Pediatric ocular melanoma: a collaborative multicenter study and meta-analysis.

PURPOSE: To investigate clinical manifestations and prognoses in pediatric patients (≤12 years old) with ocular melanoma. METHODS: This was a retrospective, multicenter cohort study with individual participant data (IPD) meta-analysis pooling available published cases, and unpublished cases from an international collaboration of seven ocular oncology centers. RESULTS: There were 133 eyes of 133 pediatric patients with choroidal or ciliary body (n = 66 [50%]), iris (n = 33 [25%]), conjunctival (n = 26 [19%]), and eyelid (n = 8 [6%]) melanoma. Overall, the mean patient age at presentation was 7 years (median, 8; range, 1-12 years), with 63 males (49%). The mean age by tumor site was 6.50 ± 3.90, 7.44 ± 3.57, 9.12 ± 2.61, and 5.63 ± 2.38 years, for choroid/ciliary body, iris, conjunctiva, and eyelid melanoma, respectively (P = 0.001). Association with ocular melanocytosis was seen in 15%, 11%, 4%, and 0%, respectively (P = 0.01). Frequency of ocular melanoma family history did not vary by tumor site (7%, 17%, 9% and 12%, resp. [P = 0.26]). After mean follow-up of 74, 85, 50, and 105 months (P = 0.65), metastasis was seen in 12%, 9%, 19%, and 13% of choroid/ciliary body, iris, conjunctiva, and eyelid melanoma, respectively. Death was reported in 5%, 3%, 8%, and 0%, respectively, with survival analysis indicating higher mortality in choroidal/ciliary body and conjunctival melanoma patients. CONCLUSIONS: Ocular melanoma in the pediatric population is rare, with unique clinical features and outcomes. Iris melanoma accounts for about one-third of pediatric uveal melanoma cases.

Clinical Features, Risk Factors, and Management of Candida Keratitis.

BACKGROUND/AIMS: To determine the clinical features, predisposing factors, and management of infectious keratitis caused by Candida spp. METHODS: Retrospective chart review. RESULTS: The medical records of 52 patients (54 eyes) with Candida keratitis were available for statistical analysis. Thinning of the corneal stroma was identified in 34 eyes (63.0%), and corneal perforation occurred in 16 eyes (29.6%). Corneal thinning and perforation were more common in Candida albicans compared with non-albicans (P-val < .001, P = .09, respectively). The most common predisposing factors for Candida keratitis were topical steroid use (21 patients, 40.4%), previous corneal transplantation (17 patients, 32.7%), and preexisting ocular surface disease (15 patients, 28.8%). Fourteen eyes (25.9%) required cyanoacrylate glue application and 10 eyes (18.5%) underwent therapeutic penetrating keratoplasty (TPK). CONCLUSION: Local immunosuppression and ocular surface disease play an important role in Candida keratitis. C. albicans appears to be more invasive compared with non-albicans spp.

CNV-Net: Segmentation, Classification and Activity Score Measurement of Choroidal Neovascularization (CNV) Using Optical Coherence Tomography Angiography (OCTA).

This paper aims to present an artificial intelligence-based algorithm for the automated segmentation of Choroidal Neovascularization (CNV) areas and to identify the presence or absence of CNV activity criteria (branching, peripheral arcade, dark halo, shape, loop and anastomoses) in OCTA images. Methods: This retrospective and cross-sectional study includes 130 OCTA images from 101 patients with treatment-naïve CNV. At baseline, OCTA volumes of 6 × 6 mm2 were obtained to develop an AI-based algorithm to evaluate the CNV activity based on five activity criteria, including tiny branching vessels, anastomoses and loops, peripheral arcades, and perilesional hypointense halos. The proposed algorithm comprises two steps. The first block includes the pre-processing and segmentation of CNVs in OCTA images using a modified U-Net network. The second block consists of five binary classification networks, each implemented with various models from scratch, and using transfer learning from pre-trained networks. Results: The proposed segmentation network yielded an averaged Dice coefficient of 0.86. The individual classifiers corresponding to the five activity criteria (branch, peripheral arcade, dark halo, shape, loop, and anastomoses) showed accuracies of 0.84, 0.81, 0.86, 0.85, and 0.82, respectively. The AI-based algorithm potentially allows the reliable detection and segmentation of CNV from OCTA alone, without the need for imaging with contrast agents. The evaluation of the activity criteria in CNV lesions obtains acceptable results, and this algorithm could enable the objective, repeatable assessment of CNV features.