Avoid color-scheme bias and vet all data before making treatment decisions
Glaucoma is the leading cause of irreversible blindness worldwide and will be an increasingly significant burden on patients as they age.1 Simply put, glaucoma is a progressive disease that causes irreversible damage to the retinal ganglion cells and their axons, altering both structure and function. As primary eye care providers, optometrists must step up to manage this burden, with the ultimate goal of enhancing patients’ lives, through reducing the risk of functional vision loss.
Although comprehensive analyses of structure and function are necessary to diagnose and manage glaucoma, spectral domain optical coherence tomography (SD-OCT) has become indispensable by revolutionizing the ability to objectively map glaucomatous tissue loss and quantify the depth of said loss. As with any evolving technology, experience has taught us to recognize that although OCT has greatly enhanced our abilities as clinicians, diagnosing and tracking disease with OCT requires understanding its limitations. ODs must recognize that humans still drive the technology and need to ensure the quality of the information it provides.
Prior to examining how to improve image quality awareness to improve decision-making, ODs should understand the basic information that they must collect from these machines to maximize their value in assessing glaucoma. Although each OCT platform has its own specific scan protocols and analyses that are not directly comparable with each other, at a bare minimum, both circumpapillary retinal nerve fiber layer (RNFL) and macular scans must be performed to confirm that a patient has glaucoma and to understand the extent and topographical distribution of structural damage present.
Most ODs, concentrating on the fact that glaucoma is an optic neuropathy, will routinely obtain scans of the optic nerve and RNFL. However, more than 30% of retinal ganglion cells (RGCs) are found in the macular region (central Å}8Åã). Unsurprisingly, early glaucoma has been shown to involve widespread shallow and deep local damage to the RNFL in the macula.2 Failure to scan the macula in suspected or known glaucoma cases can result in underassessing damage because of lack of perspective, which could lead to underinformed management. Mwanza et al reinforced this notion in 2014 when they concluded that a combination of RGC and RNFL thickness outperformed those individual parameters by themselves in diagnosing glaucoma.3
Ordering and interpreting optic nerve and macular information is the current standard in glaucoma care and must be universally embraced in optometric management. Fortunately, 1-page reports such as PanoMap (Carl Zeiss Meditec, Inc) and the Hood Report (Topcon and Heidelberg platforms) blend this information for easier and more intuitive analyses.
Watch for artifacts
When reviewing the optic nerve and macular data supplied by individual OCT platforms, maintain a “trust, but verify” mind-set. Preparing for the inevitability of errors is a more realistic approach to imaging than being surprised when errors occur.
Research shows that artifacts that affect the quality and subsequently the utility of imaging are common in clinical practice. Results from a 2016 review and analysis of this topic states that more than 25% of patients were be expected to have SD-OCT glaucoma imaging artifacts that could introduce false-positive or false-negative interpretation.4
Many factors result in artifacts or errors of which ODs must be aware to maximize the scan signal strength and reflectivity. Potential errors can be broken into 3 categories: patient dependent, operator dependent, and machine dependent.4,5 A nonexhaustive list is provided below for perspective:
– PATIENT-DEPENDENT FACTORS: Age (relative to reference database), pupil size, tear film quality, media opacities (cataracts, vitreous condensation, etc), eye movements, blinking, epiretinal membranes, myopia/ increased axial length, abnormal optic nerve insertion, parapapillary atrophy, cyclotorsion, previous congenital or acquired optic neuropathies, or macular conditions
– OPERATOR-DEPENDENT FACTORS: Poor alignment of scan (axial, rotational, centration), incorrect patient positioning, or OCT lens opacities
– MACHINE-DEPENDENT FACTORS: Inaccurate segmentation of tissue or inaccurate delineation of disc margins
After addressing artifacts that may affect data utility, ODs must understand how the various machines present normal and a deviation from that normal— and when to recognize that “green isn’t always clean.” Each instrument platform has its own normative database comprising individuals of various ages, races, and sex. Remarkably, the FDA has no standard or guidelines for the types or numbers of participants who should be included in these databases or how those data should be presented or analyzed.6
A more realistic way to classify these databases is to use the term reference rather than normative because they do not necessarily represent population norms but rather are indicative of averages of patients who met inclusion criteria as defined by individual manufacturers. Those averages are based on proprietary segmentation, scanning protocols, and thickness algorithms.7
Regardless of how the individual databases were compiled, tissue thicknesses are presented by all platforms as normal, borderline, or outside normal limits/abnormal. These thicknesses are based on the percentile distribution of database thicknesses and employ color schemes to aid in analysis. The color schemes use white as well as traffic-light colors green, yellow, and red.
White includes the 95th to the 99th percentile (above normal), green includes the 5th to the 95th (normal), yellow includes the 1st to the 5th (borderline), and red shows the 1st percentile (below normal).7 Here lies the trap: Green accounts for 90% of the tissue thickness on these machines.
From a research/probability standpoint, this may be an acceptable way of conducting business. However, relying on machine probability in clinical practice is unacceptable and can lead to a false sense of security in disease management.
A “normal” finding doesn’t mean the patient is free of disease or from progression over time. Conversely, an “abnormal” finding doesn’t mean the patient has disease, and misinterpretation of results could lead to a patient unnecessarily receiving a diagnosis of a chronic disease. For example, in the Cirrus reference database, the range of average RNFLs in the “normal” range for a 69-year old patient is 75 to 107 μm.8
Using the data
So, what is the relevance of this large range of tissue for clinical practice? Let’s step back for a moment and quickly review dynamic ranges, measurement floor, and steps. The OCT dynamic range is the usable range between the highest and lowest capturable values of tissue, based on healthy eyes. The measurement floor is the lowest detectable value beyond which measurement has no clinical value. Test-retest variability will be required multiple times in any test run because the test is repeated and this noise defines a measurement step. The test-retest variability for average RNFLs on the Cirrus platform is 4 μm.8,9
The number of steps in the overall range determines how much significant change can be detected over time: The greater the number of measurement steps in a given range, the greater the ability to capture glaucomatous changes over time.10 For example, if a dynamic range of an instrument is 20 μm and each step is 10 μm, the usefulness of that instrument is limited to detecting 2 steps of change. However, if the range is 100 μm with 10 μm steps, then 10 steps of change could be detected.
The relevance and potential wake-up call of this information are that within the “green,” a significant number of steps of RNFL progression exists, meaning there can be substantial, clinically meaningful glaucomatous loss even though the machine will categorize those changes as “normal.” As a point of reference, results from a study by Medeiros et al employing estimated RGC count at various stages of glaucoma found that a loss of 5 μm is equivalent to a loss of 100,000 RGCs. Needless to say, a lot of 5-μm steps live between 75 to 105 μm.11
An additional important point to remember is that there is a wide range of tissue thickness that could be glaucomatous regardless of the normal range within the reference database. Results from a study by Wollstein et al found that although patients with a visual field defect had an average RNFL as measured by Cirrus of 75.3 μm, patients were found to have glaucoma with a thickness of 100 μm and a patient with a thickness of 68 μm was found to be normal.12
The take-home point should be obvious: Reliance on colors for affirmation of universal normalcy for patients in the exam room is a mistake.
The amount of tissue that can be lost within the “green” is a nice segue into a brief conversation about capturing progression on OCT. It should be noted that a significant difference exists between “true” progression and “detectable” progression.13 Normal age-related attrition of RNFL is considered to be less than –1 μm per year (average of –0.54 μm per year), in which case a greater rate of change might be indicative of “true” pathological tissue loss.14
However, the OCT’s ability to detect change is limited by test-retest variability and reproducibility, which is greater than “true” change at the axonal level. On OCT printouts, RNFL measurements are provided in global averages, quadrants, sectors and clock hours. The most reproducible measurement is average/global RNFL, and the least reproducible measurement is clock hours (Figures 1-4; Table). Average/global RNFL reproducibility is approximately 4 to 5 μm, depending on the OCT platform.9,15 This has led to the informal Rule of 5, which states that if a patient with glaucoma undergoes a repeatable 5-μm or greater loss of average RNFL on consecutive scans, that loss is considered quantitative evidence of progression and might justify escalating treatment.16
Investigators recently found that the Rule of 5 must be used with caution because it is not specific for glaucoma progression between tests and performs worse longitudinally than OCT trendbased analysis.16,17 Specifically, after 5 years of semiannual testing, the use of Rule of 5 resulted in 24.8% false positives, potentially causing unnecessary treatment in patients with stable disease and unnecessary burden without benefit.
Regardless of informal rule or structured machine-based trend analyses usage, ODs cannot over-rely on OCT data alone. ODs should require a comprehensive analysis of all clinical data before intensifying a patient’s disease management.
This review of OCT in glaucoma care barely scratches the surface. Remember these guiding points in practice. First, OCT is a supplement for good clinical observation and should not be used as a sole determining factor. Second, confirming scan quality and utility always needs to be the first action before applying OCT data to decision-making. Third, avoid color scheme bias and preconceived notions of “normal” thicknesses to provide a customized analysis for the patient. Vetting disease progression requires more than simply applying easy-to-remember, informal rules. Although OCT has expanded knowledge, ODs still need to drive the machine, not let it drive them.
1. Tan NYQ, Friedman DS, Stalmans I, Ahmed IIK, Sng CCA. Glaucoma screening: where are we and where do we need to go? Curr Opin Ophthalmol. 2020;31(2):91-100. doi:10.1097/ ICU.0000000000000649
2. Hood DC. Improving our understanding, and detection, of glaucomatous damage: an approach based upon optical coherence tomography (OCT). Prog Retin Eye Res. 2017;57:46-75. doi:10.1016/j.preteyeres.2016.12.002
3. Mwanza JC, Budenz DL, Godfrey DG, et al. Diagnostic performance of optical coherence tomography ganglion cell--inner plexiform layer thickness measurements in early glaucoma. Ophthalmology. 2014;121(4):849-854. doi:10.1016/j. ophtha.2013.10.044
4. Chen JJ, Kardon RH. Avoiding clinical misinterpretation and artifacts of optical coherence tomography analysis of the optic nerve, retinal nerve fiber layer, and ganglion cell layer. J Neuroophthalmol. 2016;36(4):417-438. doi:10.1097/WNO.0000000000000422
5. Hardin JS, Taibbi G, Nelson SC, Chao D, Vizzeri G. Factors affecting Cirrus-HD OCT optic disc scan quality: a review with case examples. J Ophthalmol. 2015;2015:746150. doi:10.1155/2015/746150
6. Silverman AL, Hammel N, Khachatryan N, et al. Diagnostic accuracy of the spectralis and cirrus reference databases in differentiating between healthy and early glaucoma eyes. Ophthalmology. 2016;123(2):408-414. doi:10.1016/j. ophtha.2015.09.047
7. Sayed MS, Margolis M, Lee RK. Green disease in optical coherence tomography diagnosis of glaucoma. Curr Opin Ophthalmol. 2017;28(2):139-153. doi:10.1097/ ICU.0000000000000353
8. Cirrus HD-OCT User Manual – Models 500, 5000. Carl Zeiss Meditec Inc. 2017. Accessed April 23. 2021. https://www.zeiss.fr/ content/dam/Meditec/international/ifu/documents/cirrus-hd pct/current/2660021169012_a_cirrus_11_en_intl.pdf
9. Mwanza JC, Chang RT, Budenz DL, et al. Reproducibility of peripapillary retinal nerve fiber layer thickness and optic nerve head parameters measured with cirrus HD-OCT in glaucomatous eyes. Invest Ophthalmol Vis Sci. 2010;51(11):5724- 5730.
10. Moghimi S, Bowd C, Zangwill LM, et al. Measurement floors and dynamic ranges of OCT and OCT angiography in glaucoma. Ophthalmology. 2019;126(7):980-988. doi:10.1016/j. ophtha.2019.03.003
11. Medeiros FA, Zangwill LM, Bowd C, Mansouri K, Weinreb RN. The structure and function relationship in glaucoma: implications for detection of progression and measurement of rates of change. Invest Ophthalmol Vis Sci. 2012;53(11):6939-6946. doi:10.1167/iovs.12- 10345
12. Wollstein G, Kagemann L, Bilonick RA, et al. Retinal nerve fibre layer and visual function loss in glaucoma: the tipping point. Br J Ophthalmol. 2012;96(1):47-52. doi:10.1136/bjo.2010.196907
13. Gardiner SK, Mansberger SL, Fortune B. Time lag between functional change and loss of retinal nerve fiber layer in glaucoma. Invest Ophthalmol Vis Sci. 2020;61(13):5. doi:10.1167/iovs.61.13.5
14. Wu Z, Saunders LJ, Zangwill LM, Daga FB, Crowston JG, Medeiros FA. Impact of normal aging and progression definitions on the specificity of detecting retinal nerve fiber layer thinning. Am J Ophthalmol. 2017;181:106-113. doi:10.1016/j. ajo.2017.06.017
15. Ghasia FF, El-Dairi M, Freedman SF, Rajani A, Asrani S. Reproducibility of spectral-domain optical coherence tomography measurements in adult and pediatric glaucoma. J Glaucoma. 2015;24(1):55- 63. doi:10.1097/IJG.0b013e31829521db
16. Thompson AC, Jammal AA, Medeiros FA. Performance of the rule of 5 for detecting glaucoma progression between visits with OCT. Ophthalmol Glaucoma. 2019;2(5):319-326. doi:10.1016/j.ogla.2019.05.003
17. Thompson AC, Jammal AA, Berchuck SI, et al. Comparing the rule of 5 to trend-based analysis for detecting glaucoma progression on OCT. Ophthalmol Glaucoma. 2020;3(6):414-420. doi:10.1016/j.ogla.2020.06.005