gms | German Medical Science

Artificial Vision 2019

The International Symposium on Visual Prosthetics

13.12. - 14.12.2019, Aachen

Towards a unified set of performance outcomes for vision restoration trials

Meeting Abstract

  • Gislin Dagnelie - Departments of Ophthalmology, Johns Hopkins University, Baltimore/USA
  • A. Kartha - Departments of Ophthalmology, Johns Hopkins University, Baltimore/USA
  • R. Sadeghi - Departments of Ophthalmology, Johns Hopkins University, Baltimore/USA; Departments of Biomedical Engineering, Johns Hopkins University, Baltimore/USA
  • C. Bradley - Departments of Ophthalmology, Johns Hopkins University, Baltimore/USA
  • D. Geruschat - Departments of Ophthalmology, Johns Hopkins University, Baltimore/USA

Artificial Vision 2019. Aachen, 13.-14.12.2019. Düsseldorf: German Medical Science GMS Publishing House; 2019. Doc19artvis43

doi: 10.3205/19artvis43, urn:nbn:de:0183-19artvis432

Veröffentlicht: 10. Dezember 2019

© 2019 Dagnelie et al.
Dieser Artikel ist ein Open-Access-Artikel und steht unter den Lizenzbedingungen der Creative Commons Attribution 4.0 License (Namensnennung). Lizenz-Angaben siehe http://creativecommons.org/licenses/by/4.0/.


Gliederung

Text

Objective: The work of the HOVER taskforce has demonstrated that the need, and the ability, to reach consensus about common instruments to measure outcomes in vision restoration trials. Still, the resulting range of tests is large and diverse, and further consolidation may be possible. In view of the wide range at baseline among vision restoration trial participants – from bare light perception (LP) through ultra low vision (ULV, i.e. worse than 20/1600) up to legal blindness, 20/200 – the diverse study population (age, geography) and the potential for large treatment effects this is a major challenge.

Methods: Development of the ULV-VFQ demonstrated that a single well-developed self-report instrument can measure visual ability over a wide range, and cover all important aspect of daily vision use. Our goal in the present project was to complement the ULV-VFQ with a set of performance measures covering an equally broad range, and to rely as little as possible on environmental factors. We chose to implement 19 visual 2-, 3- and 4-AFC judgments, each at 3 difficulty levels. These activities are carried out in a virtual reality (VR) headset, and thus are called ULV-AVR. Our goal is to collect data in 200 subjects with visual acuity (VA) ranging from bare light perception to 20/400.

Results: To date we have collected measures in 30 volunteers with visual acuities ranging from LP to 20/400. Activities and subjects can be scaled along the ability/difficulty dimension using an analytic tool based on signal detection theory. Results for both activities and subjects covered the entire LP-to-20/400 range, were highly consistent, and yielded subject measures that are strongly correlated with BRVT-based VA estimates (R²=0.88).

Discussion: Covering a large ability range with a uniform set of measures is critically important for trials that may show a very large difference from baseline to follow-up, such as vision restoration through any of the approaches currently being explored. We are currently expanding the ULV-AVR with hand-eye coordination and mobility activities. This yield a broad range of measures covering most aspects of vision use.