March 2012
Volume 53, Issue 14
Free
ARVO Annual Meeting Abstract  |   March 2012
Latent Class Regression (LCR) Analysis for Detecting Glaucoma Progression
Author Affiliations & Notes
  • Gadi Wollstein
    UPMC Eye Center, Univ of Pittsburgh Sch of Med, Pittsburgh, Pennsylvania
  • Richard A. Bilonick
    UPMC Eye Center, Univ of Pittsburgh Sch of Med, Pittsburgh, Pennsylvania
  • Hiroshi Ishikawa
    UPMC Eye Center, Univ of Pittsburgh Sch of Med, Pittsburgh, Pennsylvania
  • Larry Kagemann
    UPMC Eye Center, Univ of Pittsburgh Sch of Med, Pittsburgh, Pennsylvania
  • Ian A. Sigal
    UPMC Eye Center, Univ of Pittsburgh Sch of Med, Pittsburgh, Pennsylvania
  • Jay S. Duker
    Ophthalmology, New England Eye Center, Boston, Massachusetts
  • Joel S. Schuman
    UPMC Eye Center, Univ of Pittsburgh Sch of Med, Pittsburgh, Pennsylvania
  • Footnotes
    Commercial Relationships  Gadi Wollstein, None; Richard A. Bilonick, None; Hiroshi Ishikawa, None; Larry Kagemann, None; Ian A. Sigal, None; Jay S. Duker, Carl Zeiss Meditec (F), Optovue (F), Topcon (F); Joel S. Schuman, Carl Zeiss Meditec (P)
  • Footnotes
    Support  NIH R01-EY013178, P30-EY008098; Eye and Ear Foundation (Pittsburgh, PA); Research to Prevent Blindness
Investigative Ophthalmology & Visual Science March 2012, Vol.53, 218. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Gadi Wollstein, Richard A. Bilonick, Hiroshi Ishikawa, Larry Kagemann, Ian A. Sigal, Jay S. Duker, Joel S. Schuman; Latent Class Regression (LCR) Analysis for Detecting Glaucoma Progression. Invest. Ophthalmol. Vis. Sci. 2012;53(14):218.

      Download citation file:


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

      ×
  • Supplements
Abstract
 
Purpose:
 

Detection of glaucoma progression poses a major clinical challenge. LCR is a statistical method that allows the identification of groups within population by extracting the shared latent information among various parameters. The purpose of this study was to use LCR with longitudinal structural and functional measurements to characterize glaucoma progression grouping.

 
Methods:
 

125 eyes from healthy, glaucoma suspects and glaucoma subjects were enrolled all with ≥5 reliable visual field (VF; Carl Zeiss Meditec (CZM)) tests and good quality optical coherence tomography (OCT; Stratus, CZM) scans all acquired with a maximum interval of 6 months. VF parameters used for the analysis included mean deviation (MD), pattern standard deviation (PSD) and visual field index (VFI). Mean retinal nerve fiber layer (RNFL) thickness was used from OCT. Akaike information criterion analysis determined the optimal number of groups in the LCR analysis model. LCR analysis was employed with the 4 responses handled as a function of subject age.

 
Results:
 

Five distinctive groups were detected and labeled as Stable low (low rate of change; 43 eyes) and high (38), and Progressing low (20), moderate (11) and high (13) (Figure). The groups showed a gradual steepening rate of change and intercept for both VF and RNFL for stable and progressing eyes and among their subgroups (with the exception of the Progression high group for some parameters). Most VF parameters slopes were significantly different (p<0.05) among the subgroups while RNFL was significantly different in between the Stable groups and between Stable hi and Progression moderate groups only.

 
Conclusions:
 

LCR analysis allows grouping of eyes sharing latent structural and functional information. This could be of importance in situations such as assessment of glaucoma progression - and potentially the prediction of glaucoma progression - especially since widely accepted criteria for progression are lacking.  

 
Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • visual fields • grouping and segmentation 
×
×

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.

×