I believe most of us would agree the real golden nugget with imaging is the ability to assess glaucoma progression, perhaps even more so than screening and ruling out disease. Until recently, however, we did not have a statistical package powerful enough to use for glaucoma progression analysis. The Cirrus™ HD-OCT GPA (Guided Progression Analysis) has given us this ability, primarily because the vast number of data points collected in the data cube over a short period can be used to analyze progression over time with serial scanning.
With the Cirrus HD-OCT, we are essentially measuring a 6mm x 6mm, 200 A-scans per B-scan by 200 B-scan area around the ONH. The advantages of post-processing autocentration and automatic disc margin alignment enable us to gather more than 40 million data points in 1.5 seconds.
The Cirrus HD-OCT employs a line-scanning ophthalmoscope, which provides excellent fundus details and ensures adequate registration and alignment, particularly vessel alignment, so that each scan measures the same point over time.
The key to detecting glaucoma progression lies in our ability to analyze structural changes. Unlike screening for disease or ruling out disease, where we are comparing a patient's data to a normative database, in progression analysis, the key to robust analysis is tight registration and reproducibility with low variability, low standard deviation and strong statistical analysis. The Cirrus HD-OCT is the first device with this advanced software.
Using the Cirrus HD-OCT, we can identify change in RNFL thickness through event analysis and trend analysis. Event analysis assesses changes that are beyond an expected variability at certain points compared to normative data. If a patient falls outside this area, it triggers an event that is identified as progression. Trend analysis looks at the rate of change over time, using linear regression to determine whether or not the trend is outside the expected rate of RNFL loss.
For GPA analysis, two baseline examinations are required. Thanks to newer software, we can now take two images on the same day, which I find particularly advantageous. Each subsequent scan is automatically registered to the baseline for reliable comparison of RNFL thickness over time.
The Cirrus HD-OCT uses three methods to detect RNFL defects (Figure 5). The Image Progression Analysis is helpful for detecting deep, narrow focal defects. The TSNIT Progression Graph can detect shallower, broader focal defects, and the Summary Parameter Charts detect diffuse defects.
Figure 5. The Cirrus HD-OCT uses Image Progression, TSNIT Profile Progression and Average RNFL Loss to reveal RNFL defects.
On the Image Progression Map (Figure 6), the third scan is compared to the two baselines, with yellow pixels denoting change from baseline. When a fourth image is obtained, the third and fourth examinations are compared to baseline, and change is confirmed in red. The ability to compare current and subsequent scans to baseline allows us to determine the presence and significance of change.
Figure 6. The Image Progression Map allows us to determine the presence and significance of change.
The TSNIT thickness profile progression analysis also uses color to indicate change. In Figure 7, for example, you see areas of change in yellow in the inferior temporal area. This is confirmed in the fourth scan and is shown in red. This substantiates the likelihood that this is real change.
Figure 7. The TSNIT Progression Map shows values from baselines and current examination and uses color to indicate change.
Clinicians who use the GPA feature of the Humphrey Field Analyzer (Carl Zeiss Meditec, Inc., Dublin, Calif.) are familiar with the Visual Field Index. The linear regression analyses (Figure 8) generated by the Cirrus HD-OCT are somewhat similar. The rate and significance of change is shown in text, while the RNFL thickness values for overall average, superior average and inferior average are plotted for each examination. Again, yellow markers denote change from baseline, and red markers denote change from 3 of 4 comparisons.
Figure 8. The Summary Parameter Trend Analysis shows rate and significance of change in text, while yellow and red markers denote change. Confidence intervals are shown as a gray band.
The Cirrus HD-OCT then produces a summary report (Figure 9) based on the three analyses.
Figure 9. Legend summarizes GPA analyses and indicates with a check mark if there is possible or likely loss of RNFL.
Figure 10 shows an example of a Cirrus HD-OCT report and a corresponding visual field. The slope is significant with reduced visual field index, and the GPA progression printout indicates likely progression. In this case, a nasal step is emerging in the inferior quadrant in the patient's right eye. Looking at the Cirrus printout, we see some nice structure-to-function correlation. Some RNFL loss is indicated in yellow with the fourth scan in red. This signifies a high likelihood of progression in this area, which correlates to the inferonasal step. The report also shows the values highlighted in the overall and superior thickness, but not on the TSNIT. Two checkmarks indicate likely loss based on two of the three analyses.
Figure 10. Example of a Cirrus HD-OCT report and corresponding visual field, showing structure-to-function correlation.
In this case (Figure 11), the patient's left eye is fairly stable. Nothing is highlighted on the thickness map, the TSNIT plot or the average values, and the visual field is fairly flat and stable. The right eye, however, shows progression on the visual field, with an emerging superior arcuate defect. If we think about the structural areas that would likely be affected, we would suspect something in the inferotemporal area, and we do, indeed, see it on the thickness progression map.
Figure 11. Comparing structure and function of both eyes of a patient.
To summarize, there is no question spectral domain Cirrus HD-OCT has improved resolution and speed. The hardware ensures reproducibility and registration, giving us an accurate and reliable tool for assessing RNFL change. The software has strengthened the values statistically to provide powerful analyses of structural changes. When combined with visual field changes, the Cirrus HD-OCT gives us a complete picture of each patient's status, including whether or not the disease is progressing.
