|Year : 2019 | Volume
| Issue : 1 | Page : 75-77
Reeja Choondanil Menon
Department of Ophthalmology, Nethra Eye Care Centre, Thrisshur, Kerala, India
|Date of Web Publication||15-Apr-2019|
Reeja Choondanil Menon
Department of Ophthalmology, Nethra Eye Care Centre, Thrisshur, Kerala
Source of Support: None, Conflict of Interest: None
|How to cite this article:|
Menon RC. Journal Review. Kerala J Ophthalmol 2019;31:75-7
| Integrating Macular Ganglion Cell Inner Plexiform Layer and Parapapillary Retinal Nerve Fiber Layer Measurements to Detect Glaucoma Progression|| |
Hou HW, Lin C, Leung CK. Integrating macular ganglion cell inner plexiform layer and parapapillary retinal nerve fiber layer measurements to detect glaucoma progression. Ophthalmology 2018;125:822-31.
Glaucoma, is a chronic progressive optic neuropathy, resulting in characteristic visual field loss. Strategies to detect disease progression in the early stages are vital for making decisions regarding treatment initiation and modification and hence that patients can preserve a good quality of vision throughout their life. Imaging technologies like optical coherence tomography (OCT) are reliable and extensively studied tools in quantifying the structural progression of glaucoma.
Changes in parapapillary retinal nerve fiber layer (RNFL) and macular ganglion cell-inner plexiform layer thickness have been studied and found to be useful in evaluating glaucoma progression. The significance of parapapillary RNFL lies in the fact that the axons of retinal ganglion cells converge at the disc while macula having the highest retinal ganglion cell density is reliably imaged for ganglion cell-inner plexiform layer (GCIPL) thickness. Despite this, till date, no consensus has been reached regarding the superiority of either parameter for adoption in the monitoring for progression. This is due to conflicting data which in some cases have demonstrated that parapapillary RNFL parameters are better (Oddone et al.). Others have reported macular GCIPL to be more useful in longitudinal studies.
In this prospective study, the authors investigated the relationship between progressive parapapillary retinal nerve fiber layer thinning, progressive macular GCIPL thinning, and visual field progression in patients with primary open angle glaucoma (POAG). The study establishes the feasibility of integrating the former two parameters for efficacious monitoring of early progression. One hundred and thirty-six patients with POAG (231 eyes) were recruited and followed up at approximately 4-month intervals for at least 5 years. OCT imaging of the macular GCIPL and parapapillary RNFL and standard automated perimetry was done along with routine investigations. The changes were analyzed by guided progression analysis (GPA: Carl Zeiss Meditec), an event-based change detection algorithm. Specificity of GPA for the detection of progression was ascertained by recruitment of 67 eyes of 36 healthy individuals at similar intervals for >5 years for OCT imaging of RNFL and GCIPL, visual fields as well as 26 eyes of 26 healthy individuals for 1 year followed up more closely (weekly for 8 consecutive weeks). Patients were treated during the period to maintain target intraocular pressure, analysis of progression being masked, though thickness reports were available.
Study findings showed that 53 eyes (22.3%) showed either progressive RNFL or GCIPL thinning. In 35 eyes, that had both RNFL and GCIPL thinning, 54.3% had progressive GCIPL thinning detected before RNFL thinning. About 37.1% detected vice versa. It was also found that both are connected to an increased risk of visual field progression. The risk of VF progression increased by 2.54 fold (possible progression) to 3.66 fold (likely progression) in eyes with progressive RNFL thinning and increased by 2.74 fold (possible progression) to 3.48 fold (likely progression) in eye with progressive GCIPL thinning as compared to eyes without progression of GCIPL and RNFL thinning. Furthermore, progressive RNFL and GCIPL thinning were mutually predictive. Progressive RNFL thinning leads to 2.99 fold increase in the risk of GCIPL thinning, whereas vice versa, the risk is 5.27-fold.
It has been hypothesized that the retrograde degeneration of axons due to damage initiated art lamina cribrosa in glaucoma leads to initial parapapillary RNFL thinning while shrinkage of dendrites in GCIPL before axonal damage after optic nerve injury leads to initial GCIPL thinning.
Although the study lends a strong cause to integrate the OCT GCIPL and parapapillary RNFL thickness for glaucoma monitoring, it is not without pitfalls. The reason for the temporal disparity between dendritic and axonal degeneration in glaucoma, and therefore, GCIPL and RNFL thinning is indeterminate. GCIPL analysis misses progressive thinning outside the annular area centered around the fovea. GPA cannot distinguish between age-related and disease-related thinning and progression, though significantly less percentage (9% and 4.5%) healthy eyes demonstrated progressive RNFL and GCIPL thinning, respectively. GPA performance is also linked to frequency of testing, disease progression id difficult to identify with prolonged intervals between tests. The area and rate of change of thinning are also unreported. It has been shown that trend-based progression analysis, a new algorithm outperforms GPA, the latter being an event-based analysis as the former provides a rate of progression over time. More importantly, macular disease in the form of edema, epiretinal membrane can interfere with GCIPL analysis.
In conclusion, the integration of GCIPL and RNFL thicknesses in addition to visual fields seems to be an effective tool for monitoring glaucoma progression. The use of a high speed, wide field OCT imaging system covering the macula and parapapillary region (for example, 12 mm × 9 mm in Topcon Triton swept source OCT system) in a larger scan area would improve the detection of such thinning.
| Predicting Pseudophakic Refractive Error: Interplay of Biometry Prediction Error, Anterior Chamber Depth, and Changes in Corneal Curvature|| |
Wallace HB, Misra SL, Li SS, McKelvie J. Predicting pseudophakic refractive error: Interplay of biometry prediction error, anterior chamber depth, and changes in corneal curvature. J Cataract Refract Surg 2018;44:1123-9.
To comply with patients' demands of spectacle independence, cataract surgery today assumes the role of refractive surgery. To this effect, IOL calculation formulas assume prime importance. These have been constantly evolving from the earlier regression-based formulas, which were improved on by adding factors included in their predictions, then progressing to theoretical formulas which have their foundation on geometrical optics.
Older formulas, including the Holladay I, SRK/T and Hoffer Q, required putting in axial length and K-reading. Then came the Olsen formula using four predictors: axial length, K-reading, anterior chamber depth, and lens thickness. Holladay II formula uses seven variables to predict the effective lens position. Olsen and Barrett later formulated new seven-variable formulas of their own using axial length, K-reading, anterior chamber depth, lens thickness, corneal diameter, and refraction. The Barrett Universal Formula is a more modern formula based on a theoretical eye model in which anterior chamber depth (ACD) is related to axial length and keratometry and is meant to work for all eyes. It differs from others in that the location of the principal plane of refraction of IOL is a variable in the formula. Hill-RBF (radial basis function) is a data-based formula, incorporating data from thousands of eyes. It uses existing data and artificial intelligence to predict postoperative refraction. Apart from the choice of IOL formula, postoperative refractive outcomes change due to increasing in ACD and variability in corneal curvature.
This prospective case series compared the Barrett Universal II Hill RBF and SRK/T formulas for accuracy using a predefined protocol with a single type of IOL and a single surgeon apart from evaluating the temporal effects of ACD and keratometric changes on postoperative refraction at 3 months. Patients enlisted for cataract surgery were enrolled in the study. A total of 100 eyes of 100 patients were assessed. The average age was 74.4 ± 9.1 years, mean axial length was 23.4 ± 1.1 mm, mean keratometry was 44.0 ± 1.7 D. Biometry measurements were done using partial coherence interferometry (IOLMaster500), using the Hill-RBF method or Barrett Universal II formula for IOL power calculation. IOL power selected predicted myopic postoperative refraction closest to plano, also ensuring neither formula predicted hyperopia >0.09 D.
Data were entered into Barrett Universal II formula using a program written by one of the authors. SRK/T formula was encoded into the statistical analysis and results validated against the output of IOLMaster. Biometry data were manually entered by the same investigator into online Hill–RBF calculator. The A constant entered for these calculations was 119.1. Phacoemulsification was done by temporal approach with two paracentesis incisions by the same surgeon using Provisc for capsule stabilization before implanting single-piece acrylic IOL.
Objective refraction was measured postoperatively at 1 week, 1 month, and 3 months by the same investigator using autorefractometer (combined refraction and keratometry). ACD was assessed using dual Scheimpflug analyzer. Refraction was rounded off to the nearest 0.25 D. Biometry prediction error was calculated as the actual postoperative refraction minus the refractive result predicted by each formula.
The Barrett Universal II or Hill-RBF methods are likely to lead to better refractive outcomes for patients as compared to SRK/T. According to the study, the former 2 formulas predicted the highest proportion of eyes within ± 0.5 D, ±1.0 D and ± 2.0 D at 3 months, while the SRK/T resulted in the lowest predictions in each category. The Barrett formula also predicted the highest proportion of medium length eyes (>22.0 to <24.5 mm) with postoperative refractions within 0.25, 0.5, and 1 D. At 1 week, the BU-2 had significantly less prediction error than both the other two formulae (P < 0.01). These differences evened off at 1 month (P = 0.57) and 3 months (P = 0.29), lending to the conclusion that the accuracy of the methods as the highest at 1 week. The reason for this was postulated to be changed in corneal curvature/keratometry and lens position changes.
There was found to be a mean increase of corneal curvature during the first 4 postoperative weeks. The myopic shift that could result was negated by the hyperopic shift from posterior lens movement, therefore, leading negligible change in refraction. From 4–12 postoperative weeks, small keratometric flattening was observed, leading to a hyperopic shift of a quarter of a diopter. The same could also presumably result from gradual posterior IOL displacement due to capsular contraction. The findings suggest that a progressive hyperopic shift of about 0.19 D could be expected to occur.
Study limitations include small sample size, novel and previously undescribed method of assessing ACD using a new software program and using an autorefractor instead of subjective manifest refractions.
Based on the study, the authors have concluded that presently, the Barrett Universal II formula and Hill-RBF method are the best options for IOL calculation in centers using partial coherence interferometry-based biometry. Significant increase in ACD resulted in refractive changes leading to hyperopic shift in weeks 4–12 which decreased prediction accuracy. It has been hypothesized that retained OVD, when not routinely removed from behind the IOL may alter final lens position and refractive outcome, especially since posterior lens migration was demonstrated during the study.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.