Issue: May/June 2020
Perspective from Lisa M. Young, OD, FAAO
Disclosures: Wang reports financial ties with the National Institutes of Health, the BrightFocus Foundation, the Lions Foundation, the Grimshaw-Gudewicz Foundation, Research to Prevent Blindness Inc. and Harvard Glaucoma Center of Excellence. Please see the full study for all other authors’ relevant financial disclosures.
March 26, 2020
2 min read
Save

Unsupervised AI can help identify central visual field loss patterns

Issue: May/June 2020
Perspective from Lisa M. Young, OD, FAAO
Disclosures: Wang reports financial ties with the National Institutes of Health, the BrightFocus Foundation, the Lions Foundation, the Grimshaw-Gudewicz Foundation, Research to Prevent Blindness Inc. and Harvard Glaucoma Center of Excellence. Please see the full study for all other authors’ relevant financial disclosures.
You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

Using artificial intelligence to identify central visual field loss patterns better predicted quality of life impairment in patients with glaucoma compared with using central visual field mean deviation alone, according to findings from a retrospective study.

“The preservation of central visual function is essential to the care of glaucoma patients,” Mengyu Wang, PhD, of the Schepens Eye Research Institute, Harvard Medical School, and colleagues wrote.

Wang and colleagues used visual field data from the Glaucoma Research Network consortium for their analysis. The study comprised 8,712 patients with 13,951 Humphrey 10-2 (Carl Zeiss Meditec) test results from 13,951 eyes for cross-sectional analyses and 824 patients with a minimum of five reliable 10-2 test results at 6-month intervals or more from 1,191 eyes for longitudinal analyses.

Researchers used total deviation values to determine central visual field patterns using patients' most recent 10-2 test results. They also used a 24-2 VF within a 3-month window of the 10-2 tests to stage eyes into mild, moderate or severe functional loss. To determine central visual field patterns, they applied an unsupervised AI (archetypal analysis).

The AI identified 17 central visual field patterns for eyes with all severities of functional loss. Further, 11 central visual field patterns were determined for mild glaucoma, 11 for moderate glaucoma and 16 for severe glaucoma. Several arcuate defect patterns were found among the 17 central visual field patterns, and most of the diffuse loss patterns preserved the less vulnerable zone. Similar central visual field defects across different glaucoma stages and unique central visual field defects specific to a certain glaucoma stage were observed. Researchers wrote that this could represent different subtypes of central visual field loss with different “underlying pathophysiologic features and biomarkers.

“The quantification of central visual field loss by spatial pattern analyses with archetypal analysis independently confirmed the observational findings of frequent arcuate defects and relative preservation of the inferotemporal less vulnerable zone,” they said.

“Our model could be used by clinicians to decide which patients or eyes need to be treated more aggressively to slow down visual field worsening,” they concluded. “More work is needed to demonstrate the correlation between central visual field patterns and the structural damage, perhaps using macular ganglion cell thickness maps.”