Differentiating malignant and benign eyelid lesions using deep learning.

Artificial intelligence as a screening tool for eyelid lesions will be helpful for early diagnosis of eyelid malignancies and proper decision-making. This study aimed to evaluate the performance of a deep learning model in differentiating eyelid lesions using clinical eyelid photographs in comparison with human ophthalmologists. We included 4954 photographs from 928 patients in this retrospective cross-sectional study. Images were classified into three categories: malignant lesion, benign lesion, and no lesion. Two pre-trained convolutional neural network (CNN) models, DenseNet-161 and EfficientNetV2-M architectures, were fine-tuned to classify images into three or two (malignant versus benign) categories. For a ternary classification, the mean diagnostic accuracies of the CNNs were 82.1% and 83.0% using DenseNet-161 and EfficientNetV2-M, respectively, which were inferior to those of the nine clinicians (87.0-89.5%). For the binary classification, the mean accuracies were 87.5% and 92.5% using DenseNet-161 and EfficientNetV2-M models, which was similar to that of the clinicians (85.8-90.0%). The mean AUC of the two CNN models was 0.908 and 0.950, respectively. Gradient-weighted class activation map successfully highlighted the eyelid tumors on clinical photographs. Deep learning models showed a promising performance in discriminating malignant versus benign eyelid lesions on clinical photographs, reaching the level of human observers.

Atypical Mycobacterial Infection in Anophthalmic Sockets With Porous Orbital Implant Exposure.

PURPOSE: To investigate the clinical features and risk factors of atypical mycobacterial infection in anophthalmic sockets with porous orbital implant exposure. DESIGN: Case-control study. METHODS: The medical records of all patients who had undergone surgical correction of porous orbital implant exposure were consecutively reviewed, and the patients were stratified as those with atypical mycobacterial infection (AM infection group) and others (non-AM group). RESULTS: Five and 21 patients were included in the AM infection and non-AM groups, respectively. All patients of the AM infection group had a peg or motility coupling post (MCP) and showed implant exposure around it. Following up on implant exposure, 2 patients abruptly presented with severe conjunctival injection and new lesions such as erythematous nodules or eyelid masses. They underwent immediate orbital implant exchange and atypical mycobacterial infection was diagnosed. Three patients who were not suspected of having infection underwent surgery for orbital implant exposure. Results revealed erythematous eyelid nodule or recurrent exposure shortly after surgery and patients were diagnosed with atypical mycobacterial infection. In the non-AM group, 7 (33.3%) patients underwent insertion of a peg or MCP. Statistical analysis showed that the insertion of a peg or MCP was the only risk factor showing a significant difference between the 2 groups. CONCLUSIONS: The most important underlying risk factor for atypical mycobacterial infection in the anophthalmic socket is thought to be peg- or MCP-related exposure of the porous orbital implant. Surgical removal of the infected orbital implant and long-term antibiotic medication are needed for treatment.

Long-term results of fluocinolone acetonide intravitreal implant in Behçet intractable posterior uveitis.

OBJECTIVE: To evaluate the long-term efficacy and safety of the fluocinolone acetonide intravitreal implant in patients with Behçet disease with intractable noninfectious posterior uveitis. DESIGN: Consecutive retrospective analysis. PARTICIPANTS: Eight eyes from 7 patients with Behçet uveitis who did not respond successfully to conventional treatment with topical and systemic steroids and/or systemic steroid-sparing agents were studied. METHODS: We performed a chart review of patients who were treated with a 0.59-mg fluocinolone acetonide intravitreal implant at a single centre from September 2007 through June 2009. Snellen visual acuity, control of inflammation, and the development of complications such as infection or uncontrollable intraocular pressure (IOP) elevation were evaluated. RESULTS: Mean age at implant placement was 35.3 (range 17-42) years. Mean follow-up duration was 47.8 (range 39.5-57.6) months. Postoperative visual acuity improved by more than 3 lines in 6 eyes (75%). Five patients were able to discontinue all systemic medications. Six eyes (75%) exhibited postoperative IOP spikes of more than 30 mm Hg. Five patients required glaucoma shunting surgery postoperatively for IOP control. The single phakic eye developed a visually significant posterior subcapsular lens opacification that required cataract extraction. There was 1 case of postoperative cytomegalovirus endothelitis. Infection was controlled with oral valganciclovir and topical antibiotic medication, and the patient did not require implant removal. CONCLUSIONS: The fluocinolone implant is effective in the control of intractable inflammation in Behçet uveitis. Elevation of IOP remains a major potential complication, and the possibility of infection should be considered.