Module 48

 

Home

Catalog

Contact us

 

 

Go To Post-test

 

     
 

Module 48:

Spectral Domain Optical Coherence Tomography

 

Section 3: 

Scanning Techniques
 

 

 
 

Contents: 

Software Protocol

     The topography scan

     The raster scan

     The line scan

Patient management

 

 

 
  Scanning techniques for optical coherence tomography can generally be divided into two different aspects, no matter what instrument you are using: 1. software protocol and 2. patient management.

Software Protocol

The software protocol will be different, depending upon the maker of the instrument, but some general guidelines can be helpful:

1. Once you become familiar with what the instrument can produce in terms results and reports, you will want to standardize the scanning protocol so that every operator will know what scans need to be captured, saved, and reported.  This can vary for each doctor who makes use of the instrument, or there can be one general protocol for macula scans and one general protocol for glaucoma scans. 

There are three advantages to standardization.  The first is that the operator knows what is expected.  The second is that the patient will get a sense of professionalism instead of incompetence or experimentation.  The third is that all the information needed for future evaluation and comparison to subsequent scans will have been saved.

A good example is the protocol that we use with the Opko SLO/OCT for macula scans.   We routinely capture three different scan types, because they yield different information.  Each scan type takes less than 30 seconds to perform with a cooperative patient.

   
  1. 3D retinal topography scan:  This protocol covers 30 degrees of the posterior pole with 200 line scans to produce the retinal thickness map and the 3D renderings.  You can choose a "high quality scan" which takes about 2 seconds of scan time, or you can choose a "high speed scan" which takes about 1 second of scan time.  Remember that there is a tradeoff between speed and quality.  The faster the scan rate, the fewer A-scans per unit length or area, and the lower the quality (resolution) of the image.  Also, remember that the longer the scan takes, the more motion artifacts (distortion) in the scan.   We use the "quality" setting unless the patient has a hard time fixing on the fixation cross, then the "speed" setting is used.

A report from the 3D topography scan is pictured below.  The 3D imaging can be viewed in Section 2.

   
 

   
  As technicians, I believe it is our responsibility to understand how the instrument and the software work to produce the results that it spits out.  Some doctors are very proactive and knowledgeable about the intricacies of technology, but others prefer to just be able to interpret results without knowing about the nuts and bolts.  We as technicians need to provide assistance to the physician concerning the peculiarities of the instruments we use.  OCT retinal topography is a good example.

Generally, retinal topography maps are based upon retinal thickness measurements.  Retinal thickness is defined as the vertical distance between the top of the retinal nerve fiber layer (RNFL) to the top of the retinal pigment epithelial layer (RPE).  In the image below, the top arrow points to a red line that is tracing the RNFL.  The bottom arrow is pointing to a green line that is tracing the RPE layer.

   
 
   
  The topographical map that is produced from 200 scans like the one above is pictured below, with the areas of greatest thickness (~500 microns) being in red.
   
 

   
  Now consider the scan imaged below.  Again, the top arrow points to the red line tracing the RNFL, and the bottom arrow points to the green line tracing the RPE layer.  Those of you who are experienced scanners will recognize this as an RPE detachment.  The fluid under the RPE is pushing the entire retina upward into a large bump.  You might think that this would produce a large area of elevation on a retinal topography map, but not so, because the actual retinal thickness is close to normal.
   
 
   
  The corresponding topographical map is pictured below.  Notice that the map appears close to normal (lots of blue and green), even though there is a large area of elevation on the individual scans.  The software has the capability to adjust for this by manually redrawing the RPE scan line, but it would be nice if this could be done automatically as an option, perhaps in a future software update.  I don't think you would want the software to automatically redraw the line under a PED,  because standard retinal thickness is important to know, even above a PED.
   
 

   
  2. Raster scan:  This protocol is performed for two reasons.  The first is that fewer scans are performed (36 or 72) compared to the topography scan (200), so that each individual scan is a higher resolution scan compared to the topography scan.  The second reason is the entire posterior pole is covered in a mechanical, evenly spaced fashion, so that no pathology is missed in the posterior pole.  We find this scan to be the most useful in the clinical setting because the physician can very quickly view OCT scans of the entire posterior pole.

An image from an individual raster scan is pictured below.  A slide show of an entire raster scan can be viewed in Section 2.

   
 
   
  3. Line scans:  These are individual high resolution scans that are "manually" captured.  The operator has to move the scan line on the fundus image.  We save line scans that are representative of the pathology.  The instrument saves a "loop" of up to 128 scans.  After 128 scans are saved, the older scans are written over in a continuous loop.  All scans can then be viewed, and scans that are not wanted can be deleted.  From 4 to 16 scans are saved per second (frame rate), so motion artifacts are minimal.  The scans are viewed in a movie mode, with forward, back, and pause controls.

The image of an individual line scan is pictured below.

   
 
   
   
  Patient Management

The software procedures with the Opko SD SLO/OCT are easily learned and performed.   As with most ocular imaging procedures, the key to good results is good patient management.  Although the scanning speed is much improved with spectral domain technology, the best results are still obtained when the patient's eye remains motionless.  Here are some helpfull hints:

  1. It is best to work with a dilated eye.  Because of the scanning laser technology, the Opko instrument can capture beautiful images through an undilated pupil, but dilating the pupil gives you more lateral operating room.  You don't have to continually chase after the pupil if the patient's eye is moving.
  2. Having the patient's head firmly planted in the headrest is critical.  The Z axis window is very narrow.  The Z axis is the forward and back distance between the retina and the instrument lens.  Remember that the structure that you are imaging is less than 2mm thick.  Be sure that the headrest does not move, if so, figure out how to tighten it up (loose screws?).  Be sure that the patient understands the importance of keeping the chin in the chinrest and the forehead against the bar.
  3. Good fixation by the patient is very helpful, although with the high speed technology and eye tracking (on RNFL scans), acceptable images can be captured even with an eye that has wandering fixation.  The Opko instrument has a large internal fixation cross, as well as an external fixation light for the fellow eye.
  4. Scan acquisition times for raster protocols are from 1 to 4 seconds, so warn the patient to not move or blink just before the scan starts.  Have the patient blink before each scan.  A good tear film is important for good image resolution, so keep some artificial tears handy.
  5. If there are lens opacities present, the scanner can be moved around in the X and Y axis inside the pupil to improve the scan image.  Good dilation helps.
  6. Just as with the OCT 3, a good OCT scan has no motion artifacts, no blink artifacts, and has a high signal to noise ratio.  A high signal to noise ratio is evident by good reflection from the structure.  If you are viewing the OCT scan in grayscale, the best scans are the bright, dense scans.  If you are viewing the OCT is false color, then the bests scans have a higher percentage of hot colors (red, orange, yellow).  The display has a signal indicator, but it is distracting to have to look back and forth.  The same information can be had by just observing the quality of the scan in the window.  The scanner runs continuously until you initiate a capture, so adjustments can be made in alignment to optimize the scan. Below are screen shots from the scan acquisition window, in grayscale and in false color.
 
   
   
  Below is a screenshot of the Opko SD-OCT scan acquisition screen to give you an idea of the software procedure.

1.  The patient is aligned with window #1.  The three "snowball" dots on the cornea are focused on the cornea by moving the instrument along the Z axis.  The small square in the center is aligned inside the pupil by moving the instrument on the X and Y axis with the joystick.

2. The graphic at #2 is the position of the internal fixation cross, which is centered by default, but can be moved anywhere, or aligned for optic nerve scans.

3.  Once the eye is aligned, the SLO image of the fundus appears in window #3.  The operator then clicks on the "autofocus" button at the top left.  The instrument then focuses on the retina (no diopter adjustment is needed), and then brings the scan image into window #4.

4. With the scan image in window #4, the operator uses software adjustment to tweak the alignment in the window.  The joystick is used to adjust the scan in the pupil to optimize the scan signal.  When the scan looks optimized in the window, the scan is acquired with a mouse click.

With a cooperative patient, the complete alignment process is very fast, taking about 30 seconds.

   
 
   
  One of the problems with so many producers of new spectral domain instruments is that there will be no standard software package.  For awhile at least, it may be difficult to compare quantitative measurements obtained from different instruments.  However, I believe that this problem will sort itself out.  Advantages of certain software functionalities will become apparent over time, and the various software packages will begin to look more similar than different.
   
  Back to top                   Go to Post-test