Dataset Information

Artifact Correction in Retinal Nerve Fiber Layer Thickness Maps Using Deep Learning and Its Clinical Utility in Glaucoma.

ABSTRACT:

Purpose

Correcting retinal nerve fiber layer thickness (RNFLT) artifacts in glaucoma with deep learning and evaluate its clinical usefulness.

Methods

We included 24,257 patients with optical coherence tomography and reliable visual field (VF) measurements within 30 days and 3,233 patients with reliable VF series of at least five measurements over ≥4 years. The artifacts are defined as RNFLT less than the known floor value of 50 µm. We selected 27,319 high-quality RNFLT maps with an artifact ratio (AR) of <2% as the ground truth. We created pseudo-artifacts from 21,722 low-quality RNFLT maps with AR of >5% and superimposed them on high-quality RNFLT maps to predict the artifact-free ground truth. We evaluated the impact of artifact correction on the structure-function relationship and progression forecasting.

Results

The mean absolute error and Pearson correlation of the artifact correction were 9.89 µm and 0.90 (P < 0.001), respectively. Artifact correction improved R2 for VF prediction in RNFLT maps with AR of >10% and AR of >20% up to 0.03 and 0.04 (P < 0.001), respectively. Artifact correction improved (P < 0.05) the AUC for progression prediction in RNFLT maps with AR of ≤10%, >10%, and >20%: (1) total deviation pointwise progression: 0.68 to 0.69, 0.62 to 0.63, and 0.62 to 0.64; and (2) mean deviation fast progression: 0.67 to 0.68, 0.54 to 0.60, and 0.45 to 0.56.

Conclusions

Artifact correction for RNFLTs improves VF and progression prediction in glaucoma.

Translational relevance

Our model improves clinical usability of RNFLT maps with artifacts.

SUBMITTER: Shi M

PROVIDER: S-EPMC10631515 | biostudies-literature | 2023 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Artifact Correction in Retinal Nerve Fiber Layer Thickness Maps Using Deep Learning and Its Clinical Utility in Glaucoma.

Shi Min M Sun Jessica A JA Lokhande Anagha A Tian Yu Y Luo Yan Y Elze Tobias T Shen Lucy Q LQ Wang Mengyu M

Translational vision science & technology 20231101 11

<h4>Purpose</h4>Correcting retinal nerve fiber layer thickness (RNFLT) artifacts in glaucoma with deep learning and evaluate its clinical usefulness.<h4>Methods</h4>We included 24,257 patients with optical coherence tomography and reliable visual field (VF) measurements within 30 days and 3,233 patients with reliable VF series of at least five measurements over ≥4 years. The artifacts are defined as RNFLT less than the known floor value of 50 µm. We selected 27,319 high-quality RNFLT maps with a ...[more]

PMID: 37934137

Similar Datasets

Project description:PurposeTo investigate the role of sex on retinal nerve fiber layer (RNFL) thickness at 768 circumpapillary locations based on OCT findings.DesignPopulation-based cross-sectional study.ParticipantsWe investigated 5646 eyes of 5646 healthy participants from the Leipzig Research Centre for Civilization Diseases (LIFE)-Adult Study of a predominantly white population.MethodsAll participants underwent standardized systemic assessments and ocular imaging. Circumpapillary RNFL (cRNFL) thickness was measured at 768 points equidistant from the optic nerve head using spectral-domain OCT (Spectralis; Heidelberg Engineering, Heidelberg, Germany). To control ocular magnification effects, the true scanning radius was estimated by scanning focus. Student t test was used to evaluate sex differences in cRNFL thickness globally and at each of the 768 locations. Multivariable linear regression and analysis of variance were used to evaluate individual contributions of various factors to cRNFL thickness variance.Main outcome measuresDifference in cRNFL thickness between males and females.ResultsOur population consisted of 54.8% females. The global cRNFL thickness was 1 μm thicker in females (P < 0.001). However, detailed analysis at each of the 768 locations revealed substantial location specificity of the sex effects, with RNFL thickness difference ranging from -9.98 to +8.00 μm. Females showed significantly thicker RNFLs in the temporal, superotemporal, nasal, inferonasal, and inferotemporal regions (43.6% of 768 locations), whereas males showed significantly thicker RNFLs in the superior region (13.2%). The results were similar after adjusting for age, body height, and scanning radius. The superotemporal and inferotemporal RNFL peaks shifted temporally in females by 2.4° and 1.9°, respectively. On regions with significant sex effects, sex explained more RNFL thickness variance than age, whereas the major peak locations and interpeak angle explained most of the RNFL thickness variance unexplained by sex.ConclusionsSubstantial sex effects on cRNFL thickness were found at 56.8% of all 768 circumpapillary locations, with specific patterns for different sectors. Over large regions, sex was at least as important in explaining the cRNFL thickness variance as was age, which is well established to have a substantial impact on cRNFL thickness. Including sex in the cRNFL thickness norm could therefore improve glaucoma diagnosis and monitoring.

Dataset Information

Artifact Correction in Retinal Nerve Fiber Layer Thickness Maps Using Deep Learning and Its Clinical Utility in Glaucoma.

Purpose

Methods

Results

Conclusions

Translational relevance

Publications

Artifact Correction in Retinal Nerve Fiber Layer Thickness Maps Using Deep Learning and Its Clinical Utility in Glaucoma.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets