高山 卓也

Retinal breaks are critical lesions that can cause retinal detachment and vision loss if not detected and treated early. Automated, accurate delineation of retinal breaks in ultra-widefield fundus (UWF) images remains challenging. In this study, we developed and validated a deep learning segmentation model based on the PraNet architecture to localize retinal breaks in break-positive cases. We trained and evaluated the model using a dataset comprising 34,867 UWF images of 8083 cases. Performance was assessed using image-level segmentation metrics, including accuracy, precision, recall, Intersection over Union (IoU), dice score, and centroid distance score. The model achieved an accuracy of 0.996, precision of 0.635, recall of 0.756, IoU of 0.539, dice score of 0.652, and centroid distance score of 0.081. To our knowledge, this is the first study to present pixel-level segmentation of retinal breaks in UWF images using deep learning. The proposed PraNet-based model showed high accuracy and robust segmentation performance, highlighting its potential for clinical application.

Aflibercept and brolucizumab, two anti-VEGF agents used as intravitreal injections in ophthalmology, differ significantly in molecular weight (aflibercept-115 kDa and brolucizumab-26 kDa). Using aqueous humor samples collected after drug administration, we measured and performed a comparative analysis of pharmacokinetics and half-lives of these drugs in the human eye. Since the quantification of monoclonal antibodies (mAbs) using antigen-antibody reactions, such as ELISA, is influenced by endogenous ligands or anti-drug antibodies, we employed nano-surface and molecular-orientation limited proteolysis (nSMOL), combined with liquid chromatography-tandem mass spectrometry (LC-MS/MS), for accurate measurements. Aqueous humor samples were collected from 59 eyes of 59 patients treated with aflibercept and 52 eyes of 52 patients treated with brolucizumab. Samples were obtained with a median post-injection period of 30 (range, 2-49) days for aflibercept and 28 (range, 4-60) days for brolucizumab. A population pharmacokinetic (PPK) analysis revealed that the half-life of aflibercept in human aqueous humor was significantly shorter than that of brolucizumab, 2.88 days versus 9.00 days, respectively (p = 1.16 × 10-7). Using the same mass spectrometry conditions, we calculated the half-lives of the two drugs. These results may be useful for optimizing the efficacy of these drugs in clinical practice.

PURPOSE: We report a rare case of microbial keratitis caused by Nigrospora oryzae. OBSERVATIONS: A 72-year-old Japanese woman was injured by plant debris and developed oval corneal ulcers and hypopyon in the anterior chamber. After 5 days, she complained of pain, redness, and vision loss in her left eye and was treated with antibacterial eye drops and an ointment (1.5 % levofloxacin hydrate, cefmenoxime hydrochloride, and sterilization and disinfection eye drops; SAN-IODE and ofloxacin ophthalmic ointment). Examination revealed a worsening oval corneal ulcer with Descemet's folds and a faint hypopyon. Considering the infection from soil or plants and the poor response to intensive antibacterial eye drops, topical antifungal eye drops, i.e., 1 % voriconazole eye drops, and 1 % natamycin ointment were applied. Direct microscopy of the corneal scraping with Gram staining was performed and the result was negative. Cultures from corneal scrapings showed the growth of dark colonies after several days. The colony was identified as Nigrospora oryzae by sequencing of the fungal internal transcribed spacer region. Pain and vision loss improved with improvement in corneal ulcers. The antifungal treatment was administered for 37 days. Discontinuation of the eye drops after 1 month did not result in keratitis recurrence. At the final follow-up at 70 days, the best-corrected visual acuity was 20/25, with persistent small corneal opacity. CONCLUSIONS AND IMPORTANCE: Here, we report a case of fungal keratitis caused by Nigrospora oryzae. Microbiological identification of the causes of rare infections is difficult in clinical laboratories, necessitating the use of advanced molecular techniques based on amplification and sequencing of appropriate phylogenetic markers. Nigrospora oryzae responds to topical voriconazole and natamycin.