Part 4 of 5

Early Results

Early results confirm the precision
of the Manifold®.

Our early studies confirm that the Manifold® delivers an important assessment of everyday vision.

In a published study of multiple sclerosis, Manifold®’s contrast sensitivity metrics predicted self-reported disability better than current high- and low-contrast acuity metrics.

In a study of operational performance at Wright-Patterson Air Force Base, with results presented to American Society of Aerospace Medicine, Manifold®’s contrast sensitivity metrics predicted performance in a simulation of an in-flight-refueling task better than other clinical vision measures. In addition, contrast sensitivity predicted visual discomfort during the extended stereoscopic viewing needed to complete the task.

In a study of glaucoma at the Wilmer Eye Institute, with results presented at ARVO, Manifold®’s contrast sensitivity metrics exhibited better repeatability and higher precision than a CS letter chart.

In a study of retinitis pigmentosa and potential electrostimulation therapies, with results presented at Envision and Academy of Optometry, Manifold®’s contrast sensitivity metrics were sensitive for detecting visual changes in treatment conditions that increased ocular blood flow. The metrics were more sensitive for detecting change than other endpoints. A novel analysis using signal detection theory and receiver operating characteristics was applied to evaluate the probability of change relative to baseline.

Tests showed that the Manifold® provides additional information about functional vision loss beyond that provided by previous letter CS charts.

Taken together, these results demonstrate that the Manifold® provides a measure of visual performance that correlates with subjective visual experience, but can be measured with the precision typical of objective measures.

To distinguish between the properties of test repeatability and precision, we have developed novel quantitative metrics. For example, a vision test with coarse quantization can yield repeatable results, but because the steps are so large, the results are not useful for tracking visual changes in individuals over time. Whereas repeatability is desired, it is not sufficient to track small changes over time if it is achieved through coarse quantization. We have developed metrics that distinguish the qualities of precision and repeatability.

Figure 1

The AST functional vision test can give us a better picture of what’s changing in visual health over time.

How does the Manifold® help with evaluating therapeutic candidates?

Improving the sensitivity and precision of endpoints can improve the signal-to-noise ratio at the heart of clinical trials. It gives us a better picture of what’s changing in visual health over time; it can also help us better quantify the state of vision changes before and after therapeutic interventions. We are using receiver operating characteristic (ROC) analysis—typically applied to evaluate diagnostics at a population level—to detect changes in contrast sensitivity in an individual before and after a therapeutic intervention.

  1. SENSITIVITY SPATIAL FREQUENCY 1.5 3 6 12 18 2.0 1.5 0 1.0 0.5 KEYTIME 1TIME 2TIME 3TIME 4

    Measuring the full contrast sensitivity function is valuable when tracking vision loss over time. Here, acuity testing only reveals vision loss over time (approx. 3 lines) at the highest contrast level tested.

  2. SENSITIVITY SPATIAL FREQUENCY 1.5 3 6 12 18 2.0 1.5 0 1.0 0.5 KEYTIME 1TIME 2TIME 3TIME 4

    Likewise, testing peak contrast sensitivity reveals vision loss that is isolated to low spatial frequencies: a 50% reduction (.30 decimal log units) over time.

  3. SENSITIVITY SPATIAL FREQUENCY 1.5 3 6 12 18 2.0 1.5 0 1.0 0.5 KEYTIME 1TIME 2TIME 3TIME 4

    A broader, comprehensive view of visual changes is provided by the full contrast sensitivity function. Profound loss—1 decimal log unit or 90% reduction—is evident in the mid-frequency region that is currently hidden from vision testing. This region, demonstrating an increased dynamic range of vision loss, provides a hidden substrate for treatment, with potential value for clinical trials.

  4. 0.2 0.3 0 0.1 EXPLAINED VARIANCE (R 2 ) qCSF SLOAN SNELLEN VISUALQUALITY OF LIFE

    CSF metrics provided by Manifold® correlate better with self-reported quality-of-life, measured by the composite score of the NEI-VFQ questionnaire, than high- or low-contrast acuity measures.

  5. LOG CONTRAST SENSITIVITY SPATIAL FREQUENCY PRE TREATMENTPOST TREATMENT 0.5 1 2 4 8 16 0.3 0.7 1.0 1.3 1.7 2.0 0.5 1 2 4 8 16 0.3 0.7 1.0 1.3 1.7 2.0 0.5 1 2 4 8 16 0.3 0.7 1.0 1.3 1.7 2.0 A 15 YEARS B 15 YEARS C 11 YEARS

    Precision is important for characterising individual differences in responses to treatment. For example, when evaluating recovery from blindness following extraction of congenital cataract in adolescence, improvement in contrast sensitivity can be (A) large and broad, (B) restricted to mid-to-high spatial frequencies, or (C) non-existent. Here treatment response is not obviously predicted by age or baseline vision at time of treatment. In eye care, the confident distinction between responders and non-responders is impossible without precision.

  6. AULCSF SENSITIVITY (%) 1-SPECIFICITY (%) 100 0 50 100 0 50 PRETREATMENT POSTTREATMENT 76% aROC = 76%

    This novel ROC analysis detects visual changes in individuals. Rather than represent vision metrics collected over two populations, these distributions represent metrics collected and calculated over two vision tests of the same individual: pre- and post- treatment. This novel application provides a diagnostic metric for visual change: the probability that contrast sensitivity was better in the post-condition, relative to the pre-condition.

In the end, it is important to measure not only one CSF. It is important to measure several CSFs and analyze their differences: for example, left vs right eye or pre- vs post-treatment. Ongoing studies are providing the data to develop the analytical tools to track subtle changes in visual function due to disease progression or treatment effects.

You can follow our research through the abstracts, drafts, and published papers about the Manifold® Platform and qCSF algorithm. If you want to collaborate, or learn more, please contact us.

Next: Part 5 of 5

The Manifold® Platform is here.

Continue reading ›