Analysis of the performance of the CorneAI for iOS in the classification of corneal diseases and cataracts based on journal photographs.

CorneAI for iOS is an artificial intelligence (AI) application to classify the condition of the cornea and cataract into nine categories: normal, infectious keratitis, non-infection keratitis, scar, tumor, deposit, acute primary angle closure, lens opacity, and bullous keratopathy. We evaluated its performance to classify multiple conditions of the cornea and cataract of various races in images published in the Cornea journal. The positive predictive value (PPV) of the top classification with the highest predictive score was 0.75, and the PPV for the top three classifications exceeded 0.80. For individual diseases, the highest PPVs were 0.91, 0.73, 0.42, 0.72, 0.77, and 0.55 for infectious keratitis, normal, non-infection keratitis, scar, tumor, and deposit, respectively. CorneAI for iOS achieved an area under the receiver operating characteristic curve of 0.78 (95% confidence interval [CI] 0.5-1.0) for normal, 0.76 (95% CI 0.67-0.85) for infectious keratitis, 0.81 (95% CI 0.64-0.97) for non-infection keratitis, 0.55 (95% CI 0.41-0.69) for scar, 0.62 (95% CI 0.27-0.97) for tumor, and 0.71 (95% CI 0.53-0.89) for deposit. CorneAI performed well in classifying various conditions of the cornea and cataract when used to diagnose journal images, including those with variable imaging conditions, ethnicities, and rare cases.

Adoption of Innovation in Herpes Simplex Virus Keratitis.

Herpes simplex keratitis, caused primarily by human herpes simplex virus type 1 (HSV-1), remains the most common infectious cause of unilateral blindness and vision impairment in the industrialized world. Major advances in the care of HSV keratitis have been driven in large part by the landmark Herpetic Eye Disease Study randomized clinical trials, which were among the first in ophthalmology to reflect emerging trial conventions, including multicenter subject enrollment, double-masking, placebo controls, and a priori sample size determinations. The results of these trials now form much of the evidence basis for the management of this disease. However, management patterns in clinical practice often deviate from evidence-based care. These perceived quality gaps have given rise to the evolving field of implementation science, which is concerned with the methods of promoting the application of evidence-based medicine within routine care. To overcome variations in the quality and consistency of care for HSV keratitis, a range of clinical- and technology-based innovations are proposed. The most pressing needs include the following: a rational and tractable disease classification scheme that provides an immediate link between the anatomical localization of disease (corneal epithelial, stromal, or endothelial) and the appropriate treatment, and the actualization of an electronic medical record system capable of providing evidence-based treatment algorithms at relevant points of care. The latter would also input data to population-wide disease registries to identify implementation-rich targets for quality improvement, education, and research. These innovations may allow us to reduce the human and economic burdens of this highly morbid, and often blinding, disease.

"Descemet Membrane Detachment": A Novel Concept in Diagnosis and Classification.

PURPOSE: To examine the optical coherence tomography (OCT) and histologic features of Descemet membrane detachment (DMD) to ascertain the involvement of the pre-Descemet layer (PDL). DESIGN: Retrospective, observational case series. METHODS: Clinical, histopathologic, and OCT features of a cohort of 41 cases with diagnosis of DMD from 4 centers were studied. OCT images were evaluated independently by 3 observers for number of detached layers (1 or 2), reflectivity, configuration (straight line or wavy), distance from posterior stroma, and presence or absence of a tear with any scrolling of the torn edges. Five had a histology specimen. The main outcome measure was the involvement of the PDL in DMD and its confirmation by histology. RESULTS: Three types of DMD were identified: type 1, where the PDL and DM were detached together; type 2, where only the DM was detached; and mixed, where the PDL and DM were detached but also separated from each other. These were further found to be rhegmatogenous or nonrhegmatogenous depending on the presence of absence of a tear in DM or both layers. Histology confirmed involvement of PDL in all 5 cases and showed it to be infiltrated by cells in 3 of 5 cases. CONCLUSIONS: The PDL is involved in DMD. This fact significantly changes our understanding of DMD and could have implications for management. The detached PDL can be infiltrated with cells. A prospective study in relation to etiology and types of DMD is needed.

Global Consensus on Definition, Classification, Diagnosis, and Staging of Limbal Stem Cell Deficiency.

PURPOSE: Despite extensive knowledge gained over the last 3 decades regarding limbal stem cell deficiency (LSCD), the disease is not clearly defined, and there is lack of agreement on the diagnostic criteria, staging, and classification system among treating physicians and research scientists working on this field. There is therefore an unmet need to obtain global consensus on the definition, classification, diagnosis, and staging of LSCD. METHODS: A Limbal Stem Cell Working Group was first established by The Cornea Society in 2012. The Working Group was divided into subcommittees. Four face-to-face meetings, frequent email discussions, and teleconferences were conducted since then to obtain agreement on a strategic plan and methodology from all participants after a comprehensive literature search, and final agreement was reached on the definition, classification, diagnosis, and staging of LSCD. A writing group was formed to draft the current manuscript, which has been extensively revised to reflect the consensus of the Working Group. RESULTS: A consensus was reached on the definition, classification, diagnosis, and staging of LSCD. The clinical presentation and diagnostic criteria of LSCD were clarified, and a staging system of LSCD based on clinical presentation was established. CONCLUSIONS: This global consensus provides a comprehensive framework for the definition, classification, diagnosis, and staging of LSCD. The newly established criteria will aid in the correct diagnosis and formulation of an appropriate treatment for different stages of LSCD, which will facilitate a better understanding of the condition and help with clinical management, research, and clinical trials in this area.

In vivo microscopic and optical coherence tomography classification of neurotrophic keratopathy.

Neurotrophic keratopathy (NK) is a rare degenerative corneal disorder characterized by instability of epithelial integrity with consequent epithelial defects that can worsen up to persistent epithelial defects with stromal melting and ulceration. The pathogenesis of NK springs from a variable degree of damage to the trigeminal nerve plexus, leading to a reduction or total loss of corneal sensitivity. Mackie classification (1995) distinguishes three stages of NK, based on the severity of clinical presentation. The technological innovations in corneal diagnostic imaging allow clinicians to accurately study the morphometry and morphology of corneal structure with microscopic resolution. In this study, 45 patients affected by NK at different stages underwent in vivo confocal microscopy (IVCM) and anterior segment optical coherence tomography (AS-OCT) with particular attention to analyze subbasal nerve plexus fibers and the stromal structure. At the light of IVCM and AS-OCT observations, we propose a different staging of NK with respect to the Mackie's classification that takes into account the severity of subbasal nerve fibers damage and the extension in depth of stromal ulceration; this classification better defines, at the time of diagnosis, the cellular and structural alterations in the affected corneas, with possible prognostic and therapeutic values in the management of NK.

Classification of Limbal Stem Cell Deficiency Using Clinical and Confocal Grading.

PURPOSE: To grade the severity of limbal stem cell deficiency (LSCD) based on the extent of clinical presentation and central corneal basal epithelial cell density (BCD). METHODS: This is a retrospective observational comparative study of 48 eyes of 35 patients with LSCD and 9 eyes of 7 normal subjects (controls). Confocal images of the central cornea were acquired. A clinical scoring system was created based on the extent of limbal and corneal surface involvement. LSCD was graded as mild, moderate, and severe stages based on the clinical scores. The degree of BCD reduction was given a score of 0 to 3. RESULTS: Compared with BCD in controls, BCD decreased by 23.0%, 40.4%, and 69.5% in the mild, moderate, and severe stages of LSCD classified by the clinical scoring system, respectively. The degree of BCD reduction was positively correlated with larger limbal and corneal surface involvement and when the central visual axis was affected (all P ≤ 0.0005). Mean corrected distance visual acuity logarithm of the minimum angle of resolution was 0.0 ± 0.0 in control eyes, 0.2 ± 0.5 in mild LSCD, 0.6 ± 0.4 in moderate LSCD, and 1.6 ± 1.1 in severe LSCD (P < 0.0001). There was a significant correlation between a higher clinical score and corrected distance visual acuity logarithm of the minimum angle of resolution (rho = 0.82; P < 0.0001) and a greater decrease in BCD (rho = -0.78; P < 0.0001). CONCLUSIONS: A clinical scoring system was developed to assess the extent of clinical presentation of LSCD. A classification system to grade the severity of LSCD can be established by combining the BCD score with the clinical score.

Anterior segment optical coherence tomography for the diagnosis of corneal dystrophies according to the IC3D classification.

Corneal dystrophies are categorized according to the International Committee for Classification of Corneal Dystrophies (IC3D) classification, and their treatment depends on the affected structures and layer of the cornea. Therefore, estimating the depth and extent of the morphological changes due to the specific dystrophy is crucial when deciding between different treatment options. Besides superficial laser treatments and penetrating keratoplasty, minimal invasive lamellar keratoplasties such as Descemet membrane endothelial keratoplasty, deep anterior lamellar keratoplasty, or Descemet stripping automated keratoplasty have become increasingly popular to exchange the specific opaque layers in dystrophic eyes. To determine the morphological changes of the cornea in the different dystrophies, in addition to slit-lamp examination, anterior segment optical coherence tomography has become an important tool with nearly histological resolution. Nonetheless, only a few case series describe the characteristics of changes seen on anterior segment optical coherence tomography. Therefore, we summarize anterior segment optical coherence tomography signs and correlate with slit-lamp examination, as well as the histopathological findings, of corneal dystrophies according to the IC3D classification.

[Corneal dystrophies]. / Les dystrophies de cornée.

Degenerative or hereditary corneal diseases are sometimes difficult to discriminate. Corneal dystrophies affect approximately 0.09 % of the population. They are identified by the IC3D classification based on their phenotype, genotype and evidence gathered for their diagnosis. Practically, the ophthalmologist manages functional symptoms, such as recurrent erosions, visual loss and amblyopia, photophobia, foreign body sensation, and sometimes pain and aesthetic concerns. Medical treatments consist of drops to promote healing, ointments, hyperosmotic agents and bandage contact lenses. Less invasive surgical treatments are used as second line therapy (phototherapeutic keratectomy, lamellar keratectomy). More invasive procedures may eventually be utilized (lamellar or penetrating keratoplasty). Anterior lamellar or endothelial keratoplasty are now preferred to penetrating keratoplasty, although the latter still remains the only possible option in some cases. Some rare dystrophies require coordinated and comprehensive medical care.

A case of recurrent, grade 3 corneal choriostoma after surgical excision.

We report a case of congenital grade 3 corneal choriostoma accompanied by anterior staphyloma. The tumor was excised by deep lamellar dissection, and the defect was filled with multilayer amniotic membranes. On histologic examination, the excised tumor contained demislike connective tissues, dermal adnexa (hair follicles and sebaceous gland), and adipose tissue, and was lined by keratinizing squamous epithelium. One month after surgery, the tumor recurred and progressively enlarged.

Conservative Surgical Management of a Grade III Corneal Dermoid.

PURPOSE: To present a rare, unique, grade III corneal dermoid treated with conservative surgical management. METHODS: An 8-month-old boy was admitted to our clinic with a lesion on the central cornea of the right eye and a corneal mass in the left eye. The mass had a skin-like surface and protruded to the outside of the palpebral fissure. MRI (magnetic resonance imaging) of the orbits showed a large cyst covering most of the cornea and the absence of the anterior chamber and lens. The dermoid was excised and sent for histopathological examination. The cornea was reconstructed using the partial thickness scleral graft. RESULTS: The histopathology report confirmed the diagnosis of a grade III corneal dermoid. On examination 1 year after the surgery, the orbital and globe volumes were the same, and the intraocular pressure was normal. CONCLUSIONS: This is likely the first report of an unusual case in which a grade III corneal dermoid with an anterior staphyloma was treated by anterior segment reconstruction using an autologous partial thickness scleral graft. We believe that this technique has the advantage that it stimulates orbital and facial development and has good cosmetic results.

Severity of acute ocular involvement is independently associated with time to resolution of ocular disease in toxic epidermal necrolysis patients.

PURPOSE: Toxic epidermal necrolysis (TEN) is a severe, life-threatening mucocutaneous disorder that frequently involves the ocular surface. This study aims to investigate the natural history and resolution of acute ocular involvement in patients with TEN admitted to the intensive care unit (ICU). METHODS: Case notes of patients admitted to ICU with TEN at a tertiary referral centre in a 9-year period were retrospectively reviewed. Patients' characteristics, severity of ocular involvement, SCORTEN systemic severity score and treatment were correlated with resolution of ocular involvement and time to resolution. RESULTS: Nine out of 10 (90%) patients had ocular involvement with 4 graded as mild, 2 as moderate and 3 as severe. All had bilateral ocular disease. The median length of hospital stay was 28 days and the median time to resolution of ocular involvement was 19 days. Four out of 9 (44%) patients still had active ocular disease at the time of discharge. Only older age (p=0.032) and a milder grade of ocular disease (p=0.001) were significantly associated with resolution of ocular disease. In a multivariable Cox-regression model, only a milder grade of ocular disease remained independently associated with time to resolution of ocular disease (p=0.006). CONCLUSIONS: Grading of acute ocular disease severity does not reflect systemic disease severity and is significantly associated with resolution and time to resolution of ocular involvement in TEN. The high rate of ocular involvement in patients with TEN and relatively large proportion of patients with active disease on discharge reiterates the need for constant ophthalmological monitoring of these patients.

Prevalence of chronic ocular complications in Stevens-Johnson syndrome and toxic epidermal necrolysis.

PURPOSE: The aim of this study is to identify and grade the severity of chronic ocular complications in patients who suffered from Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN) treated in Cape Town, South Africa. METHODS: A total of 54 patients with SJS or TEN for 6 months or longer were examined. The ocular complications were classified into corneal, eyelid and conjunctival complications. The complications were graded from 0 to 3 depending on the severity. RESULTS: A total of 108 eyes were included in the study. Medications caused SJS or TEN in all cases, and the most common associated drugs were anti-retroviral medications. 59.3% of patients were HIV-positive, with CD4 counts ranging from 6 to 521. Although only 11% of patients with SJS or TEN had acute ocular complications during the initial illness, 89% developed chronic ocular complications. Loss of the palisades of Vogt was the most common corneal complication. Among the six components of conjunctival and eyelid complications, irregularity of the mucocutaneous junction abnormalities was the most common, followed by mild conjunctival hyperemia. There was no statistically significant difference in the severity of chronic ocular complications between HIV-positive and HIV-negative patients (P = 0.4). In addition, the severity of chronic ocular complications was not statistically significantly associated with visual acuity loss (P = 0.3). CONCLUSION: We conclude that almost 90% of patients who are diagnosed with SJS or TEN will develop chronic ocular complications. Unless eyelids are severely affected, most chronic complications are mild to moderate ocular surface abnormalities and not necessarily vision-threatening complications.

Reliability of 4 clinical grading systems for corneal staining.

PURPOSE: To determine the interobserver and intraobserver reliability of 4 clinical grading systems for corneal staining. DESIGN: Retrospective, observational study. METHODS: One hundred twenty-two photographs of corneal erosions from variable ocular surface diseases were graded by 11 ophthalmologists. Each image was graded with 4 grading systems: the Oxford scheme, the National Eye Institute-recommended system, the area-density combination index, and the Sjögren's International Collaborative Clinical Alliance ocular staining score. Grading was repeated after 1 week to evaluate repeatability. Interobserver and intraobserver reliability were evaluated using intraclass correlation coefficients (ICCs). To determine the degree of agreement based on the severity of corneal staining, the relationship between the variance and the score using each grading system was evaluated with linear regression. RESULTS: Interobserver reliability for the 4 grading systems was excellent, with ICCs ranging from 0.981 to 0.991. The intraobserver repeatability of the 4 grading systems also was excellent, with ICCs ranging from 0.939 to 0.998. The National Eye Institute-recommended system showed the best reliability and repeatability. There was no definite correlation between variance and score in the Oxford scheme (Y = 0.006X + 0.284; R(2) = 0.002) or the Sjögren's International Collaborative Clinical Alliance ocular staining score grading system (Y = -0.068X + 0.595; R(2) = 0.109). However, there was a significant correlation between variance and score in the National Eye Institute-recommended system (Y = 0.210X + 0.965; R(2) = 0.144) and in the area-density combination index (Y = 0.187X + 0.279; R(2) = 0.178); the variance increased with the corneal staining score. CONCLUSIONS: The 4 grading systems may be useful for evaluation of corneal staining independent of disease conditions and grading individuals.

Conjunctivochalasis as a contributing factor for the development of ocular surface disease in medically treated glaucoma patients.

PURPOSE: To evaluate the impact of conjunctivochalasis (CCh) and its severity on the ocular surface parameters of glaucoma subjects treated with topical antiglaucomatous medication. MATERIALS AND METHODS: One hundred patients with the clinical diagnosis of glaucoma were recruited for this study. CCh was graded based on the extent of inferior lid margin involvement as follows: 1=single (temporal) location, 2=two locations (nasal and temporal), and 3=whole lid. For all the subjects, the break-up time (BUT), lissamine green (LG) staining, and Schirmer test (under topical anesthesia) was performed for both eyes. Ocular Surface Disease Index (OSDI) questionnaire scores were also noted for each subject. Student t test, Mann-Whitney U test, and Kruskal-Wallis test was used for statistical evaluations. RESULTS: Sixty-three subjects had evidence of CCh; 32 had grade 1, 24 had grade 2, 7 had grade 3 CCh, and 37 glaucoma patients had no evidence of CCh. There were significant differences in the BUT scores (7.2±2.7 vs. 10.1±2.4 s, P<0.001), Schirmer values (7.7±3.9 vs. 13.3.±4.0 mm, P<0.001), LG staining score (1.6±1.0 vs. 0.3±0.5, P<0.001), and the OSDI scores (19.4±17.2 vs. 6.7±5.2, P<0.001) between patients who had CCh and those without CCh. The BUT scores and Schirmer test values of patients with grades 2 and 3 CCh were significantly lower than those of grade 1 CCh and those without CCh (P<0.001). The LG grading and OSDI scores were significantly higher in grade 2 and 3 patients as compared with those with grade 1 CCh and those without CCh (P<0.001). CONCLUSIONS: The functional characteristics of the ocular surface appear to be adversely influenced by the presence and the extent of CCh in glaucoma patients.

The corneal transplant score: a simple corneal graft candidate calculator.

BACKGROUND: Shortage of corneas for transplantation has created long waiting lists in most countries. Transplant calculators are available for many organs. The purpose of this study is to describe a simple automatic scoring system for keratoplasty recipient candidates, based on several parameters that we consider most relevant for tissue allocation, and to compare the system's accuracy in predicting decisions made by a cornea specialist. METHODS: Twenty pairs of candidate data were randomly created on an electronic spreadsheet. A single priority score was computed from the data of each candidate. A cornea surgeon and the automated system then decided independently which candidate in each pair should have surgery if only a single cornea was available. RESULTS: The scoring system can calculate values between 0 (lowest priority) and 18 (highest priority) for each candidate. Average score value in our randomly created cohort was 6.35 ± 2.38 (mean ± SD), range 1.28 to 10.76. Average score difference between the candidates in each pair was 3.12 ± 2.10, range 0.08 to 8.45. The manual scoring process, although theoretical, was mentally and emotionally demanding for the surgeon. Agreement was achieved between the human decision and the calculated value in 19 of 20 pairs. Disagreement was reached in the pair with the lowest score difference (0.08). CONCLUSIONS: With worldwide donor cornea shortage, waiting for transplantation can be long. Manual sorting of priority for transplantation in a long waiting list is difficult, time-consuming and prone to error. The suggested system may help achieve a justified distribution of available tissue.

Classification of secondary corneal amyloidosis and involvement of lactoferrin.

PURPOSE: To classify secondary corneal amyloidosis (SCA) by its clinical appearance, to analyze the demographics of the patients, and to determine the involvement of lactoferrin. DESIGN: Retrospective, observational, noncomparative, multicenter study. PARTICIPANTS: Twenty-nine eyes of 29 patients diagnosed with SCA by corneal specialists at 9 ophthalmologic institutions in Japan were studied. METHODS: The clinical appearance of SCA was determined by slit-lamp biomicroscopy and was classified into 3 types. The demographics of the patients, for example, age, gender, and the duration of the basic disease (trichiasis, keratoconus, and unknown), were determined for each clinical type. Surgically excised tissues were stained with Congo red and antilactoferrin antibody. The postoperative prognosis also was determined. MAIN OUTCOME MEASURES: Clinical appearance of the 3 types of SCA, along with the gender, age, and duration of the basic diseases were determined. RESULTS: Classification of SCA into 3 types based on clinical appearance found 21 cases with gelatinous drop-like dystrophy (GDLD)-like appearance (GDLD type), 3 cases with lattice corneal dystrophy (LCD)-like appearance (LCD type), and 5 cases with the combined type. Patients with the GDLD type were younger (average age: 40.9 years for the GDLD type, 74.3 years for the LCD type, and 46.8 years for the combined type), predominantly women (85.7% for the GDLD type, 33.3% for the LCD type, and 60% for the combined type), and had the basic disease over a longer time (average duration: 22.1 years for the GDLD type, 14.0 for the LCD type, and 11.4 for the combined type). The distribution of the basic diseases (trichiasis vs. keratoconus vs. unknown) was not significantly different for each type. Surgical treatments, for example, phototherapeutic keratectomy, lamellar keratoplasty, and simple keratectomy, resulted in a good resolution in all surgically treated cases. One subject dropped out of the study. Spontaneous resolution was seen in one subject after epilation of the cilia. Amorphous materials in the excised tissues showed positive staining results by Congo red and by antilactoferrin antibody. CONCLUSIONS: Secondary corneal amyloidosis can be classified into 3 clinical types based on its clinical appearance. Larger numbers of females and lactoferrin expression were seen in all 3 types. FINANCIAL DISCLOSURE(S): The author(s) have no proprietary or commercial interest in any materials discussed in this article.

The 7-year cumulative incidence of cornea guttata and morphological changes in the corneal endothelium in the Reykjavik Eye Study.

PURPOSE: To examine the corneal endothelium and establish the 7-year cumulative incidence of cornea guttata (CG). METHODS: Population-based prospective cohort study with 573 participants (third wave of the Reykjavik Eye Study (RES) in 2008). Four hundred and thirty-seven subjects had either right or left eyes available for analysis after excluding confounding eye conditions. The baseline for eyes at risk for developing CG is the second wave of the RES in 2001. Participants underwent specular microscopy and a standardized eye examination. RESULTS: The cumulative 7-year incidence of CG in either eye was estimated as a 95% confidence interval for the expected value for both genders combined (15-23%), for males (8-18%) and for females (19-29%). In right eye only, the 7-year cumulative incidence for both genders combined was estimated to be 6-11%. For genders combined and for males only, the data indicated no correlation between 7-year cumulated incidence and age at baseline. In women, however, the change of 7-year incidence for CG in at least one eye appeared to be correlated to age at baseline. Reduction of endothelial cell density for corneas with CG at baseline was found [CI (0.95)-132 ± 94]. CONCLUSION: The cumulative 7-year incidence of primary central CG for a middle-aged and older Caucasian population without history of potentially confounding eye disease has been established. Women tend to have higher incidence if onset occurs at middle age. If CG is present, the cell density and the cell size variation decrease within a 7-year period.

A review of corneal melting following kerato-refractive surgery.

Corneal melting is a rare complication that may occur following a number of different types of surgery. Keratolysis may lead to scarring, irregular astigmatism, photophobia and decreased vision. This article reviews the incidence, pathophysiology and treatment of this condition in kerato-refractive surgery.