A snapshot of current research and development initiatives spanning genetics, imaging, devices, and patient-centered tools.
High myopia (≤ −6.00 D or axial length ≥ 26 mm) carries substantially elevated risks of sight-threatening complications including retinal detachment, myopic macular degeneration, and glaucoma. Our work investigates the genetic architecture of pathological axial elongation through large-scale GWAS, whole-genome sequencing, and epidemiological surveys in East Asian populations.
Whole-genome sequencing of extreme myopia families has identified a large duplication on chromosome 3 in a multigenerational pedigree. Overexpression of genes within the duplicated region was shown to induce myopia-like axial elongation in zebrafish, and we are currently generating conditional knock-in mice to dissect the underlying mechanism.
These studies aim to identify the epidemiological and genetic risk factors for the development of pathological myopia, and to uncover modifiable determinants and predictive biomarkers for disease progression and its sight-threatening complications.
Pachychoroid spectrum disease — encompassing central serous chorioretinopathy (CSC), pachychoroid neovasculopathy (PNV), and polypoidal choroidal vasculopathy (PCV) — is unified by dilated outer choroidal vessels and increased choroidal venous congestion. Our research focuses on elucidating the genetic determinants, natural history, and optimal treatment strategies across this spectrum, using multimodal retinal imaging including OCT angiography.
We were the first to demonstrate that a substantial proportion of Asian patients clinically diagnosed with AMD harbor PNV as the underlying neovascular subtype — a finding that has reshaped understanding of macular neovascularization in Asian populations. We are also contributing to the APAO diagnostic criteria for pachychoroid drusen, an emerging entity within this spectrum.
On the clinical side, we run Japan's only dedicated CSC specialty clinic, and are currently conducting a physician-initiated clinical trial of reduced-fluence PDT (rf-PDT) — the REPLAY trial — to support regulatory approval of this treatment for CSC in Japan.
Conventional glaucoma drainage devices are passive and do not permit postoperative adjustment of intraocular pressure (IOP), risking hypotony or inadequate pressure control. We are developing a next-generation adjustable microfluidic glaucoma implant that enables non-invasive, postoperative fine-tuning of aqueous humor outflow resistance, with the goal of achieving non-inferiority to trabeculectomy while maintaining a minimally invasive profile.
The device incorporates a compact adjustment mechanism developed in collaboration with the Institute for Integrated Cell-Material Sciences (iCeMS) at Kyoto University, whose expertise in micro-scale materials and fabrication has been central to realizing a clinically viable implant form factor.
The project has attracted over ¥55M in competitive public funding, including the AMED Bridging Research Program (Seeds H — Interdisciplinary Fusion), Seeds A, and the Next-Generation Healthcare Startup Support Program, reflecting its translational promise.
Standard automated perimetry relies on subjective button-press responses and is susceptible to patient fatigue and fixation errors. The Gaze Analyzing Perimeter (GAP) is a head-mounted device that tracks eye movements at up to 240 fps and determines stimulus visibility entirely from gaze behavior — eliminating button responses.
The core principle is that when a newly presented stimulus is visible, gaze shifts directly and linearly toward it; when the stimulus falls within a scotoma or is not perceived, gaze either moves elsewhere or fails to reach the target. This reflexive gaze response provides an objective, response-independent measure of visual sensitivity across the visual field.
Clinical studies have demonstrated strong correlation with the Humphrey Field Analyzer, with particular promise for objective visual field assessment in glaucoma screening, monitoring, and populations where reliable button-press responses are difficult to obtain.
Eye movement patterns are disrupted in Alzheimer's disease and mild cognitive impairment (MCI), often preceding clinical diagnosis. Using the Nagahama Study cohort of nearly 10,000 participants, we evaluated multiple gaze parameters — including mean gaze reaction speed (time from stimulus onset to gaze shift), latency, number of saccades, first gain, and end gain — and identified several that show significant associations with cognitive function.
Building on these findings, we are developing a program to detect MCI from gaze behavior during passive viewing of video content, a low-burden approach requiring no active response from the subject. This project has been funded by the AMED Medical Engineering Collaboration and AI Implementation Research Program with approximately ¥300M in competitive funding, and a physician-initiated clinical trial is planned as the next step toward regulatory validation.
Established by the Japanese Ophthalmological Society in 2017, JOIR is a national real-world database that automatically collects fundus photographs, OCT images, and visual field data from 22+ university hospitals and affiliated facilities via secure cloud transfer. As of March 2026, the registry holds over 1.5 million fundus photographs, 600,000+ OCT images, and associated clinical data — one of the largest ophthalmic imaging repositories in the world.
Pre-trained AI models developed from JOIR data are publicly available, supporting broader research use across institutions. A landmark milestone was reached in October 2025, when the first Software as a Medical Device (SaMD) developed through this project received regulatory approval in Japan — the first such approval in Japanese ophthalmology.