPCR). RESULTS: In total, we identified 56 differentially expressed miRNAs in the AH of intraocular tuberculosis (IOTB) patients compared to controls. We selected four significantly dysregulated miRNAs (miR-423-5p, miR-328-3p, miR-21-5p, and miR-16-5p) based on the RT-qPCR validation and predicted their gene targets. We developed a miRNA-targets regulatory network by combining pathways of interest and genes associated with TB. We identified that these four miRNAs might play an important role in IOTB pathogenesis via tuberculosis-associated pathways; PI3K-Akt signaling, autophagy and MAPK pathway. CONCLUSIONS: For the first time, this study identifies the dysregulation of four miRNAs in the AH of IOTB patients using the ultra-low input small-RNA sequencing approach. Further target prediction and validation identify the role of these miRNAs in tuberculosis pathogenesis via tuberculosis-related pathways. This study identifies miRNAs as potentially ideal biomarkers in the aqueous humor of IOTB patients.

Differentiation of Active Corneal Infections from Healed Scars Using Deep Learning.

PURPOSE: To develop and evaluate an automated, portable algorithm to differentiate active corneal ulcers from healed scars using only external photographs. DESIGN: A convolutional neural network was trained and tested using photographs of corneal ulcers and scars. PARTICIPANTS: De-identified photographs of corneal ulcers were obtained from the Steroids for Corneal Ulcers Trial (SCUT), Mycotic Ulcer Treatment Trial (MUTT), and Byers Eye Institute at Stanford University. METHODS: Photographs of corneal ulcers (n = 1313) and scars (n = 1132) from the SCUT and MUTT were used to train a convolutional neural network (CNN). The CNN was tested on 2 different patient populations from eye clinics in India (n = 200) and the Byers Eye Institute at Stanford University (n = 101). Accuracy was evaluated against gold standard clinical classifications. Feature importances for the trained model were visualized using gradient-weighted class activation mapping. MAIN OUTCOME MEASURES: Accuracy of the CNN was assessed via F1 score. The area under the receiver operating characteristic (ROC) curve (AUC) was used to measure the precision-recall trade-off. RESULTS: The CNN correctly classified 115 of 123 active ulcers and 65 of 77 scars in patients with corneal ulcer from India (F1 score, 92.0% [95% confidence interval (CI), 88.2%-95.8%]; sensitivity, 93.5% [95% CI, 89.1%-97.9%]; specificity, 84.42% [95% CI, 79.42%-89.42%]; ROC: AUC, 0.9731). The CNN correctly classified 43 of 55 active ulcers and 42 of 46 scars in patients with corneal ulcers from Northern California (F1 score, 84.3% [95% CI, 77.2%-91.4%]; sensitivity, 78.2% [95% CI, 67.3%-89.1%]; specificity, 91.3% [95% CI, 85.8%-96.8%]; ROC: AUC, 0.9474). The CNN visualizations correlated with clinically relevant features such as corneal infiltrate, hypopyon, and conjunctival injection. CONCLUSIONS: The CNN classified corneal ulcers and scars with high accuracy and generalized to patient populations outside of its training data. The CNN focused on clinically relevant features when it made a diagnosis. The CNN demonstrated potential as an inexpensive diagnostic approach that may aid triage in communities with limited access to eye care.