arXiv Science⌕ Search

arXiv · 2611.18760

Information about Optic Nerve Head Biomechanics is Already Coded in Your Clinical Charts and Can be Used to Predict Glaucoma Progression

Abstract

Glaucoma patients routinely undergo optical coherence tomography (OCT) imaging and intraocular pressure (IOP) assessment as part of standard care. Because IOP typically differs across visits, it is possible to map visit-to-visit ONH strains by registering OCT volumes acquired under different IOPs. These strains reflect ONH elasticity and, over longer intervals, also capture growth and remodeling of the ONH. In this study, we evaluated whether such visit-to-visit ONH strain signatures could serve as predictors of glaucoma progression. We retrospectively analyzed 525 eyes from 430 glaucoma patients (238 fast and 287 slow progressors), defining fast progression by a visual field index slope of -1.5% per year or worse. 3D strain maps were derived from baseline and follow-up OCT scans using AI. Predictions using T=1 to T=4 follow-up visits were evaluated by five-fold cross-validation in a long-term cohort (T=1: 1.30 years; T=4: 5.26 years) and short-term testing (T=1: 0.46 years; T=4: 1.25 years). Overall, IOP differed between paired visits in 92% of cases. Long-term T=1 prediction achieved an area under the receiver operating characteristic curve (AUC) of 0.74 and accuracy of 0.72. A single short-term follow-up (T=1) yielded an AUC of 0.75 and accuracy of 0.71; four follow-ups yielded an AUC of 0.87 and accuracy of 0.82. SHapley Additive exPlanations (SHAP) identified neuroretinal rim compression and prelaminar elongation associated with fast progression, with substantial inter-eye variability. Overall, visit-to-visit ONH strains derived from routine OCT predicted glaucoma progression after a single follow-up visit at a mean interval of 0.46 years. These findings suggest that short-term OCT/IOP assessments could provide valuable insights into ONH biomechanics, identify eyes at risk of rapid progression, and reveal distinct strain signatures associated with disease progression.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Yibo Chen, Patipol Tiyajamorn, Charis Y. N. Chiang, Deepta Ghate, Thanadet Chuangsuwanich, Michaël J. A. Girard. 2026-11-11. Information about Optic Nerve Head Biomechanics is Already Coded in Your Clinical Charts and Can be Used to Predict Glaucoma Progression. https://arxiv.science/abs/2611.18760

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite arXiv Science by Indiciae: BibTeX · RIS.

KEEP EXPLORING

Related papers

Physics-Aware Zero-Shot Learning for Low-Dose CT from a Single Sinogram

Low-dose computed tomography (CT) imaging remains a persistent challenge, especially under heterogeneous acquisition conditions. Zero-shot learning, which enables models to learn directly from individual images without external supervision, represents a compelling approach to mitigate distribution mismatches arising from variations in scanner hardware, imaging protocols, and patient demographics. However, most existing zero-shot approaches employ down-sampling operations that reduce spatial resolution and cause domain shifts, which become particularly evident when applied to raw sinogram data. Here, we propose a physics-aware zero-shot CT denoising framework to exploit the conjugate symmetry property intrinsic to CT sinograms. Conjugate X-ray pencil beams traverse identical anatomical paths and ideally yield the same expected attenuation values but with independent noise realizations. By randomly mapping sinogram values between conjugate-ray pairs, we synthesize structurally consistent yet noise-diverse sinogram pairs, enabling self-supervised network training without external data. The resulting lightweight denoising network demonstrates strong generalization across low-dose simulation, clinical helical CT, and preclinical photon-counting CT datasets. Quantitative and qualitative results demonstrate competitive denoising performance, with favorable noise suppression and structural preservation across diverse acquisition settings. This work establishes a generalizable and data-efficient paradigm for CT denoising that integrates physical redundancy into learning-based image reconstruction. The code for the proposed method is openly available at: https://github.com/shiyongyi/Physics-Aware-Zero-Shot-CT-Sinogram-Denoising

eess.IV↗

QUTCC: Quantile Uncertainty Training and Conformal Calibration for Imaging Inverse Problems

While deep learning offers tremendous promise for scientific and medical imaging, any failures and hallucinations (predictions that do not coincide with reality) are hard to pinpoint and can have serious downstream consequences. Uncertainty estimation techniques, such as conformal prediction, can help by predicting statistically valid error bars for a model's prediction. However, popular conformal prediction methods were not designed for high-dimensional image-valued problems and do not take into account spatial correlations within an image during conformal calibration, resulting in larger-than-necessary uncertainty intervals. We propose a practical simultaneous quantile regression method that enables non-linear, spatially-adaptive scaling during conformal calibration. Our method, QUTCC uses a U-Net architecture with a quantile embedding to learn a full conditional quantile distribution during training, and then leverages this non-linear, learned function for spatially-adaptive conformal calibration. At test time, our method can efficiently estimate uncertainty intervals with pixel-marginal coverage guarantees. In addition, QUTCC can also predict pixel-wise conditional probability density estimates without any built-in distributional assumptions. We evaluate our method on several denoising problems, accelerated magnetic resonance imaging, and quantitative phase microscopy. Our method consistently produces tighter uncertainty intervals than prior conformal methods at the same coverage level, can predict plausible conditional distributions for different tasks, and in some cases, high-uncertainty regions can help us locate hallucinations in a model's prediction.

eess.IV↗

GroundGazer: An optical reference system for planar localization with millimeter accuracy at low cost

Highly accurate indoor localization systems with absolute mm positioning accuracy are currently expensive. They include laser trackers, total stations, and motion capture systems relying on multiple high-end cameras. In this work, we introduce a high-accuracy, planar indoor localization system named GroundGazer (GG) for autonomous mobile robots (AMRs). GG estimates the AMR's planar position with mm and its heading with sub-degree accuracy. The system requires only a monocular (fisheye) camera, a chessboard floor, and an optional laser diode. Our system is simple and low-cost due to the chessboard floor, robust, scalable to multiple robots, and extendable to 3D position and orientation estimation.

eess.IV↗