Home Print this page Email this page Users Online: 233
Home About us Editorial board Search Ahead of print Current issue Archives Submit article Instructions Subscribe Contacts Login 


 
 Table of Contents  
COMMISSIONED ARTICLE
Year : 2017  |  Volume : 5  |  Issue : 1  |  Page : 51-63

Optical coherence tomography in glaucoma-I


1 Samartha Clinic, Mumbai, Maharashtra, India
2 P. D. Hinduja National Hospital and MRC, Mumbai, Maharashtra, India

Date of Submission22-Nov-2016
Date of Acceptance22-Nov-2016
Date of Web Publication6-Dec-2016

Correspondence Address:
Barun K Nayak
P. D. Hinduja National Hospital and MRC, Veer Savarkar Marg, Mumbai - 400 016, Maharashtra
India
Login to access the Email id

Source of Support: None, Conflict of Interest: None


DOI: 10.4103/2320-3897.195312

Rights and Permissions
  Abstract 

It is a well-established fact that more than 40% ganglion cells are lost before the white-on-white perimetry shows defects. Hence, the importance of modalities which can image the structures of optic nerve and surrounding retinal nerve fiber layer in the early diagnosis of glaucoma cannot be overemphasized. However, the appropriate interpretation of these imaging modalities is of paramount importance. Out of the three machines such as HRT, scanning laser polarimeter (GDx VCC), and the OCT, the OCT has made a remarkable progress and can be incorporated judiciously in the clinical practice of glaucoma management. The purpose of this article is to explain the interpretation and its clinical application in an appropriate manner in the management of patients with glaucoma.

Keywords: Glaucoma, optical coherence tomographer, retinal nerve fiber layer


How to cite this article:
Dharwadkar S, Nayak BK. Optical coherence tomography in glaucoma-I. J Clin Ophthalmol Res 2017;5:51-63

How to cite this URL:
Dharwadkar S, Nayak BK. Optical coherence tomography in glaucoma-I. J Clin Ophthalmol Res [serial online] 2017 [cited 2020 Jul 5];5:51-63. Available from: http://www.jcor.in/text.asp?2017/5/1/51/195312

The foundation of the current optical coherence tomographer (OCT) was laid in early 1900 by Albert Abraham Michelson and his invention of the interferometer for which he was awarded the Nobel Prize in 1907. This was later pressed into in vivo imaging by the pioneering work of David Huang et al. in the 1990s in the form of time domain (TD) OCT [Figure 1]a. This instrument had the interferometer as its "heart" and used the time delay of reflected light from tissue to determine its spatial location, with the use of a moving reference mirror [Figure 1]b.
Figure 1: (a and b) Time-domain optical coherence tomographer

Click here to view


Another landmark was achieved with the use of light wavelengths and their interference pattern rather than the time delay to determine the spatial location and this was the basis of the "Fourier transformation" based spectral domain (SD) OCT [Figure 2]a. Like the Michelson interferometer in the TD-OCT, the spectrometer [Figure 2]b was the heart of the SD-OCT.
Figure 2: (a and b) Spectral-domain optical coherence tomographer

Click here to view


After the initial launch of the OCT in the early 2000, a lot has changed in the field of ophthalmic imaging related to glaucoma. Among the other devices introduced for imaging, namely the Heidelberg retina tomograph (HRT) and the scanning laser polarimeter (GDx VCC), the OCT has stood the test of time and has undergone a quantum leap from its initial beginnings. This being said that the HRT was the first ophthalmic device commercially available for the quantitative assessment of optic disc from the historical standpoint.

OCT, by consensus, has now become the tool of choice for the diagnosis and structural follow-up of glaucoma on imaging. The HRT uses an arbitrary reference plane in relation to which the thickness of the optic nerve rim and related data are calculated whereas the GDx generates data by measuring change in polarization of transmitted light and correlates it to the neural element (nerve fiber layer [NFL]) thickness.

As against all this, OCT measures the thickness of tissue - an actual structural measure. OCT data thus are cross-sectional and are not interpolated. OCT can also measure multiple structures, including the NFL, optic nerve head (ONH), and macula (in the effect of the entire length of the ganglion cell) and in doing so allows their correlation which helps in better decision-making. The HRT can measure only the ONH parameters and the GDx can measure only the NFL.

From the modest TD machine that held sway for the initial 4 years of its existence, the SD technology has virtually replaced it as the imaging tool of choice in glaucoma due to its speed and resolution. [1],[2]

Imaging occupies a unique place in the treatment of glaucoma as it is the only objective way of quantitatively assessing the damage to the ONH and the NFL over time and hence estimating disease progression. OCT findings also help reassure the patient by showing him/her concrete evidence on paper of his/her current condition or the extent of the loss (as easy numerical values) which can make him/her more inclined to change or accept more aggressive therapy as the need maybe.

Improvements in the resolution, acquisition speed, and hence the reproducibility over the generations of the machine have led to the OCT becoming one of the reliable methods to assess retinal structures [Figure 3]a. Having said that, there are certain prerequisites that make this reliability more robust and not all machines qualify to be used for glaucoma management with the same level of confidence.
Figure 3: (a) Evolution of optical coherence tomographer imaging. (b) Division of the macular region into its layers with high-resolution optical coherence tomographer

Click here to view


The objective nature of imaging as against the psycho-physical nature of visual field testing in addition to the redundancy of the retinal nerve fibers before a field defect is evident, helps it score over the standard white-on-white perimetry tests in the early stages of disease.

Cutting edge technology and the use of modern-day segmentation (the division of the image into various parts/subcomponents by computer algorithm that creates tissue boundaries based on contrast) protocols in OCT have led to subcomponent analysis in the diagnosis of glaucoma (e.g., ganglion cell complex [GCC], ganglion cell-inner plexiform layer [GC-IPL], and total macular thickness) in addition to the conventional use of imaging of the ONH and NFL [Figure 3]b. This has also been accompanied by the increased use of OCT for the bleb and angle imaging with the anterior segment module in many machines, expanding their utility in glaucoma management.

The refinement of technology in the past few years in the imaging devices has seen a significant improvement in the hardware and the software of the OCT, whereas in the other two devices (GDx VCC and the HRT), it has essentially been a software upgrade with increase in population-specific normative database and addition of newer statistical analytical tools.

The OCT has transitioned from the TD to the SD and now the swept source (SS) in the past 14 years with a phenomenal change in the imaging speeds and resolution. It has offset to a great extent the effect of saccades and motion artifacts that have been the Achilles heel of imaging. This has also been accompanied by some software modifications and increase in the population-specific normative databases similar to the other devices. All the above led to a resultant improvement in repeatability, reproducibility, and hence reliability for glaucoma diagnostics and monitoring of progression. [3]

As we are dealing with small measurements (e.g., say around 100-160 μ in NFL), the ability to resolve and measure even a minute change at the same location time and again reliably, is very critical. This is particularly important as the ultimate aim of all these tools would be to establish progression with reasonable certainty. Unlike in retinal scans where a particular value of tissue thickness would not be important numerically to the micron level per se, a similar change would be used to predict progression and estimate the effect of treatment in glaucoma patients.

In this article, we have tried to highlight the basics of the OCT, correct use of the OCT (related to the posterior segment) in the management of glaucoma, the advantages, the drawbacks of this technology, how best to use it, and the interpretation of an OCT printout in relation to glaucoma. The article would be restricted to posterior segment images, and the anterior segment tools would be dealt with in a subsequent issue.


  Clinically Relevant Basic and Technical Aspects Top


The evolution of the OCT started with the TD-OCT, which was the first device to get the optical section of the optic disc and the parapapillary NFL.

The core physics of a TD-OCT involved a light-emitting diode, moving mirror and a sample, and a reference arm [Figure 1]b. The time delay and the strength of the optical echo in relation to a reference arm were used to obtain the tissue sections after the postacquisition processing. The scanning and the processing time was long, and the axial and lateral resolutions were low resulting in images that had blurred margins and the machine found difficulty in adjusting for microsaccades during examination.

The SD-OCT that was introduced in 2006 replaced the method of signal transduction (time delay-based imaging) by the interference-based method (spectrometer) that resulted in a drastic increase in the speed of scans. Refined optics resulted in higher and better resolution and hence greater reproducibility.

Some of the manufacturers introduced the gaze tracking, reproducible re-scan placement, and alignment features in addition to improved resolution and speed that took the accuracy of the OCT to levels that could reasonably be used to look for structural progression. This was also accompanied by improvement in the segmentation algorithms that detected the interfaces between two layers of tissue that were used to place the measurement lines in an automated fashion.

The improvement in the resolution and the segmentation protocols allowed for the use of retinal subcomponent analysis in glaucoma diagnostics. Currently, the most followable parameter on the OCT is the parapaillary NFL as per most of the studies carried out till date. [4]

It is, however, to be remembered that the segmentation protocols differ with the manufacturers of the different instruments and are patented and hence not interchangeable. The Cirrus uses the GC-IPL complex for the measurement, the Optovue uses the GCC that includes the GC-IPL and the NFL at the macula, and the Spectralis uses the total macular thickness for their macular analysis. Latest software with the Spectralis can segment every layer at the macula as shown in [Figure 3]b.

The latest addition to the large portfolio of OCT is the SS OCT that uses a source of light of a longer wavelength that allows to imaging deeper than the conventional SD-OCT and gives the ability to visualize the lamina cribrosa and choroidal vasculature, the so-called extended depth imaging. This has been accompanied by a 10-fold increase in the scan speed over the SD-OCT. The axial resolution, however, is nearly the same as that of the good SD-OCT machines (i.e., 4-5 μ).


  Interpretation of the Scans Top


The scans obtained with different instruments differ in the data and their presentation depends on the hardware and the software of the machine. It is advisable to get familiar with the machine being used at the outset, with particular attention to the individual nuances specific to that machine. It is also highly recommended to have an informed and able technician to discuss these aspects clearly during the installation process including the preferred scanning protocols for that machine.

For the sake of illustration in this article, the printouts of the Cirrus and the Spectralis machines will be described for the optic disc, NFL, and the macular scans. Each machine has various permutations for the printouts, out of which the most commonly used printouts will be detailed here.

The Cirrus has software to present the optic disc and the NFL parameters (ONH and the NFL oculus uterque [OU] scan report) in the same scan as shown in [Figure 4]. The description of the printout should be done under the following headings [Figure 4]:
Figure 4: Cirrus scan printout

Click here to view


  1. Patient information: This forms the first part of information for the printout and consists of the age, sex, examination date, time, date of birth, and the patients' registration number. This needs to be fed into the machine once and recalled for the subsequent scans later. A mistake in the entry of patient information during subsequent scans will lead to the machine considering it as another patient and not using it for comparisons
  2. Quality of the image: The quality score of the image is calculated using proprietary methods and is designated differently for the different instruments. The acceptability standard (quality) of the images for consideration in progression analysis or interpretation of a single scan is defined by the manufacturer and is to be adhered to. The best possible and consistent quality of image is to be aimed at; during acquisition of serial scans, to minimize errors of judgment in interpretation [5]
    It is prudent to remember changes in the ocular media occurring in the individual patient over a period of follow-ups and factor it in the interpretation. Gross changes in the image quality due to cataract, for example, should prompt acquisition of the images after cataract extraction as media opacities have a large bearing on scan parameters and final result.
  3. It shows three tomography scan images, two for the disc in horizontal and vertical meridian above (with the demarcated limits of disc which is the Bruch's membrane opening [BMO]) and the retinal nerve fiber layer (RNFL) image below that has marked segmentation lines. The segmentation line location should be inspected for propriety before continuing to interpret the scans further. In cases of doubt over the segmentation, the image should be seen on the instrument screen in magnified view. The printout may sometimes be indistinct, especially for the thinned NFL, and its verification may be prudent to certify accurate segmentation. Any changes therein could be made if deemed necessary to avoid errors in subsequent interpretation
  4. NFL thickness map (Cirrus) is a topographical display of the NFL. A typically normal map shows an hour glass shape of yellow and red colors around the disc with the color scale in microns on the left of the image for reference
  5. RNFL deviation map: It has an enface fundus image and shows the machine-derived boundary of the cup and the disc and also the calculation circle placement for the NFL. It depicts the deviation from the normative database in the form of color-coded superpixels. This map is only available in the Cirrus machine
  6. Table of parameters of the optic disc and the NFL in comparison to the normative database (Cirrus). A distinct color coding is given to the acquired values depends on their statistical relation to the values of the database of age-matched normals used and stored in that machine. It is to be noted that all the machines may not have a normative database representative of the examined population. It may also not be exhaustive enough to cover all the anatomic variations. Hence, stand-alone use of the classification for decision to classify as normal or disease would have limitations and is best avoided. These should only be used as a guideline. A working knowledge of the database of one's own machine would help in interpretation. The values of the thinnest 1% of normative database are marked red and considered "outside normal limits." The values that fall between 1% and 5% and including 5% are labeled in yellow and considered borderline/suspect. From >5% to 95% is the wide range for normal values and marked in green. The values >95% of the range is labeled white and indicates the thickest range of normative database. This statistical rule is uniform for all the machines with database installed irrespective of make. However, the absolute actual reference values and the color code may differ on the composition of the database and machine make
    The most useful parameters to follow on the Cirrus were found to be the vertical cup disc ratio, the rim area and the mean NFL thickness, NFL in the lower temporal zone, and average lower zone thickness in many studies till date. The values on the display that are not compared to the normative data or lie outside its scope are shaded gray. If the disc area is <1.3 or >2.5 mm 2 , most of the interpretation goes out of the scope of the normative database. Normative database is not available for patients under 18 years of age in Cirrus.
    This table of parameters for the disc is generated by using the termination of the BMO as the limit of the disc margin and the shortest perpendicular from that point to the internal limiting membrane (ILM) (minimum bandwidth) to define the cup limit in each scanned slice. It is to be remembered that the limits of the BMO may not match the limits of the disc seen clinically.
    The extracted tomograms, horizontal and vertical for each eye at the bottom of the report, can give an idea if these landmarks have been defined correctly by the machine software after acquisition. An ideal scan should show the exact location of the BMO and the shortest perpendicular to the ILM in the above images both marked in color. The region above that point on the ILM is the rim and the region below is considered the cup for the sake of calculation.
  7. Curves depicting neuroretinal rim and the RNFL measurement in temporal superior nasal inferior temporal (TSNIT) format compared to the normative data (color coded) are given in the center with a continuous and a hatched line representing the respective curves for either eye. It would be useful to verify change if any from the previous scan by looking up the segmentation on the instrument screen rather than the printout as small changes might be missed
  8. RNFL quadrant and clock hour average are given below the NFL maps and color coded in the same scales as rest of the report (based on their P value). They specify the location of the pathology sector clock hour wise. Their importance in early defects is the fact that a small clock hour defect may be ironed out if the sectoral average (larger area) is considered and hence both have to be considered together. The area of pathology (e.g., NFL defect) can be correlated to the expected clock hour of change on the image.


The scans in the Spectralis OU printout [Figure 5] in the latest software upgrade have the noise reduction rating and the NFL TSNIT curves along with the inter-eye asymmetry and the quadrant and the sectoral analysis. A noteworthy fact in the latest version is the availability to choose one of the three scan circles, i.e., 3.5, 4.1, and 4.7 for measurement of the NFL to bypass peripapillary pathology that can affect it.
Figure 5: Spectralis printout

Click here to view


Being an OCT-scanning laser ophthalmoscope (SLO), the image of the fundus on the printout is a simultaneous infrared image acquired during the scanning process that has superimposed scan circles with the one utilized for analysis being highlighted (most commonly the 3.5 circle, that is, innermost is used).

Among the major differences are the scan quality scores that are depicted differently with respect to the Cirrus and that there is no calculation of the disc parameters except the ONH height profile, the mean rim width, and the BMO area (representative of the disc size) of the disc. Tabulation of numerical values of the disc and the rim area and the cup-to-disc ratio are not included unlike the Cirrus device and there is no RNFL deviation map. RNFL parameters and the sector-wise values and averages are given but are not in clock hour format. The methods of interpretation of the scan and the sequence, however, remain the same in the two machines.

An important part of the imaging in glaucoma is the monitoring for progression. These machines have methods to detect and depict progression. The fundamental requirement for this functionality is to have repeatable measurements at the same place in the same eye accurately throughout the time of follow-up. A very few machines have the ability of fast eye tracking and registration, cyclotorsion compensation, and very fast scanning speeds to achieve this. The above two machines among a few others have this capability and have been studied to give reproducible results with minimal intra- and inter-test variability.

The Cirrus has the software called the Glaucoma Progression Analysis on similar lines as that of the Humphrey perimeter [Figure 6a] where two baseline scans are later compared to the follow-up scans. A parametric table is also provided in the analysis and color coded to depict change over time, the colors indicative of the likelihood of progression [Figure 6b], with their significance being mentioned in the lower right of the printout. The machine thus is capable of giving a trend as well as an event analysis like the one in perimetry and is a good feature to follow-up the disease.
Figure 6a: Cirrus progression printout

Click here to view
Figure 6b: Parametric projection

Click here to view


The Spectralis [Figure 7], however, takes a single baseline scan and does a pointwise comparison of each follow-up scan to the original baseline over a period of time to pick up the change in the NFL profile as shown in the figures. The Spectralis affords only an event-based analysis for change in NFL profile.
Figure 7: Spectralis follow up printout

Click here to view



  Macular and Ganglion Cell Analysis Top


The improvement in the resolution of the devices coupled with accurate segmentation protocols has led to the ability to examine the ganglion cells and their changes in cases of glaucoma and other pathologies. Various instruments have different patented protocols for scanning the macular region to assess the changes in early glaucoma as rationally these changes should anatomically predate the changes in the NFL.

In cases that had a structurally normal macular region, the macular assessment by the SD-OCT has been proven to compare favorably with the NFL measurements and correlate to the visual field defects. [6],[7]

The Cirrus segments the GCC including the IPL and the ganglion cell bodies but excludes the NFL. The Optovue includes the NFL too, whereas the Spectralis measures the total macular thickness.

In the Cirrus platform, the device acquires data from a cube through a square grid of 6 mm × 6 mm of 128 B scans, each consisting of 512 A scans. The GCC algorithm measures thickness of the macular GCC in a 6 mm × 6 mm × 2 mm annulus centered on the fovea that has a inner vertical diameter of 1 mm and outer diameter of 4 mm. This is done to avoid the areas difficult to segment and measure accurately that lie within and outside this annulus. [8]

The data thus obtained is divided into six parts and the mean thickness and the minimum mean thickness values are generated and presented in the analysis as shown in [Figure 8]. These data are also color coded in a manner similar to the ONH-RNFL protocol for statistical probabilities. The earliest changes have been noted in the mean minimum thickness in the inferotemporal sector according to most of the studies till date.
Figure 8: Methods of macular segmentation

Click here to view


In the Spectralis, the entire macular thickness from the retinal pigment epithelium to the ILM is presented in the analysis superimposed on the SLO image of the macula, and a macular asymmetry map of the upper and the lower half of the macula is also generated as shown in [Figure 8]. The statistical significance with respect to normative data, however, is not calculated in this machine. Their newer software is capable of segmenting and presenting analysis of each individual layer of the macula.

Both the above devices can generate a composite display to include the optic disc, NFL, and the macular scans in one printout to help clinically correlate the structural changes to each other and to the visual fields [Figure 9]. The forum is proprietary software by Zeiss that integrates the visual field printout to the OCT images for structure-function correlation [Figure 10].
Figure 9: Combination printout

Click here to view
Figure 10: Zeiss forum printout

Click here to view


The major drawback of using macular analysis is the wide range of comorbid conditions that affect this area, they can affect their utility on long-term follow-up (e.g., diabetic macular edema, foveal traction, hole formation, and cystoids macular edema,). All these changes can affect the anatomy and interfere with segmentation protocols and alter the macular thickness and act as confounders to reduce the specificity of the test.


  Artifacts in Imaging Top


Despite these remarkable advancements, evidence indicates that SD-OCT imaging artifacts are a relatively common finding in clinical practice. [9],[10],[11],[12] For convenience, we can classify factors affecting OCT scan quality as patient dependent (media opacities, pupil diameter), operator dependent (centration, signal strength of acquired image), and device dependent (acquisition speed and database related).

The scan quality can be affected in eyes with a pupil diameter lesser than 2 mm. Recent studies did not find significant changes in RNFL thickness before and after dilation, suggesting that Cirrus HD-OCT results should not be affected by pupil size. But as a rule, it would be better to dilate to decrease the variability, especially if artifacts are seen with normal pupil size. [13],[14]

OCT studies have shown that dry eye and/or cataract diminish scan quality index and decrease RNFL thickness measurement.

The deleterious effects of cataract on OCT scan quality are more difficult to overcome, unless cataract surgery is performed and it is best to avoid imaging in the presence of catarctous process in imaging axis. [15],[16]

Newer OCT acquisition time is <2 s, making these systems excellent for routine clinical use. However, blinks may still occur during the acquisition time. In the absence of an eye tracking system, the acquisition process continues uninterrupted even in the presence of blinks. This leads to a transient loss of data, which is proportional to the duration of the blink. It is reflected as a loss of data on the final analyzed image.

Patients are commonly instructed not to blink during camera alignment and scan acquisition; however, this may cause tear film evaporation and breakup, especially in patients with preexisting ocular surface disorders or drug toxicities. Therefore, careful observation of the live funduscopic "en-face" image and the OCT tomograms is recommended as it may reveal tear film disruption on the live image and signal degradation with color attenuation on the OCT tomograms.

Patients should be encouraged to continue physiologic blinking during alignment and blink just before the scan so as to tide over the blink during acquisition process. [17]

Floaters and other vitreous opacities decrease the scan quality by interfering with the light beam path [Figure 11]. The retinal blood vessels also cause shadowing artifact [Figure 12]a and though a normal occurrence a close note is to be made at the segmentation line below the vessel, especially if it falls in an abnormal area. Floater affects the OCT measures when located within the scan area. When a floater is located on the scan circle, a vertical shadow of signal attenuation/interruption is visible in the corresponding area of the circular tomogram. A floater outside the scanning circle would not usually matter.
Figure 11: Artifacts

Click here to view
Figure 12: (a) Effect of blood vessels (shadowing on segmentation). (b) Vessel breaks on the B-scan image indicate problems in scan acquisition (motion artifact). (c) Segmentation errors (artifact) in high myopia -note that the lines do not confirm to the retinal nerve fiber layer but are irregular

Click here to view


It can be moved away by asking the patient to produce brief to-and-fro eye movements immediately before scan acquisition and acquire at the time the floater exits the scan path. The same can be repeated on all subsequent scans.

In addition to the floaters that cause posterior shadowing, other vitreopathies may be responsible for RNFL thickness changes. Notably, the posterior vitreous detachment can produce the significant changes in the NFL measurements and should be kept in mind while seeking an explanation for sudden and unexpected changes in the TSNIT profile. The attachments of the vitreous face if seen in previous scans can give a clue to their occurrence.

Motion artifacts result from eye movements [Figure 12]b, such as saccades, during scan acquisition. They typically appear on the en-face image as horizontal shifts/breakage of blood vessels' path.

Improvements in SD-OCT scanning speed and acquisition time have reduced the likelihood of motion artifacts in modern-day OCT machines. However, eye movements still represent a potential problem for devices lacking an eye tracking system or motion correction algorithms.

Scan should be repeated, particularly with motion artifacts intersecting the scan circle and/or the optic disc area.

The signal-to-noise ratio, or in other words, the strength of the light signal backscattered by ocular structure is calculated and has been conventionally used as an objective measure of scan quality. Scan quality scores vary as per the machine and the manufacturer. Therefore, signal strength maximization nearest to the highest value suggested by the maker and repeatable scans of the same signal strength have to be aimed for.

The factors such as dry eye and media, improper objective lens cleaning, or poor image centration may also play a role and are perfectly and easily avoidable in responsible hands.

Several studies have shown that scans with better signal strength are associated with higher RNFL thickness measures. [18],[19]

OCT image truncation may commonly occur in myopic eyes with steep retinal curvature or peripaillary staphylomas [Figure 12]c.

Other causes include improper distance between the eye and the device due to incorrect patient positioning or axial misalignment of the OCT scanning head.

Improper patient positioning was also important in previous generation machines but no longer an issue with tracking, auto-rescan, and alignment features of newer machines.

Most machines provide disc parameters such as the average and vertical cup-to-disc ratios and cup volume through automated delineation of the optic disc and cup margins. Adequate optic disc assessment relies on the ability of the automated algorithm to identify the termination of Bruch's membrane, corresponding to the optic disc edge.

Improper identification of the disc limits as seen clinically may be due to the large areas of parapapillary atrophy that is variable and progressive in patients with myopia and glaucoma. The machines identify the BMO as the limit of the clinical disc and hence in cases with a large gamma zone or vessels located at the edge of the disc, the parameters may take a beating although the NFL and macula may still give a good clue if the changes do not encroach the scan areas.

Segmentation errors due to improper identification of the limits of the structures by the segmentation algorithms can occur due to shadowing artifacts, presence of optically dense membranes, and tractional bands apart from highly myopic eyes that have poor signal strength on the images. Looking at the scans before looking at their interpretation unravels these changes and help decision-making.

Improper database vis-a-vis the patient ethnicity can lead to faulty classification in spite of having normal measurements. A working knowledge of the database of the instrument can prevent an overkill or underdiagnosis.

Although the artifacts seem plenty, almost all are avoidable, and with meticulous technique and proper knowledge, the device can be used in a robust manner.


  Incorporation of Optical Coherence Tomographer into Practice Top


While incorporating this device into practice, one should remember the following:

  1. It is important to dilate the pupil as a standard protocol for imaging (unless contraindicated) for consistent imaging quality [9]
  2. It is essential for a machine used for glaucoma assessment to have fast eye tracking, autoalignment, and fast scan acquisition speeds
  3. It is important and mandatory to use the same machine, same place, and reproducible image quality (with attention to media haze) to ensure a meaningful evaluation and follow-up
  4. Printout formats that do not include actual scan images should not be used
  5. Use of progression software is ideal if not mandatory similar to visual field tests. Remember at all times that it is the progression that we are looking for
  6. Confirm progression on repeat scans in cases of suspect in scan quality and clinically correlate on fields, especially if perimetric glaucoma, although the changes in both modalities may not be parallel to one another and structural damage precedes functional damage on most occasions. Posterior vitreous changes have to be paid attention to and looked for, in cases of unexplained changes on scans
  7. Structure function printouts are helpful but not mandatory. They serve as a good education and documentation tool. [10]


Drawbacks of optical coherence tomographer imaging in glaucoma

  1. Artifacts and anomalies of the scanning process affect all the machines due to inherent drawbacks or limitations in the machine hardware or anatomical variation in the individual patients. It is important to recognize the common sources of error and avoid them in clinical practice
    Common examples are as follows:
    1. The effect of blood vessels and floaters on segmentation due to their shadowing
    2. Effect of large parapapillary staphyloma on the quality of image due to the uneven acquisition plane that results in abnormal segmentation
    3. Floor effect due to the residual glial tissue that leads to a plateau in the NFL loss in cases with advanced glaucomatous damage
    4. The effect of retinal pathology on the measurements, for example, epiretinal membranes, macular edema, vascular occlusions, and postvitreo retinal surgery
  2. Progression on the OCT is not defined unlike the case of visual field tests and hence it has to be subjectively interpreted keeping in mind the limits of resolution of the individual machine
  3. All OCT devices do not have the ability of fast gaze track and alignment hardware and software for progression, and hence have a limitation for their use in glaucoma follow-ups.



  Conclusion Top


Meticulous attention to detail and adequate knowledge of the machine and its usage can ensure proper use of this device for glaucoma diagnosis and progression with reasonable certainty.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

 
  References Top

1.
Sung KR, Kim DY, Park SB, Kook MS. Comparison of retinal nerve fiber layer thickness measured by Cirrus HD and Stratus optical coherence tomography. Ophthalmology 2009;116:1264-70.e1.  Back to cited text no. 1
    
2.
Kim JS, Ishikawa H, Sung KR, Xu J, Wollstein G, Bilonick RA, et al. Retinal nerve fibre layer thickness measurement reproducibility improved with spectral domain optical coherence tomography. Br J Ophthalmol 2009;93:1057-63.  Back to cited text no. 2
    
3.
Soltani-Moghadam R, Alizadeh Y, Kazemnezhad Leili E, Absari Haghighi M. Reproducibility of peripapillary retinal nerve fiber layer thickness measurements with Cirrus HD-OCT in glaucomatous eyes. Int J Ophthalmol 2015;8:113-7.  Back to cited text no. 3
    
4.
Bussel II, Wollstein G, Schuman JS. OCT for glaucoma diagnosis, screening and detection of glaucoma progression. Br J Ophthalmol 2014;98 Suppl 2:ii15-9.  Back to cited text no. 4
    
5.
Cheung CY, Leung CK, Lin D, Pang CP, Lam DS. Relationship between retinal nerve fiber layer measurement and signal strength in optical coherence tomography. Ophthalmology 2008;115:1347-51.e1-2.  Back to cited text no. 5
    
6.
Kim NR, Lee ES, Seong GJ, Kim JH, An HG, Kim CY. Structure-function relationship and diagnostic value of macular ganglion cell complex measurement using Fourier-domain OCT in glaucoma. Invest Ophthalmol Vis Sci 2010;51:4646-51.  Back to cited text no. 6
    
7.
Shin HY, Park HY, Jung KI, Park CK. Comparative study of macular ganglion cell-inner plexiform layer and peripapillary retinal nerve fiber layer measurement: Structure-function analysis. Invest Ophthalmol Vis Sci 2013;54:7344-53.  Back to cited text no. 7
    
8.
Raza AS, Cho J, de Moraes CG, Wang M, Zhang X, Kardon RH, et al. Retinal ganglion cell layer thickness and local visual field sensitivity in glaucoma. Arch Ophthalmol 2011;129:1529-36.  Back to cited text no. 8
    
9.
Asrani S, Essaid L, Alder BD, Santiago-Turla C. Artifacts in spectral-domain optical coherence tomography measurements in glaucoma. JAMA Ophthalmol 2014;132:396-402.  Back to cited text no. 9
    
10.
Keane PA, Mand PS, Liakopoulos S, Walsh AC, Sadda SR. Accuracy of retinal thickness measurements obtained with Cirrus optical coherence tomography. Br J Ophthalmol 2009;93:1461-7.  Back to cited text no. 10
    
11.
Ho J, Castro DP, Castro LC, Chen Y, Liu J, Mattox C, et al. Clinical assessment of mirror artifacts in spectral-domain optical coherence tomography. Invest Ophthalmol Vis Sci 2010;51:3714-20.  Back to cited text no. 11
    
12.
Han IC, Jaffe GJ. Evaluation of artifacts associated with macular spectral-domain optical coherence tomography. Ophthalmology 2010;117:1177-89.e4.  Back to cited text no. 12
    
13.
Savini G, Carbonelli M, Parisi V, Barboni P. Effect of pupil dilation on retinal nerve fibre layer thickness measurements and their repeatability with Cirrus HD-OCT. Eye (Lond) 2010;24:1503-8.  Back to cited text no. 13
    
14.
Massa GC, Vidotti VG, Cremasco F, Lupinacci AP, Costa VP. Influence of pupil dilation on retinal nerve fibre layer measurements with spectral domain OCT. Eye (Lond) 2010;24:1498-502.  Back to cited text no. 14
    
15.
Mwanza JC, Bhorade AM, Sekhon N, McSoley JJ, Yoo SH, Feuer WJ, et al. Effect of cataract and its removal on signal strength and peripapillary retinal nerve fiber layer optical coherence tomography measurements. J Glaucoma 2011;20:37-43.  Back to cited text no. 15
    
16.
Bambo MP, Garcia-Martin E, Otin S, Sancho E, Fuertes I, Herrero R, et al. Influence of cataract surgery on repeatability and measurements of spectral domain optical coherence tomography. Br J Ophthalmol 2014;98:52-8.  Back to cited text no. 16
    
17.
Stein DM, Wollstein G, Ishikawa H, Hertzmark E, Noecker RJ, Schuman JS. Effect of corneal drying on optical coherence tomography. Ophthalmology 2006;113:985-91.  Back to cited text no. 17
    
18.
Vizzeri G, Bowd C, Medeiros FA, Weinreb RN, Zangwill LM. Effect of signal strength and improper alignment on the variability of stratus optical coherence tomography retinal nerve fiber layer thickness measurements. Am J Ophthalmol 2009;148:249-55.e1.  Back to cited text no. 18
    
19.
Wu Z, Vazeen M, Varma R, Chopra V, Walsh AC, LaBree LD, et al. Factors associated with variability in retinal nerve fiber layer thickness measurements obtained by optical coherence tomography. Ophthalmology 2007;114:1505-12.  Back to cited text no. 19
    


    Figures

  [Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6a], [Figure 6b], [Figure 7], [Figure 8], [Figure 9], [Figure 10], [Figure 11], [Figure 12]


This article has been cited by
1 Three dimensional neuro-retinal rim thickness and retinal nerve fiber layer thickness using high-definition optical coherence tomography for open-angle glaucoma
Suresh Subramaniam,Jin Wook Jeoung,Won June Lee,Young Kook Kim,Ki Ho Park
Japanese Journal of Ophthalmology. 2018;
[Pubmed] | [DOI]
2 A longitudinal observation of a patient with normal tension glaucoma
Hlupheka L. Sithole
African Vision and Eye Health. 2017; 76(1)
[Pubmed] | [DOI]



 

Top
 
 
  Search
 
Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

 
  In this article
Abstract
Clinically Relev...
Interpretation o...
Macular and Gang...
Artifacts in Imaging
Incorporation of...
Conclusion
References
Article Figures

 Article Access Statistics
    Viewed7121    
    Printed33    
    Emailed0    
    PDF Downloaded1013    
    Comments [Add]    
    Cited by others 2    

Recommend this journal