June 2023
Volume 64, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2023
Glaucoma subtyping based on visual field progression using unsupervised machine learning
Author Affiliations & Notes
  • Xiaoqin Huang
    The University of Tennessee Health Science Center, Memphis, Tennessee, United States
  • Asma Poursoroush
    The University of Tennessee Health Science Center, Memphis, Tennessee, United States
  • Jian Sun
    National Institute of Allergy and Infectious Diseases, Bethesda, Maryland, United States
  • Louis R Pasquale
    Ophthalmology, Icahn School of Medicine at Mount Sinai, New York, New York, United States
  • Michael Boland
    Department of Ophthalmology, Massachusetts Eye and Ear, Boston, Massachusetts, United States
  • Chris A Johnson
    Department of Ophthalmology and Visual Sciences, University of Iowa Hospitals and Clinics, Iowa City, Iowa, United States
  • Siamak Yousefi
    The University of Tennessee Health Science Center, Memphis, Tennessee, United States
  • Footnotes
    Commercial Relationships   Xiaoqin Huang None; Asma Poursoroush None; Jian Sun None; Louis Pasquale None; Michael Boland None; Chris Johnson None; Siamak Yousefi Remidio, Code F (Financial Support), NIH/NEI, Code F (Financial Support), Research to Prevent Blindness, Code F (Financial Support), Bright Focus Foundation, Code F (Financial Support)
  • Footnotes
    Support  NIH EY033005, NIH EY031725
Investigative Ophthalmology & Visual Science June 2023, Vol.64, 368. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Xiaoqin Huang, Asma Poursoroush, Jian Sun, Louis R Pasquale, Michael Boland, Chris A Johnson, Siamak Yousefi; Glaucoma subtyping based on visual field progression using unsupervised machine learning. Invest. Ophthalmol. Vis. Sci. 2023;64(8):368.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : To identify glaucoma subtypes with different rates of visual field (VF) progression and to characterize subtypes with slow and rapid glaucoma progression based on unsupervised models.

Methods : We developed a latent class mixed model (LCMM) to identify glaucoma subgroups based on mean deviation (MD) trajectories. We characterized the subgroups based on demographic, clinical, ocular, and VF factors at the baseline. We identified risk factors driving rapid glaucoma progression using generalized estimating equation (GEE) and performed survival analysis. The results were justified by visualizing VF defects and clinical characterization at specific visits.

Results : The LCMM model discovered four clusters of eyes including Improvers, Stables, Slow progressors, and Rapid progressors based on their mean of MD decline (0.08, -0.06, -0.21, and -0.45 dB/year, respectively). The number of eyes in clusters were 794 (25%), 1675 (54%), 531 (17%) and 133 (4%), respectively. Eyes with rapid VF progression had higher baseline age, intraocular pressure (IOP), pattern standard deviation (PSD) and refractive error (RE), but lower central corneal thickness (CCT). Rapid progression was associated with calcium channel blockers, gender, heart disease history, diabetes history, African American race, stroke history, and history of migraine headaches.

Conclusions : Unsupervised LCMM can identify glaucoma subtypes based on VF worsening. LCMM-identified rapid glaucoma progressors have some characteristics that were known previously as well as several novel characteristics, including higher history of stroke, heart disease, diabetes, and history of using calcium channel blockers. Those with rapid VF progression were more from African American race, were more likely males, and had a higher incidence of glaucoma conversion. Such glaucoma subtyping can provide guidance for adjusting treatment plans to slow vision loss and improve the quality of life of patients with glaucoma.

This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.

 

Figure 1. The smoothed trajectory of mean deviation (MD) in four clusters. 1-Improvers, 2-Stables, 3-Slow progressors, 4-Rapid progressors.

Figure 1. The smoothed trajectory of mean deviation (MD) in four clusters. 1-Improvers, 2-Stables, 3-Slow progressors, 4-Rapid progressors.

 

Figure 2. Chord diagrams showing clinical variables by cluster.
Calcium: Calcium channel blockers, HeartHX: Heart disease history, DbHX: Diabetes History, OtherHX: Other disease history, RaceB: African American Race, StkHX: Stroke history.

Figure 2. Chord diagrams showing clinical variables by cluster.
Calcium: Calcium channel blockers, HeartHX: Heart disease history, DbHX: Diabetes History, OtherHX: Other disease history, RaceB: African American Race, StkHX: Stroke history.

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×