 |
Module 38 |
||
|
|
|
||
|
Module 38: |
Optical Coherence Tomography |
||
|
|
Part 2: Quantitative Analysis and scanning techniques | ||
|
|
|||
|
Freezing and selecting the scan The Fast Macular Thickness Scan The FMT scan and quantitative analysis Reproducibility issues with the FMT scan Retinal Thickness/Volume Change Analysis
|
|||
|
As discussed in the previous module, the Zeiss OCT 3 allows both qualitative and quantitative analysis of the retina.
Qualitative analysis involves describing or identifying morphological changes and anomalous structures in the retina. Morphology is the study of forms and structures of organisms. An anomaly is a deviation from what is normal (an irregularity). Qualitative analysis was discussed in the first OCT module.
Quantitative analysis involves measurements of the retina, specifically retinal thickness and volume, and nerve fiber layer thickness. This is possible because the OCT 3 software is able to identify and "trace" two key layers of the retina, the NFL and RPE.
The software can then measure the distance between these two layers, which represents retinal thickness. Multiple thickness measurements can be interpolated into a volume measurement of the retina. The software can also measure nerve fiber layer thickness.
A good scan image must be obtained in order for the software measurement analysis to work. The software must be able to differentiate the nerve fiber layer (NFL) and the the retinal pigment epithelium (RPE) layer. The differentiation is possible because these layers are more highly reflective than the other layers of the retina. A good scan optimizes the reflectivity from these layers. This module also discuss techniques for optimizing OCT scans.
For basic information concerning the operation of the Zeiss OCT 3, please refer to the manual that comes with the instrument. This discussion assumes the reader has some basic knowledge of how to operate the instrument.
The OCT is similar to a retina camera in that clear media is needed for a good OCT scan. Corneal and lens opacities can significantly degrade the image. A good quality tear layer is also very helpful in obtaining a good image. Have the patient blink frequently, but advise the patient of when a scan is being captured so that he/she can "not blink" at the time of acquisition. It is a good practice to instill an artificial tear drop in the eye before the session begins, and as needed during the course of the exam.
If the pupil is well dilated, the joystick can be manipulated in the up/down and left/right directions to help the OCT "see around" some opacities.
The video window is used to align the patient and to view the retina during the exam. It is not necessary to have a good image in the video window in order to acquire a good scan. In fact, sometimes the best images are obtained when there is a poor image in the video window. You don't want to sacrifice the quality of the scan in order to get a good video image.
A good video image can be useful because it can give you an idea of approximately where the scan was on the retina. I say approximately, because the video image was not saved at the same time that the scan image was saved, allowing for eye movement to create a disparity.
In the video image below, you can faintly make out the pattern of the radial line scan over the macula. A darker retinal video image makes it easier to make out the position of the scan.
The video image has several control settings that can be accessed from the "video and lamp parameters" tab on the scan acquisition screen.
I find that checking the "dark" eye color button works the best for many scans. The video brightness and contrast, and the lamp intensity can be changed to maximize the video image. If you are getting a poor video image, or none at all, these settings are the first you should check.
A good quality OCT scan has good reflectivity from edge to edge. The "hotter" colors (orange, red, white, yellow) are maximized. The two scans below both have good reflectivity and good focus. The top one is a high resolution scan, and the bottom scan is one of six lower resolution scans that make up a radial line scan. Notice the good focus on the second scan that shows the vitreous interface. Generally, the retina should be in the lower portion of the scan window so that the vitreous can be images as well. Manipulate the joystick so that the scan is as horizontal as possible. This is not always easy to do.
The scan below is in focus, but the right side of the scan lacks good reflectivity. This is because the edge of the scan is catching the iris. Attention to alignment within the pupil will produce a better scan.
Alignment and focus are used to maximize the quality of the scan. As discussed in the manual, alignment begins with centering and zooming in on the "football" shaped reflex in the video image. The initial lens-to-subject distance is achieved when the retinal image fills the video screen. This is similar to the image you see when doing retina photography. At this point, your attention should shift to the scan window.
If an image is not visible in the scan window, you should click on the "z-offset optimize" button on the "scan parameter" tab. Once the scan image is visible, it can be move with the z-offset arrows.
At this point, the "optimize polarize" button should be clicked.
This should automatically refine the focus on the retina, and you should see an increase in the "hot" colors in the scan, as illustrated below. The top image is before optimization, the bottom is after.
Reflectivity may be further enhanced by moving the focus knob on the side of the OCT unit.
This is particularly true if the patient is significantly far sighted or near sighted. Changing this knob may be helpful even if the patient does not have a significant refractive error. The polarization should be optimized again after this knob has been changed.
At this point, the "scan mode" button is clicked so that you have a full resolution image of the scan(s) in the scan window. From this point on, maintaining a good scan image is a matter of adjusting the unit with the joystick if necessary to compensate for movement by the patient that may degrade alignment. Encourage the patient to blink until you are ready to freeze the image.
Freezing and selecting the scan
The image can be frozen with a video image with a flash or without a flash. Try it both ways. Freezing without a flash is easier on the patient, and there may be little difference in the video image. These control buttons are next to the scan mode button, or the buttons near the joystick can be used.
When ready to freeze the image, tell the patient not blink and let the scan go through several passes before freezing the image. The software saves the last 8 scan passes for review. Always use the review button (on the bottom row of buttons). This allows you to pick the best scan from the series of scan passes, as pictured below.
When you click on the thumbnail at the bottom, the current set of scans is displayed in the windows above. The software also tells you what the signal strength was for that particular scan, on a scale from 1 to 10, with 10 being maximum signal strength. You can only save one scan from the group. Check (click on the thumbnails) each scan or scan group. Save the scan that has high signal strength, consistent quality from edge to edge, and no blink or movement artifacts.
Signal strength 5 or above is generally acceptable, but if you are getting readings below 7, be sure to scan several times to try for a better signal strength.
The OCT scan can sometimes be improved by changing the "noise" and "range" settings on the "OCT Image" tab. The default settings are indicated by blue markers on the scale.
Most instruments produce the best scans when set with the default values. Increasing the noise level will produce "hotter' colors in the scan, but the noise artifacts will also increase. Noise artifacts are those "snowflakes" that you see in the dark areas of the scan (e.g. the vitreous), as picture below.
The Fast Macular Thickness Scan
The Fast Macular Thickness Scan (FMTS or FMT scan) has become a standard OCT retina scan.
It has two important functions.
The FMTS and Quantitative Analysis (thickness and volume)
Many macular diseases involve the presence of edema (fluid) either in the intra-retinal space, or in the sub-retinal spaces. One way to follow the progress of the disease and the various treatments is to try to gauge the amount of fluid present over time. Fluorescein angiography has been used for this purpose. OCT has given us another tool. The OCT allows us to measure retinal thickness (an edematous retina will have increase thickness) by measuring the distance between the nerve fiber layer (NFL) and the retinal pigment epithelium layer (RPE). A volume measurement can be obtained by combining the thickness data from more than one OCT scan.
The FMTS protocol consists of 6 scans lines oriented radial and spaced equally around a center point. Each of the six scans is saved as a separate scan line.
Using the retinal thickness analysis tool, the software then traces a line along the NFL layer and a line along the RPE layer. The software can recognize these layers because they are highly reflective (hot colors). Thus, the importance of maximizing the hot colors.
The software then measures the distance between the two lines and a graph is produced which compares the measured thickness to the thickness of a normal retina.
Each of the six scans can be reviewed by clicking on the slider bar to the left, and any or all of them can be printed out for the patient's record. When reviewing each scan, it is important to examine the position of the white line on the NFL and the white line on the RPE. If the scan has a weak signal, the software may make a mistake in following the layers, as illustrated in the screen shot below.
If a better scan can obtained, the defective scan should be discarded. It is a good idea to obtain two or three of these scans at the same time, so that the chances are better of getting an acceptable quality scan.
Be aware that the OCT 3 retinal thickness analysis does not measure retinal elevation. Take for example this eye with a pigment epithelial detachment (PED) pictured below. The arrow on the left would represent retinal elevation, from the choroid, through the fluid space of the PED, to the nerve fiber level. The arrow on the right shows what the analysis measures, defined by the distance from the RPE (which is detached) to the NFL.
Below is pictured the OCT 3 retinal thickness analysis for this eye. The align process that is performed on the scan image for the retinal thickness analysis distorts the image.
Retinal volume analysis can be performed on the same FMT scan that you performed retinal thickness analysis on. The volume analysis software takes the data from the six scans and combines the thickness measurements into a volume measurement for the circle that encloses the scan lines.
Upon examining the diagram above, you will notice that there is much area within the circle that is not actually measured for thickness. The software "interpolates" the data within these areas. This means that it makes a guess based upon the data from the two adjacent scan lines. The more space there is between the scan lines, the less likely that the interpolation will be accurate. Therefore the data from the inner portion of the circle (green) will be more accurate than the data from the outer portion of the circle (yellow). By the same reasoning, the data from the center point will be the most accurate of all, since each of the scan lines passes through the center point (more on this later).
The retinal volume analysis software records the center thickness measurement on the printout. Interpolated thickness measurements are given for the pieces of the pie in the outer areas and inner areas of the circle. An overall volume measurement is given, based on the thickness measurements.
Before performing retinal volume analysis on FMT scan, it is important to use the retinal thickness analysis on the same FMT scan so that each of the six scans can be examined to be sure that the software has correctly identified the NFL and RPE layers on each scan. It is best not to use an FMT scan with layer identification errors as this will introduce significant error into the volume analysis.
Below is pictured a retinal map analysis on a scan with a layer identification error (white arrow). Notice that the software correctly traced the RPE layer on the right portion of the scan, but the trace mistakenly jumped up to the NFL on the far left portion of the scan. This created a measurement error that is represented by the wedge shaped blue area on the thickness map (black arrow).
Reproducibility issues with the FMTS
Since the scans must be acquired sequentially instead of simultaneously, lower resolution "fast" scans (128 scans per line vs. 512 scans per line) are used in order to minimize the acquisition time, which minimizes the inherent errors caused by eye movement during scan acquisition. In this case, "fast" means about 2 seconds, which is still plenty of time for patient movement.
For a given FMT scan, the center thickness measurement will be the same for each of the six scans, if the eye does not move during the scan. The retinal volume analysis software gives us a standard deviation of the center thickness measurement for the six scans. This is the +/- number that follows the center thickness number, which is an average of the center point thickness of the six scans. The standard deviation is a statistical measure that tells you how closely the individual numbers are grouped around the average. In this case, smaller is better. A small standard deviation number indicates little patient movement during the scan, which means the scans in the group will give a more accurate volume measurement than a scan group with a larger standard deviation.
Generally a scan group with a standard deviation number within 10% of the thickness number is considered to be acceptably accurate. See the example below. The arrow points to the center thickness average and the standard deviation on a Retinal Map Analysis printout. The section is enlarged in the image directly below. The center thickness is given as "508 +/- 49". Since the standard deviation number "49" is within 10% of the average thickness of "508", this scan might be considered to be acceptably accurate.
It is always best to capture more than one FMT scan. For volume measurements, it is a good practice to save at least 2 FMT scans. If all other factors are equal, you would consider the most accurate scan to be the one with the lowest center thickness deviation.
One of the most useful functions of the OCT 3 is the ability to take volume measurement over time. For example, a FMT scan before treatment for AMD, and FMT scans at various intervals after treatment. Successful treatment should be followed by a decrease in retinal thickness and volume.
Along with patient movement during the scan, another reproducibility issue with the OCT 3 is the ability (or inability) to scan the exactly the same area on the retina in serial scans over time. In other words, how do you know that you are scanning the same spot that you scanned last month? The short answer is that you don't know. The technical terminology for this kind of error is "lack of registration".
The long answer is that for the measurements to be useful, you don't have to scan exactly the same spot. Measurements of 508 center thickness and 450 center thickness one month later may be within the margin of error. You can't really say that the retina is getting thinner, because you may be measuring different spots, or the patient may be moving. But measurements of 508 initially, and 200 one month later almost certainly do mean the retina is getting thinner, assuming that you are measuring close to the same area.
There are ways to increase the likelihood that you are scanning the same area on repeat scans. The most useful and obvious technique is to rely on the patient's central fixation, if the patient has good fixation. The patient simply looks at the central fixation target. The fixation target can be moved around within the scan area, and the scan can be moved with the fixation target, or independently from the fixation target. The dropdown menu on the scan parameter tab allows you to choose the different options. The target and the scan are moved within the video window with the mouse by clicking and dragging.
Probably the most useful scans are the scans that are centered on the fovea. This is the default position of each scan when it is initiated. For example, the default position for the center of the FMT scan is over the fixation target. In this situation, the likelihood of a repeatable scan is very good, assuming the eye has good fixation.
A center line feature can be activated by simply clicking inside the scan window. This places a white line through the center of the frame, as show below. This line aids in centering the image on the fovea with each scan pass..
There may be occasions when you are scanning a retinal lesion that does not involve the fovea. In this case the fixation target and the center of the scan will be in different positions within the video window. How would you be able to increase the likelihood of a repeatable scan in this case? The OCT 3 software provides a "repeat" scan feature that will place the fixation target and the scan in same place in the video window as they were in a selected previous scan.
What if the patient has poor fixation? If the patient has good control of eye movements, she can simply be asked to look into the center the scan box.
The OCT 3 software provides another aid for repeatability: the landmark. Selecting from dropdown menu on the "scan parameter" tab, the scan can be moved to the area of interest and the landmark can be positioned over some easily identifiable landmark such as a vessel crossing. The repeat feature can be used on subsequent scans, and the landmark and the scan can be moved as a unit (this is a dropdown menu choice) until the landmark is over the previously selected vessel crossing.
Retinal Thickness/Volume Change Analysis
Two FMT scans on the same eye, but taken on different dates, can be selected at the same time while holding down the "ctrl" key. "Retinal Thickness/Vol Change is chosen from the analysis tab.
The analysis will give you a "change map" as pictured below, showing the difference between the two scans.
When evaluating the glaucoma suspect or the glaucoma patient, two parameters that the ophthalmologist is interested in are the characteristics of the optic nerve cup and the thickness of the nerve fiber layer surrounding the optic nerve head.
The optic cup profile can be evaluated by capturing a "Fast Optic Disc" scan. The patient fixes on the target, which is automatically placed at the edge of the scan window so that the optic nerve is viewed toward the center of the video window. The operator then moves the scan so that the star pattern is centered on the optic nerve head. Centering can be aided by clicking on the scan window to view the white centering lines.
The optic nerve scan can be analyzed with the "optic nerve head analysis" protocol, as pictured below.
Nerve fiber layer thickness can be evaluated with the "Fast RNFL Thickness" scan. This is a circular scan that requires the operator to place the circle so that the center of the circle is centered on the optic nerve head.
The analysis software places lines on the top and bottom of the nerve fiber layer and the distance between the two lines is interpreted to be the thickness of the nerve fiber layer (white arrows). As with the retinal thickness and volume measurements, it is important to maximize the reflectivity of the scan, using the same scan techniques as discussed earlier.
The scan circle in this next image is not well centered.
Care must be take to make sure that the image is captured with the circle centered on the optic nerve. Patient movement can make this a challenge. The placement of the circle can make a big difference in the analysis of the nerve fiber layer thickness, as pictured below. These two scans (OD) are of a normal eye. The scan in the first analysis is well centered and the RNFL thickness falls within the normal range. The scan in the second analysis is of the same eye (OD), but the scan is not well centered. The analysis is abnormal (black arrows).
Glaucoma scanning with the OCT 3 suffers from the same problems that macular quantitative analysis scans suffer from, lack of registration and questionable repeatability. Until improvements in the hardware and software improve or eliminate these problems, operator skill will play a major roll in the quality of OCT scanning.
References:
Brancato R. and Lumbroso B. Guide to Optical Coherence Tomography Interpretation. Rome: Innovation-News-Communication, 2004.
Schuman J., Puliafito C., and Fujimoto J. Ocular Coherence Tomography of Ocular Diseases. Thorofare NJ: Slack Inc., 2004.
|
|||
| Back to top Go to Post-Test | |||
