June 2022
Volume 63, Issue 7
Open Access
ARVO Annual Meeting Abstract  |   June 2022
Machine learning prediction of limited early response to anti-VEGF therapy in neovascular age-related macular degeneration in routine clinical practice.
Author Affiliations & Notes
  • Scott W Perkins
    Cleveland Clinic Lerner College of Medicine of Case Western Reserve University, Cleveland, Ohio, United States
  • Anna K Wu
    Center for Ophthalmic Bioinformatics, Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
    Case Western Reserve University School of Medicine, Cleveland, Ohio, United States
  • Rishi P Singh
    Center for Ophthalmic Bioinformatics, Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
  • Footnotes
    Commercial Relationships   Scott Perkins None; Anna Wu None; Rishi Singh Apellis, Aerie, Graybug, Code F (Financial Support), Novartis, Genentech, Regeneron, Alcon, Bausch and Lomb, 41 Gyroscope, Code I (Personal Financial Interest)
  • Footnotes
    Support  This study was supported in part by the NIH-NEI P30 Core Grant (IP30EY025585), Unrestricted Grants from The Research to Prevent Blindness, Inc., and Cleveland Eye Bank Foundation awarded to the Cole Eye Institute.
Investigative Ophthalmology & Visual Science June 2022, Vol.63, 322 – F0153. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Scott W Perkins, Anna K Wu, Rishi P Singh; Machine learning prediction of limited early response to anti-VEGF therapy in neovascular age-related macular degeneration in routine clinical practice.. Invest. Ophthalmol. Vis. Sci. 2022;63(7):322 – F0153.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : Patients with neovascular age-related macular degeneration (nAMD) have varying responses to anti-vascular endothelial growth factor (anti-VEGF) treatment. Those with residual intraretinal fluid (IRF) and/or subretinal fluid (SRF) after intensive anti-VEGF injections in the first 3 months of therapy have been defined as having limited early response (LER) to treatment. Prior post-hoc studies have associated LER status with reduced long-term visual acuity. Predicting LER from baseline data may be useful in identifying treatment-resistant patients who would benefit from more intensive treatment or targeted selection of anti-VEGF therapy.

Methods : Data was obtained from a retrospective chart review of 286 treatment naïve patients diagnosed with nAMD. Age, best visual acuity (BVA), central subfield thickness (CST), and optical coherence tomography (OCT) data were obtained at baseline, three, six, and twelve months. IRF and SRF were quantified from OCT data using the Notal OCT Analyzer machine learning algorithm (Tel Aviv, Israel). Data were pre-processed by one-hot encoding and min-max scaling. Predictive models were trained and evaluated using a 10-fold cross-validated approach.

Results : LER was predicted from baseline age, BVA, IRF, SRF, and CST by ridge logistic regression, k nearest neighbors classification, and radial basis function kernel support vector classification with 10-fold cross-validated accuracy of 0.66, 0.63, and 0.64, respectively. Area under the cross-validated receiver operating characteristic curve was 0.63, 0.59, and 0.62, respectively. T-distributed stochastic neighbor embedding did not show clusters separated by LER status.

Conclusions : These results demonstrate the potential of machine learning techniques for predicting nAMD outcomes in routine clinical practice. While age, BVA, IRF, SRF, and CST confer predictive accuracy, incorporation of additional variables may further improve predictive models. This would be beneficial for supporting physician decisions regarding anti-VEGF treatment in nAMD. Further studies could incorporate treatment data to determine which anti-VEGF therapies may be most suitable for patients at risk of LER.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.

×
×

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.

×