Infectious keratitis (IK) is among the top 5 leading causes of blindness globally. Early diagnosis is key to guide an appropriate therapy to avoid complications such as vision impairment and blindness. Culture of corneal scrapes is the initial diagnostic test to grow and identify the causal organism. Alternative diagnostic tools include imaging such as with optical coherence tomography (OCT) and in vivo confocal microscopy (IVCM). OCT’s advantage is its ability to accurately determine the depth and extent of the corneal ulceration, infiltrates and haze; therefore, characterizing the severity and progression of the infection. However, it is not a preferred choice in the diagnostic tool package for infectious keratitis. IVCM is a great aid in the diagnosis of fungal and Acanthamoeba keratitis with overall sensitivities of 66-74% and 80-100%, and specificity of 78-100% and 84-100%, respectively. Recently the use of deep learning (DL) models in IK has been shown to aid diagnosis via image recognition. Most of the studies that have developed DL models to diagnose the different types of IK have utilised photographs from digital camaras, slit-lamp images, or IVCM images. Some studies have used extremely efficient single DL algorithms to train their models and others used ensemble approaches. This technology is likely to assist in the diagnosis of IK in some years. Further work is needed to examine and validate the clinical performance of the DL models in the real-world setting and to evaluate whether this technology improves patient clinical outcomes. The scope of this review was to provide a recent update on the diagnostic imaging tools in IK

Acanthamoeba keratitis (AK) is a severe, rare protozoal infection of the cornea. Acanthamoeba can survive in diverse habitats and at extreme temperatures. AK is mostly seen in contact lens wearers, whose lenses have become contaminated, or a history of water exposure, and in those without contact lens wear, who have experienced recent eye trauma involving contaminated soil or water. Infection usually results in severe eye pain, photophobia, inflammation, and corneal epithelial defects. The pathophysiology of this infection is multifactorial, including the production of cytotoxic proteases by Acanthamoeba that degrades the corneal epithelial basement membrane and induce the death of ocular surface cells, resulting in degradation of the collagen-rich corneal stroma. AK can be prevented by avoiding the risk factors, which include avoiding water contact such as swimming or showering in contact lenses, and wearing protective goggles when working on the land. AK is mostly treated with an antimicrobial therapy of biguanides alone or in combination with diaminidines, although the commercial availability of these medicines is variable. Other than anti-amoeba therapies, targeting host immune pathways in Acanthamoeba disease may lead to the development of vaccines or antibody therapeutics which could transform the management of AK.

Fungal keratitis (FK) is an invasive infection of the cornea primarily associated with Aspergillus and Fusarium species. FK is treated empirically with a limited selection of topical antifungals with varying levels of success. Though clinical infections are typically characterized by a dense network of mature mycelium, traditional models used to test antifungal susceptibility of FK isolates exclusively evaluate susceptibility in fungal cultures derived from asexual spores known as conidia. The purpose of this study was to characterize differences in fungal response when topical antifungal treatment is initiated at progressive phases of fungal development. We compared efficacy of voriconazole and luliconazole against in vitro cultures of A. flavus and F. keratoplasticum at 0, 24, and 48 h of fungal development. A porcine cadaver corneal model was used to compare antifungal efficacy of voriconazole and luliconazole in ex vivo tissue cultures of A. flavus and F. keratoplasticum at 0, 24, and 48 h of fungal development. Our results demonstrate phase-dependent susceptibility of both A. flavus and F. keratoplasticum to both azoles in vitro as well as ex vivo. We conclude that traditional antifungal susceptibility testing with conidial suspensions does not correlate with fungal susceptibility in cultures of a more advanced developmental phase. A revised method of antifungal susceptibility testing that evaluates hyphal susceptibility may better predict fungal response in the clinical setting where treatment is often delayed until days after the initial insult.

Infectious keratitis (IK) represents a major cause of corneal blindness. This study aims to investigate the demographics, risk factors, microbiological characteristics and antibiotic susceptibility patterns of IK in Taiwan over the past 15 years. A retrospective population-based study was conducted using the Chang Gung Research Database. Patients with IK were identified by diagnostic codes for corneal ulcer from 2004 to 2019. Of 7807 included subjects, 45.2% of patients had positive corneal cultures. The proportion of contact lens-related IK declined, while that of IK related to systemic diseases grew. The percentage of isolated gram-positive bacteria surpassed that of gram-negative bacteria in the 15-year period. The prevalence of Pseudomonas aeruginosa showed a decreasing trend (p = 0.004), whereas coagulase-negative Staphylococcus (CNS) and Propionibacterium species were increasingly detected (p < 0.001). Overall, the trend of antibiotic susceptibility of both gram-positive and gram-negative bacteria did not change throughout the study period. The susceptibility to the test antibiotics maintained over 90% in gram-negative isolates during 15 years. Vancomycin preserved 100% susceptibility to all gram-positive isolates. Since most tested antibiotics exhibited stable susceptibility over decades, this study reinforced that fluoroquinolones and fortified vancomycin continue to be good empiric therapies for treating bacterial keratitis in Taiwan.

Pythium insidiosum(PI) can cause sight threatening keratitis which is managed commonly by performing penetrating keratoplasty. This article is a retrospective review to assess outcomes of keratoplasty performed in patients diagnosed with PI keratitis. (2) Methods: Pre-operative, intra operative and post-operative data of patients diagnosed with PI keratitis and who underwent keratoplasty for their condition from January 2020 to December 2021 were col-lected from the central patient database of a tertiary eye care hospital in India. The data were an-alyzed for anatomic success, elimination of infection, graft survival, incidence of repeat kerato-plasty, final visual acuity and varied complications. (3) Results: In total, 16 eyes underwent pene-trating keratoplasty for PI keratitis during the study period. Mean time to keratoplasty from on-set of symptoms was 31.3 days and mean graft size was 10.4 mm. Nine out of the 16 cases had recurrence of infection following surgery, six of which required a repeat keratoplasty for elimina-tion of infection. Out of these 6 patients, one patient underwent 2 repeat keratoplasties. Mean graft size for repeat keratoplasty performed in recurrent cases was 11.7 mm. Globe was successfully salvaged in 14 out of 16 patients (87.5 %). Endo-exudates, graft infiltration, graft dehiscence, secondary glaucoma and retinal detachment were the various complications noted after kerato-plasty. (4) Conclusion: Keratoplasty remains the choice of treatment in PI keratitis, however re-currence of disease and graft failure are common. Large sized grafts, meticulous per-operative removal of infection, adjuvant cryotherapy, and intraoperative and post-operative use of antibi-otics can help in improving outcome of keratoplasty in these patients.

In this study, we aimed to develop a deep learning model for identifying bacterial keratitis (BK) and fungal keratitis (FK) by using slit-lamp images. We retrospectively collected slit-lamp images of patients with culture-proven microbial keratitis between January 1, 2010, and December 31, 2019, from two medical centers in Taiwan. We constructed a deep learning algorithm, consisting of a segmentation model for cropping cornea images and a classification model that applies convolutional neural networks to differentiate between FK and BK. The model performance was evaluated and presented as the area under the curve (AUC) of the receiver operating characteristic curves. A gradient-weighted class activation mapping technique was used to plot the heatmap of the model. By using 1330 images from 580 patients, the deep learning algorithm achieved an average diagnostic accuracy of 80.00%. The diagnostic accuracy for BK ranged from 79.59% to 95.91% and that for FK ranged from 26.31% to 63.15%. DenseNet169 showed the best model performance, with an AUC of 0.78 for both BK and FK. The heat maps revealed that the model was able to identify the corneal infiltrations. The model showed better diagnostic accuracy than the previously reported diagnostic performance of both general ophthalmologists and corneal specialists.

Tissue resident memory T cells (TRM) were first described in 2009. While initially the major focus was on CD8 TRM, there has been recently an increased interest in defining the phenotype and the role of CD4 TRM in diseases. Circulating CD4 T cells seed tissue CD4 TRM, but there also appears to be an equilibrium between CD4 TRM and blood CD4 T cells. CD4 TRM are more mobile than CD8 TRM, usually localized deeper within the dermis/lamina propria and yet may exhibit synergy with CD8 TRM in disease control. This has been demonstrated in herpes simplex infections in mice. In human recurrent herpes infections, both CD4 and CD8 TRM persisting between lesions may control asymptomatic shedding through interferon gamma secretion, although this has been more clearly shown for CD8 T cells. The exact role of the CD4/CD8 TRM axis in the trigeminal ganglia and/or cornea in controlling recurrent herpetic keratitis is unknown. In HIV, CD4 TRM have now been shown to be a major target for productive and latent infection in cervix. In HSV and HIV co-infections, CD4 TRM persisting in the dermis support HIV replication. Further understanding of the role of CD4 TRM and their induction by vaccines may help control sexual transmission by both viruses.

Acanthamoeba keratitis (AK) is a blinding corneal infection caused by the protozoan Acanthamoeba. The long-term course of AK suggests the host immunity could not kill Acanthamoeba rapidly. The immune status is still unclear in the late stage of AK. The comparative transcriptome analysis was made based on the bulk RNA sequencing of cornea tissues from AK patients and donors. Differentially expressed genes and enriched signaling pathways were calculated. CIBERSORT algorithm was used for Immune infiltration analysis of cornea tissue between AK and normal controls. A total of 2668 differentially expressed genes, including 1477 upregulated genes and 1191 downregulated genes, were detected. Gene Ontology analysis revealed that the pathways were significantly enriched in leukocyte migration, regulation of T cell activation, the external side of plasma membrane, collagen-containing extracellular matrix, immune receptor activity and cytokine binding. KEGG pathway analysis showed that the pathways were significantly enriched in the cytokine-cytokine receptor interaction, hematopoietic cell lineage and Staphylococcus aureus infection pathway. The immune infiltration profiles varied little between AK and normal controls. Compared with normal tissue, cornea tissue of AK contained a higher proportion of M0 macrophages and CD8 T cells, while resting memory CD4 T cells contributed to a relatively lower portion (P < 0.05). Finally, the expression levels of cell markers and SLAMF7/STAT6 pathway were confirmed by histopathology examinations, RT-qPCR and western blot.

Purpose. Moraxella is an ocular bacterial pathogen isolated in cases of keratitis, conjunctivitis, and endophthalmitis. Gram-negative brick-shaped diplobacilli from ocular specimens, and slow growth in culture, are early indications of Moraxella ocular infection; however, identifying Moraxella to species can be complex and inconsistent. Methods. In this study, bacteria consistent with Moraxella were identified to species using: 1) DNA sequencing coupled with vancomycin susceptibility, 2) MALDI-TOF Mass Spectrometry, and 3) Biolog ID System. Study samples consisted of 9 ATCC Moraxella controls, 82 isolates from keratitis, 21 isolates from conjunctivitis, and 4 isolates from endophthalmitis. Results. The ATCC controls were correctly identified. For keratitis, 66 (80.5%) were identified as M. nonliquefaciens, 7 (9.0%) as M. lacunata, 5 (6%) as M. osloensis, 2 (2.5%) as Acinetobacter lwoffi, 1 (1.0%) as M. bovis/nonliquefaciens, and 1 (1.0%) as M. osloensis/nonliquefaciens. For conjunctivitis, 9 (43.0%) were identified as M. osloensis, 6 (29.0%) as M. nonliquefaciens, 3 (14.3%) as Roseomonas, 2 (9.5%) as Acinetobacter (parvus, junii), and 1 (4.5%) as M. catarrhalis/M. nonliquefaciens. From endophthalmitis, 3 of 4 of the isolates were M. nonliquefaciens. Overall, M. nonliquefaciens and M. osloensis were identified in 70% (75 of 107) and 13% (14 of 107) of cases, respectively, totaling 83% (89 of 107). Conclusions. M. nonliquefaciens and M. osloensis are important bacterial pathogens of the eye as determined by DNA sequencing, MALDI-TOF MS, and Biolog. Although Moraxella catarrhalis is a clinical pathogen, other species of Moraxella appear to have a prominent role in eye infections.

Fungal keratitis (FK) is a serious ocular infection that often poses significant diagnostic and therapeutic dilemmas. This study aimed to examine the causes, clinical characteristics, outcomes, and prognostic factors of FK in the UK. All culture-positive and culture-negative presumed FK (with complete data) that presented to Queen’s Medical Centre, Nottingham, and Queen Victoria Hospital, East Grinstead, between 2011 and 2020 were included. A total of 117 patients (n=117 eyes) with FK were included in this study. The mean age was 59.0±19.6 years (range, 4-92 years) and 51.3% patients were female. Fifty-three fungal isolates were identified from 52 (44.4%) culture-positive cases, with Candida spp. (33, 62.3%), Fusarium spp. (9, 17.0%), and Aspergillus spp. (5, 9.4%) being the most common organisms. Ocular surface disease (60, 51.3%), prior corneal surgery (44, 37.6%), and systemic immunosuppression (42, 35.9%) were the three most common risk factors. Hospitalisation for intensive treatment was required for 95 (81.2%) patients, with a duration of 18.9±16.3 days. Sixty-six (56.4%) patients required additional surgical interventions for eradicating the infection. Emergency therapeutic/tectonic keratoplasty was performed in 29 (24.8%) cases, though 13 (44.8%) of them failed at final follow-up. The final corrected-distance-visual-acuity (CDVA) was 1.67±1.08 logMAR. Multivariable logistic regression analyses demonstrated increased age, large infiltrate size (>3mm), and poor presenting CDVA (<1.0 logMAR) as significant negative predictive factors for poor visual outcome (CDVA of <1.0 logMAR) and poor corneal healing (>60 days of healing time or occurrence of corneal perforation requiring emergency keratoplasty; all p<0.05). In conclusion, FK represents a difficult-to-treat ocular infection that often results in poor visual outcome, with a high need for surgical interventions. Innovative treatment strategies are urgently required to tackle this unmet need.

Machine learning (ML) has a large capacity to learn and analyze a large volume of data. This study aimed to train different algorithms to discriminate between healthy and pathologic corneal images by evaluating digitally processed spectral-domain optical coherence tomography (SD-OCT) corneal images. A set of 22 SD-OCT images belonging to a random set of corneal pathologies was compared to 71 healthy corneas (control group). A binary classification method was applied; three approaches of ML were used. Once all images were analyzed, representative areas from every digital image were also processed and analyzed for a statistical feature comparison between healthy and pathologic corneas. The best performance was obtained from transfer learning - support vector machine (TL-SVM) (AUROC = 0.94, SPE 88%, SEN 100%) and transfer learning – random forest (TL- RF) method (AUROC = 0.92, SPE 84%, SEN 100%), followed by convolutional neural network (CNN) (AUROC = 0.84, SPE 77%, SEN 91%) and random forest (AUROC = 0.77, SPE 60%, SEN 95%). The highest diagnostic accuracy in classifying corneal images was achieved with the TL-SVM and the TL-RF models. In image classification, CNN was a strong predictor. This pilot experimental study developed a systematic mechanized system to discern pathologic from healthy corneas.