Ensemble Machine Learning Approaches for Automated Fungal Keratitis Diagnosis Using In Vivo Confocal Microscopy Images

In vivo confocal microscopy (IVCM) is an emerging technology that provides minimally invasive, high resolution, steady-state assessment of the ocular surface at the cellular level. Several challenges still remain but, at present, IVCM may be considered a promising technique for clinical diagnosis and management. This mini-review summarizes some key findings in IVCM of the ocular surface, focusing on recent and promising attempts to move “from bench to bedside”. IVCM allows prompt diagnosis, disease course follow-up, and management of potentially blinding atypical forms of infectious processes, such as acanthamoeba and fungal keratitis. This technology has improved our knowledge of corneal alterations and some of the processes that affect the visual outcome after lamellar keratoplasty and excimer keratorefractive surgery. In dry eye disease, IVCM has provided new information on the whole-ocular surface morphofunctional unit. It has also improved understanding of pathophysiologic mechanisms and helped in the assessment of prognosis and treatment. IVCM is particularly useful in the study of corneal nerves, enabling description of the morphology, density, and disease- or surgically induced alterations of nerves, particularly the subbasal nerve plexus. In glaucoma, IVCM constitutes an important aid to evaluate filtering blebs, to better understand the conjunctival wound healing process, and to assess corneal changes induced by topical antiglaucoma medications and their preservatives. IVCM has significantly enhanced our understanding of the ocular response to contact lens wear. It has provided new perspectives at a cellular level on a wide range of contact lens complications, revealing findings that were not previously possible to image in the living human eye. The final section of this mini-review provides a focus on advances in confocal microscopy imaging. These include 2D wide-field mapping, 3D reconstruction of the cornea and automated image analysis.

In Vivo Confocal Microscopy and Anterior Segment Optical Coherence Tomography Analysis of the Cornea in Nephropathic Cystinosis

En-face Optical Coherence Tomography as a Novel Tool for Exploring the Ocular Surface: A Pilot Comparative Study to Conventional B-Scans and in Vivo Confocal Microscopy

With the advancement of skin research, today's consumer has increased access to technological information about ageing skin and hair care products. As a result, there is a rapidly increasing demand for proof of efficacy of these products. Recognizing these demands has led to the development and validation of many clinical methods to measure and quantify ageing skin and the effects of anti-ageing treatments. Many of the current testing methods used to research and evaluate anti-ageing product claim to employ sophisticated instruments alongside more traditional clinical methods. Intelligent use of combined clinical methods has enabled the development of technologically advanced consumer products providing enhanced efficacy and performance. Of non-invasive methods for the assessment and quantification of ageing skin, there is a plethora of tools available to the clinical researcher as defined by key clinically observed ageing parameters: skin roughness and surface texture; fine lines and wrinkles; skin pigmentation; skin colour; firmness and elasticity; hair loss; and proliferative lesions. Furthermore, many clinical procedures for the evaluation of ageing skin treatments are combined with invasive procedures, which enable added-value to claims (such as identification and alteration of biochemical markers), particularly in those cases where perception of product effect needs additional support. As discussed herein, clinical methods used in the assessment of skin ageing are many and require a disciplined approach to their use in such investigations.

In vivo confocal microscopy (IVCM) imaging is increasingly popular in ocular surface disease diagnosis and management. We conducted a systematic review to update the use of IVCM in the diagnosis and treatment of dry eye and meibomian gland dysfunction (MGD). A literature review was conducted on IVCM studies in MGD, dry eye disease, systemic disease causing dry eye, dry eye in glaucoma patients, contact lens-associated ocular conditions, graft-versus-host disease, and Sjogren’s syndrome-related dry eye. The articles were identified through PubMed and a total number of 63 eligible publications were analyzed in detail. All primary research studies on confocal microscopy on dry eye and related conditions from 2017 onwards were included. The reports were reviewed for their contribution to the existing literature as well as potential biases and drawbacks. Despite limitations such as small field of view, lack of population-based norms, and lack of standardization of image acquisition, interpretation, and quantification, IVCM is useful as a complementary technique for clinical diagnosis in various ocular surface disorders related to dry eye. With advances in hardware and software in the near future, it has the potential for further practical impact.

Objectives:To analyze and assess compatibility of trabeculectomy filtering bleb characteristics and appearances using biomicroscopy, anterior segment optical coherence tomography (AS-OCT) and in vivo confocal microscopy (IVCM).Materials and Methods:Twenty-eight eyes of 28 patients who underwent glaucoma filtering surgery with mitomycin C in our clinic between 2009 and 2013 were evaluated. Morphological appearances of the blebs on slit-lamp biomicroscopy were defined according to the Moorfields bleb classification system. For the internal tissue assessment of blebs, AS-OCT and IVCM were performed. Bleb biometric parameters such as length, height and bleb wall thickness were assessed by AS-OCT; conjunctival epithelial-stromal cyst, structural network of conjunctival stroma and vascularisation were examined with IVCM. The relation between biomicroscopic morphological staging and bleb characteristics detected on AS-OCT and IVCM were assessed.Results:The mean age of the 28 patients (16 male, 12 female) was 57.2±15.9 (19 to 79) years. The mean time elapsed between surgery and examination was 29.2±19.2 (6 to 68) months. According to biomicroscopic appearance, 17 (60.7%) blebs were functional (13 diffuse, 4 microcystic), whereas 11 (39.3%) blebs were non-functional (9 flat, 2 encapsulated). In the comparison of non-functional and functional blebs, functional blebs were found to be superior in terms of biometric parameters on AS-OCT assessment (p<0.05). Higher number of epithelial and stromal cysts and less vascularisation were detected by IVCM in functional blebs when compared with non-functional blebs (p<0.05).Conclusion:Biomicroscopic appearances and characteristics on AS-OCT and IVCM of filtration blebs are consistent with each other. Besides biomicroscopic examination, which is an easy and practical method for determining bleb morphology, cross-sectional images obtained by AS-OCT and IVCM provide objective data regarding internal structure and functional features of blebs.

To investigate the characteristics of in vitro culture and in vivo confocal microscopy (IVCM) in patients with fungal keratitis (FK) presented in a tertiary referral hospital in central China. In this noncomparative retrospective study, patients with the diagnosis of FK between October 2021 and November 2022 were reviewed. An IVCM and fungal culture (corneal scraping specimens) were performed, and the characteristics were analyzed. During October 2021 and November 2022, 85 patients were diagnosed with FK. From 63 culture-positive cases, 8 species of fungus were identified. The proportions of isolated fungal species were Fusarium and Aspergillus equally accounting for 33.3% (21 of 63), Alternaria 9.5% (6 of 63), Curvularia 6.3% (4 of 63), Scedosporium apiospermum 6.3% (4 of 63), Paecilomyces lilacinus 3.2% (2 of 63), Exserohilum 3.2% (2 of 63), and Candida 4.8% (3 of 63), respectively. In positive culture cases, IVCM was found to be positive for hyphae or spores in 61 of 63 patients (96.8%). Different fungal species had a variety of cultural characteristics and IVCM manifestations. In a tertiary referral hospital in central China, Fusarium species, Aspergillus species, and Alternaria species were the 3 most common isolated fungal pathogens, and the proportion of Aspergillus species was significantly higher than that in other regions of China. Careful lesion depth examination by IVCM and OCT should be taken before lamellar keratoplasty to avoid postoperative recurrence. Identifying the IVCM image and culture characteristics will facilitate rapid diagnosis and proper treatment, but IVCM cannot yet replace fungal cultures to distinguish between different fungal species.

A new era of ocular imaging has recently begun with the advent of in vivo confocal microscopy (IVCM), shedding more light on the pathophysiology, diagnosis, and potential treatment strategies for dry eye disease. IVCM is a noninvasive and powerful tool that allows detection of changes in ocular surface epithelium, immune and inflammatory cells, corneal nerves, keratocytes, and meibomian gland structures on a cellular level. Ocular surface structures in dry eye-related conditions have been assessed and alterations have been quantified using IVCM. IVCM may aid in the assessment of dry eye disease prognosis and treatment, as well as lead to improved understanding of the pathophysiological mechanisms in this complex disease. Further, due to visualization of subclinical findings, IVCM may allow detection of disease at much earlier stages and allow stratification of patients for clinical trials. Finally, by providing an objective methodology to monitor treatment efficacy, image-guided therapy may allow the possibility of tailoring treatment based on cellular changes, rather than on clinical changes alone.

Endothelial plaque is an important sign of fungal keratitis and is related to diagnosis, surgical indications, and prognosis. However, bacterial keratitis sometimes involves fibrin formation on the back corneal surface, similar to endothelial plaques. Because corneal infiltration interferes with precise observation of the posterior corneal plaque, distinguishing pathogens with a slitlamp is difficult. We hope to assist clinicians in early diagnosis and timely treatment by observing the connection state of endothelial plaques and the corneal endothelium through anterior segment optical coherence tomography (AS-OCT) and the different forms of endothelial plaques in infectious keratopathy through in vivo confocal microscopy (IVCM). We analyzed 52 patients in the Eye Hospital of the First Affiliated Hospital of Harbin Medical University who were clearly diagnosed with fungal or bacterial keratitis with endothelial plaques. All patients underwent AS-OCT and IVCM on admission. According to the smear, IVCM, or fungal and bacterial culture results, the patients were diagnosed with fungal (28 patients) or bacterial keratitis (24 patients). AS-OCT in 25 patients diagnosed with fungal keratitis revealed that the corneal endothelium-endothelial plaque boundary was unclear and wavy, and 24 patients had unclear cell boundaries and a large number of compactly distributed inflammatory cells in the endothelial layer according to IVCM. AS-OCT in 23 patients diagnosed with bacterial keratitis revealed clear corneal endothelium-endothelial plaque boundaries, and insufficient endothelial cell boundaries with a large number of visible and scattered inflammatory cell structures were observed through IVCM in 22 patients. Corneal endothelial plaque detection by AS-OCT and IVCM can be used for early diagnosis of infectious keratitis.

To develop a feasible method to image eyelid margin structures using in vivo confocal microscopy (IVCM) for use in clinical research. Second, to assess the association between IVCM and meibography images. IVCM was performed on the central upper eyelid margin of 13 healthy participants (31 ± 5 years). Overall morphology montages (1600 × 1600 μm) were created of 3 participants. Single frames (400 × 400 μm) of 10 participants were imaged to determine the feasibility of measuring eyelid features. Meibography was performed with EASYTEARview+ in the same 10 participants. ImageJ software was used to quantify image structures. In the montages, structures of rete ridges, meibomian gland openings, and the lid wiper region were observed. The maximum possible montage size, using multiple single frames, was approximately 5200 × 1500 × 150 μm in the X, Y, and Z directions, respectively. The mean number, density, area, perimeter, and shortest and longest diameters of rete ridges of the 9 nonoverlapped frames were 12 ± 2/frame, 73 ± 5/mm, 2504 ± 403 μm, 250 ± 33 μm, 40 ± 6 μm, and 84 ± 13 μm, respectively. Sampling analysis determined at least 5 nonoverlapped frames were necessary to accurately represent the parameters of the ridges. The mean areas of 3 meibomian openings were 785 ± 784 μm, 1036 ± 963 μm, 950 ± 1071 μm, 848 ± 954 μm, 737 ± 831 μm, 735 ± 743 μm, and from 30 μm to 130 μm at 20-μm depth intervals, respectively. No significant association between IVCM and meibography parameters (P = 0.53) was found. Imaging rete ridges with IVCM should include at least 5 nonoverlapping single frames in the upper eyelid margin. At least 3 openings imaged between 30 and 130 μm at 20-μm depth intervals are recommended to determine the opening area.

Background and Purpose In vivo confocal microscopy (IVCM) is a non-invasive imaging technique that allows morphological analysis as a diagnostic approach of the cornea in real time, thus providing a suspected diagnosis of fungal or amoebic keratitis immediately, whereas culture or PCR require several days or even weeks. Since these infections are rare, it is difficult for ophthalmologists to gain the experience necessary to differentiate infection from normal findings or artefacts. The purpose of this project was to establish a simulator, on which physicians could practice as well as acquiring a database of IVCM images of fungal or amoebic keratitis and respective analyses. Patients and methods An IVCM simulator was set up with cadaver human corneas, infected with either acanthamoeba, candida or aspergillus. Twenty-one ophthalmologists were trained in IVC microscopy first in a Dry Lab, then practically on the simulator. For evaluation, the participants were asked to fill out a standardized questionnaire, with a pre- and post-course self-assessment. Results The self-assessed theoretical and practical skills in differentiating infectious from non-infectious keratitis in IVCM significantly increased (p = 0.0001, p = 0.0002, respectively). The barrier to use this technique decreased (p = 0.0474). Conclusion A very simple protocol based on a model of ex vivo corneal mycotic and amoebic infections can be used to train novices in the structured approach and diagnostic use of IVCM for corneal infections.

To quantitatively analyze laser scanning in vivo confocal microscopy (IVCM) images of all corneal layers in contact lens-wearing and noncontact lens-wearing keratoconus patients. The study population included rigid gas permeable (RGP) contact lens-wearing keratoconus patients (group 1; N = 29), keratoconus patients who did not wear contact lenses (group 2; N = 30), and subjects who neither had keratoconus nor wore contact lenses (group 3; N = 30), with groups 2 and 3 matched to group 1 by age and sex. The central cornea was examined with IVCM in all subjects. The mean duration of contact lens wear was 5.50 ± 3.68 years (range, 2-15 years). Eyes with keratoconus showed significantly lower basal epithelial cell and anterior and posterior stromal keratocyte densities, as well as subbasal nerve fiber density, nerve branch density, and nerve fiber length compared with healthy control subjects. Furthermore, compared with group 2, group 1 had significantly lower basal epithelial cell density (4920 ± 476 cells/mm(2) vs. 4503 ± 461 cells/mm(2), P = 0.001) and anterior stromal keratocyte density (561 ± 91 cells/mm(2) vs. 464 ± 55 cells/mm(2), P < 0.001), but there was no significant difference for posterior stromal keratocyte density (P = 0.808), endothelial cell density (P = 0.699), or subbasal nerve fiber density (P = 0.142), nerve branch density (P = 0.614), and nerve fiber length (P = 0.850). Significant corneal microstructural abnormalities were observed in eyes with keratoconus. RGP contact lens wear was associated with a further reduction in the basal epithelial cell and anterior stromal keratocyte densities, but with no effect on posterior stromal keratocyte density, endothelial cell density, or corneal nerve morphology.

ObjectiveTo examine the reliability of clinical examination and in vivo confocal microscopy (IVCM) in distinguishing ocular surface squamous neoplasia (OSSN) from benign conjunctival lesions.DesignCase-control study.ParticipantsSixty individuals with conjunctival lesions (OSSN and benign) and 60 age-matched controls with normal conjunctiva presenting to Kilimanjaro Christian Medical Centre, Moshi, Tanzania.MethodsParticipants were examined and photographed, and IVCM was performed. Patients with conjunctival lesions were offered excisional biopsy with histopathology and a human immunodeficiency virus (HIV) test. The IVCM images were read masked to the clinical appearance and pathology results. Images were graded for several specific features and given an overall categorization (normal, benign, or malignant). A group of 8 ophthalmologists were shown photographs of conjunctival lesions and asked to independently classify as OSSN or benign.Main Outcome MeasuresComparison of the histopathology diagnosis with the clinical and IVCM diagnosis.ResultsFifty-two cases underwent excisional biopsy with histopathology; 34 were on the OSSN spectrum, 17 were benign, and 1 was lymphoma. The cases and controls had comparable demographic profiles. Human immunodeficiency syndrome infection was more common in OSSN compared with benign cases (58.8% vs. 5.6%; odds ratio, 24.3, 95% confidence interval [CI], 2.8–204; P = 0.003). Clinically, OSSN lesions more frequently exhibited feeder vessels and tended to have more leukoplakia and a gelatinous appearance. Overall, the ophthalmologists showed moderate agreement with the histology result (average kappa = 0.51; 95% CI, 0.36–0.64). The masked grading of IVCM images reliably distinguished normal conjunctiva. However, IVCM was unable to reliably distinguish between benign lesions and OSSN because of an overlap in their appearance (kappa = 0.44; 95% CI, 0.32–0.57). No single feature was significantly more frequent in OSSN compared with benign lesions. The sensitivity and specificity of IVCM for distinguishing OSSN from benign conjunctival lesions were 38.5% and 66.7%, respectively.ConclusionsIn East Africa, conjunctival pathology is relatively common and can present significant diagnostic challenges for the clinician. In this study, neither clinical examination nor IVCM was found to reliably distinguish OSSN from benign conjunctival pathology because of an overlap in the features of these groups. Therefore, IVCM cannot currently replace histopathology, and management decisions should continue to rely on careful clinical assessment supported by histopathology as indicated.

Purpose To describe the findings of in vivo confocal microscopy (IVCM) and anterior segment optical coherence tomography (AS-OCT) in a case of bilateral acute microcystic epitheliopathy after daily soft contact lens wear. Methods IVCM and AS-OCT were used in the course of the bilateral epitheliopathy of a 23-year-old patient at the acute stage of the disease and at recovery after four days of treatment. The images were analyzed and compared. Results On AS-OCT of the right eye, general hyperreflectivity and the increased thickness of the central corneal epithelium to 150 μm with numerous hyporeflective small, oval cysts were revealed and resolved completely at day 4 after diagnosis and treatment. AS-OCT scans of the left eye revealed oval shaped, hyperreflective material within the corneal epithelium. IVCM of both eyes showed numerous microcysts of different sizes filled with hyperreflective material mostly within superficial epithelial layers. Epithelial cells showed a polymorphism along with disruption of a cytoarchitecture. Basal epithelial cells and anterior stroma showed inflammatory changes. Posterior corneal stroma and endothelium presented normal morphology. Conclusions Microcystic keratitis appeared as localized microcystic inflammation of epithelium on AS-OCT, which was confirmed by IVCM. Both IVCM and AS-OCT are helpful diagnostic instruments in case of cystic inflammation of corneal epithelium.

To compare densities of corneal epithelial dendritic cells (DCs), corneal subbasal nerves, and conjunctival epithelial immune cells (EICs) between patients with dry eye disease (DED) with and without graft-versus-host disease (GVHD) using in vivo confocal microscopy (IVCM). This study included 54 patients who had moderate to severe DED either associated with (n = 33) or without (n = 21) chronic GVHD. In addition to evaluating clinical parameters of DED, images obtained by laser-scanning IVCM of the central cornea and superior tarsal conjunctiva were analyzed to measure densities of corneal epithelial DCs, corneal subbasal nerves, and conjunctival EICs. Although there were no significant differences between GVHD and non-GVHD groups in symptom scores, the GVHD group had significantly worse corneal fluorescein staining, tear break-up time, and Schirmer's scores than the non-GVHD group. Corneal epithelial DC density, corneal subbasal nerve density, and conjunctival EIC density were 148 ± 135 cells/mm2, 16.3 ± 6.1 mm/mm2, and 670 ± 267 cells/mm2, respectively, in the GVHD group; and 122 ± 99 cells/mm2, 18.3 ± 5.1 mm/mm2, and 572 ± 271 cells/mm2, respectively, in the non-GVHD group. After adjusting for clinical parameters, including the DED severity, none of the IVCM parameters was significantly different between the GVHD versus non-GVHD groups (P = 0.82, P = 0.21, and P = 0.60, respectively). In GVHD-associated DED, cellular changes in the cornea and conjunctiva observed by IVCM were similar to those seen in patients who have non-GVHD dry eye with the same level of disease severity. Therefore, corneal and conjunctival IVCM findings in GVHD-associated DED are possibly reflective of the local disease (DED) severity rather than the underlying systemic disease process.