Full-Field Stimulus Test for Visual Function Assessment in Ultra-Low Vision with Retinitis Pigmentosa

This study demonstrates that the Full-field Stimulus Test (FST) effectively quantifies residual visual function in patients with ultra-low vision due to retinitis pigmentosa, showing strong correlations with performance-based daily activity tasks and supporting its use as a clinically meaningful endpoint for therapeutic trials.

The Gist

The Big Picture: Measuring the "Dimmest Light"

Imagine you are trying to judge how well a person can see in a world that is almost pitch black. For people with a severe eye disease called Retinitis Pigmentosa (RP), their vision has faded so much that standard eye charts (the ones with the big "E" at the top) are useless. They might only see a hand waving, a finger counting, or nothing at all.

The researchers wanted to find a better way to measure if these patients have any light left in their eyes, and if that tiny bit of light actually helps them do daily tasks. They used a tool called the Full-Field Stimulus Test (FST).

Think of the FST not as a test of sharpness (like reading small letters), but as a test of sensitivity (like a night-vision camera). Instead of asking, "Can you read this word?" it asks, "Did you see a flash of light anywhere in your entire field of view?"

The Problem with Old Rulers

Usually, doctors measure vision with a ruler that stops at "Counting Fingers." If you can't count fingers, the ruler breaks. You are just labeled "No Light Perception" (NLP).

The problem is that two people labeled "No Light Perception" might be very different. One might be truly blind, while the other might just be missing the "ruler" to measure their tiny bit of remaining sight. This study tried to build a new, more sensitive ruler that can measure those tiny, faint glimmers of light that standard tests miss.

How They Tested It

The team studied 35 patients with severe RP. They did three main things:

1. The Light Flash Test (FST): After sitting in a dark room for 45 minutes (to let their eyes get super sensitive), patients were shown brief flashes of white light. The machine measured the dimmest flash they could detect.
2. The "Real Life" Obstacle Course: To see if the light flashes actually mattered, they gave patients two types of tasks:
   • The Table Game: Patients had to find and identify objects (like cups or geometric shapes) on a table.
   • The Walking Game: Patients had to walk toward a "door" (a black cloth on a wall) or follow a white line on a black floor.
3. The Questionnaires: They asked patients how their vision affected their daily life and happiness.

What They Found

1. The New Ruler Works
The FST was able to detect light in patients who were officially labeled as having "No Light Perception." In fact, 64% of the patients who were supposed to be completely blind could actually see the flashes. It's like finding out that a room thought to be pitch black actually has a tiny nightlight on.

2. The Connection to Real Life
This is the most important part. The researchers found a strong link between the FST numbers and the real-life games.

• The Analogy: Imagine the FST score is the "battery level" of a flashlight. The Table and Walking games are the tasks the flashlight is used for.
• The Result: When the "battery" (FST score) got higher, the patients were much better at finding objects on the table and walking toward the door.
• The Threshold: They found a specific "battery level" (a specific FST score) where patients suddenly went from being unable to do the tasks to being able to do them. It's like a light switch: once the light gets bright enough, the room becomes usable.

3. The Questionnaires Were Silent
Interestingly, the patients' answers on the questionnaires (how they felt about their vision) did not match the test scores.

• Why? The paper suggests that people who have lived with severe vision loss for a long time get very good at adapting. They develop strategies and habits that help them cope. So, even if their "light sensitivity" improves slightly, they might not feel a huge difference in their daily happiness because they were already managing well. It's like a person who has learned to navigate a dark house by touch; if you turn on a dim light, they might not notice the change in their ability to get around, even though the light is there.

4. The Test is Reliable
They tested the patients three times over a few weeks. The results were very consistent. If a patient saw a certain level of light on Monday, they saw a very similar level on Wednesday. This proves the test is stable and not just a fluke.

The Bottom Line

This study shows that the Full-Field Stimulus Test (FST) is a powerful tool for measuring vision in people who are almost blind.

• It can detect light where standard tests see nothing.
• It predicts whether a patient can actually perform daily tasks like finding objects or walking safely.
• It provides a precise number that doctors can use to track if a treatment is working, even if the patient's vision hasn't improved enough to read a standard eye chart yet.

In short, the FST acts like a sensitive microphone that can hear the faintest whispers of vision, proving that even in "ultra-low vision," there is often still a signal worth measuring.

Technical Summary

1. Problem Statement

• Clinical Challenge: Assessing visual function in patients with Ultra-Low Vision (ULV), specifically those with advanced Retinitis Pigmentosa (RP) who have visual acuity (VA) worse than Counting Fingers (CF), is a significant hurdle in therapeutic development.
• Limitations of Current Metrics: Conventional VA testing (e.g., ETDRS charts) is categorical and insufficient for this population. It fails to quantify residual vision in patients classified as No Light Perception (NLP), Light Perception (LP), or Hand Motion (HM), and cannot detect subtle changes following interventions like gene therapy, optogenetics, or stem cell treatments.
• Need for New Endpoints: There is a critical need for quantitative, continuous endpoints that can measure residual retinal sensitivity and correlate with real-world functional abilities to serve as primary or supportive outcomes in clinical trials.

2. Methodology

• Study Design: A single-center, prospective, observational, cross-sectional study conducted at Keio University School of Medicine (Tokyo, Japan).
• Participants: 35 patients (70 eyes) diagnosed with RP and VA in the worse-seeing eye $\le$ CF (LogMAR $\ge$ 2.0).
  • VA Distribution: NLP (14 eyes), LP (32 eyes), HM (19 eyes), CF or better (5 eyes).
• Procedures:
  • Full-Field Stimulus Test (FST): Performed monocularly using the Espion Ganzfeld Profile System after 45 minutes of dark adaptation. Stimuli were 4-ms white-light flashes. Thresholds were derived from the frequency-of-seeing curve (Weibull function). Measurements were repeated across three visits (spaced $\ge$ 1 week apart) to assess reproducibility.
  • Functional Vision Tasks (Binocular):
    • Table Tests: Adapted from retinal prosthesis and optogenetics trials. Included identifying tableware/objects on a black cloth (Table Test A) and locating/touching objects or cups with varying contrast (Table Test B).
    • Mobility Tests: Adapted from prosthesis studies. Included a "Door Task" (locating a simulated door) and a "Line Task" (following a white line on a black floor) under six ambient illuminance levels (400 lx to 1 lx).
  • Questionnaires: NEI VFQ-25, Impact of Vision Impairment (IVI), and IVI-VLV (administered by interviewers).
• Statistical Analysis:
  • Comparison of FST thresholds across VA groups (Kruskal-Wallis/Dunn's test).
  • Correlation analysis between FST thresholds and functional task scores (Spearman's rank correlation).
  • Receiver-Operating Characteristic (ROC) analysis to determine FST cutoffs that predict the transition from zero to non-zero functional performance.
  • Assessment of test-retest reliability using Bland-Altman plots and Coefficient of Repeatability (CoR).

3. Key Contributions

• Granular Quantification in NLP: Demonstrated that FST can detect measurable residual light sensitivity in 64.3% of eyes clinically classified as NLP, providing a quantitative metric where conventional VA offers none.
• Anchoring to Daily Function: Successfully linked psychophysical FST thresholds to performance-based measures of Activities of Daily Living (ADL), specifically object recognition and mobility.
• Establishment of Clinical Cutoffs: Identified specific FST threshold ranges that correspond to the ability to perform basic functional tasks, moving beyond abstract sensitivity to practical utility.
• Validation of Reliability: Confirmed high test-retest reliability and lack of systematic learning effects over three visits, supporting FST's viability for longitudinal clinical trials.

4. Key Results

• FST Thresholds by VA Group:
  • Significant differences were found across VA categories ($p < 0.001$).
  • Median thresholds (log cd·s/m²): NLP (1.13), LP (-0.27), HM (-1.13), CF or better (-2.82).
  • Heterogeneity: Substantial variability existed within groups (spanning >3 log units), suggesting FST offers finer stratification than categorical VA labels.
• Correlations with Functional Vision:
  • Strong Negative Correlations: FST thresholds showed significant correlations with Table Test performance ($r = -0.46$ to $-0.70$) and Mobility Test performance ($r = -0.65$). Lower (better) FST thresholds correlated with better task performance.
  • No Correlation with QoL: No significant association was found between FST thresholds and patient-reported outcome (PRO) questionnaire scores (NEI VFQ-25, IVI, IVI-VLV), likely due to "response shift" and adaptation in long-standing ULV.
• Reproducibility:
  • Test-retest variability showed no systematic bias.
  • Coefficient of Repeatability (CoR) was $\pm 0.66$ to $\pm 0.82$ log cd·s/m².
  • Intraclass Correlation Coefficient (ICC) was 0.962, indicating excellent reliability.
• ROC Analysis (Functional Cutoffs):
  • FST thresholds of $-1.75$ to $-0.87$ log cd·s/m² were identified as the transition points where patients began achieving non-zero scores on functional tasks.
  • Mobility: Door Task cutoff at $-1.75$ (Sensitivity 100%, Specificity 96.4%).
  • Table Tasks: Cutoffs ranged from $-0.87$ to $-1.00$.

5. Significance

• Regulatory and Clinical Utility: The study supports FST as a clinically meaningful, scalable, and standardized endpoint for Phase 3 trials in ULV (e.g., gene therapy, optogenetics). It addresses the regulatory requirement for endpoints that reflect ADL capabilities.
• Stratification Tool: FST allows for the inclusion of NLP patients in trials who were previously considered unmeasurable, expanding the evaluable population for vision restoration therapies.
• Defining Meaningful Change: The study suggests that a change of $\ge 1.0$ log cd·s/m² in FST thresholds likely represents a clinically meaningful improvement (exceeding the noise of test-retest variability), while changes of $\approx 0.5$ log units may represent the smallest detectable difference.
• Future Directions: The findings advocate for anchoring FST improvements to performance-based functional tasks in future interventional studies to validate the clinical relevance of visual sensitivity gains.