Microperimetry in Retinal Diseases

Abstract

Retinal microperimetry (MP) is a procedure that assesses the retinal sensitivity while the fundus is directly observed, and an eye tracker system is active to compensate for involuntary eye movements during testing. With this system, the sensitivity of a small locus can be accurately determined, and it has become an established ophthalmic test for retinal specialists. Macular diseases are characterized by chorioretinal changes; therefore, the condition of the retina and choroid requires careful and detailed evaluations to perform effective therapy. Age-related macular degeneration is a representative retinal disease in which the macular function has been evaluated by the visual acuity throughout the course of the disease process. However, the visual acuity represents the physiological function of only the central fovea, and the function of the surrounding macular area has not been sufficiently evaluated throughout the different stages of the macula disease process. The new technique of MP can compensate for such limitations by being able to test the same sites of the macular area repeatedly. This is especially useful in the recent management of age-related macular degeneration or diabetic macular edema during anti-vascular endothelial growth factor treatments because MP can assess the effectiveness of the treatment. MP examinations are also valuable in diagnosing Stargardt disease as they can detect visual impairments before any abnormalities are found in the retinal images. The visual function needs to be carefully assessed along with morphologic observations by optical coherence tomography. In addition, the assessment of retinal sensitivity is useful in the presurgical or postsurgical evaluations.

INTRODUCTION

The macular area of the visual field is the most sensitive and critical region of the visual field, and conventional visual acuity tests have been the primary method of assessing the function of the fovea. Since it is not sufficient to assess only the fovea in macular diseases, new techniques and devices that evaluate the function of this critical area have been used and new ones have been recently developed.

To fulfill the unmet need for macular assessments, retinal microperimetry (MP) has emerged as a functional examination of the posterior retina. In this method, the retinal sensitivity threshold is determined as it is in static visual field tests, but an eye-tracking system is incorporated into the device so that the same retinal area can be repeatedly tested. Compared with static visual filed tests such as the Humphrey Field Analyzer (Carl Zeiss), retinal microperimetric instruments such as MP-3 (NIDEK) or MAIA (iCare) are principally developed to examine macular diseases with the optimized program.

Following the early period of this new technique of determining the sensitivity of specific loci in eyes with a macular hole or subretinal neovascularization,^1,2 MP is now widely used by retinal specialists worldwide. Recently, several commercial models of MP that combine MP with fundus imaging have been developed. Any macular disorders caused by pathologic choroidal/retinal neovascularization, degenerative chorioretinal diseases, and macular edema can be carefully evaluated by these techniques and devices. They are especially helpful in determining the effectiveness of new treatments.^3,4 This new method can be conducted easily with automated eye tracking and is now widely used in routine examinations to monitor treatments with anti-vascular endothelial growth factors (VEGF).⁵ For example, in diabetic macular edema (DME), the retinal sensitivities determined by MP coupled with optical coherence tomography (OCT) biomarkers and multifocal electroretinograms are used as functional markers for the entire macula region. Thus, earlier studies have shown that MP, together with these other methods, were able to show improvements in sensitivity following anti-VEGF therapy by ranibizumab or aflibercept in eyes with DME.^6,7 Furthermore, the macular function assessed by MP combined with structural OCT angiography (OCTA) metrics and capillary dropout areas showed low sensitivities in eyes with ischemic DME.⁸ In addition, the retinal sensitivity determined by MP and the vascular density of both the superficial and deep capillary plexuses determined by OCTA were both significantly correlated with the stage of diabetic retinopathy (DR).⁹

We present the results of recent studies of retinal MP that have provided important information on certain macular diseases. The macular diseases studied were age-related macular degeneration (AMD), Stargardt disease, DR, pathologic myopia,¹⁰ and various macular conditions that develop after surgery.

Microperimetric Instruments

Most of the early MP measurements were obtained with a scanning laser ophthalmoscope (Rodenstock, Germany), and this was followed by the first commercial, the MP-1 (NIDEK). At present, several manufacturers have developed instruments such as the MP-3 (NIDEK), MAIA (iCare), HENSON (Elektron Eye Tech), and the multimodal OCT/SLO device (OPKO Health). The MP-3 (NIDEK) device is the most advanced instrument for MP (Fig. 1). MP is a subjective examination, as are other static visual field tests and visual acuity tests. The maximum visual field is 40 degrees, and it is used with an eye-tracking system. The retinal sensitivity threshold of the MP-3 ranges from 0 to 34 decibels (dB) and for MP-1 from 0 to 20 dB. One limitation of MP-3 is the lack of a normative database and any comparative analysis software as present in the Humphrey Field Analyzer. However, it has the advantage that the points tested can be spaced at 0.1 degrees by the examiner, and the results can be overlaid on a color or red-free fundus image. These features are especially useful in assessing the retinal function in macular diseases because the damaged area and pattern of macular diseases vary and are not uniform as in glaucoma. The time required for a single MP-3 examination ranges from a few minutes to more than 8 minutes, depending on the number of points tested. Usually, 30 to 80 points within a 20-degree field are examined. In addition, the size of examination points can be set to 1 of the 5 Goldman perimetry (I, II, III, IV, or V) sizes. The basis for selecting these specifications is the need to obtain a more precise assessment of each macular area.

FIGURE 1: A, The appearance of microperimeter MP-3 (NIDEK). B, 32 points of examination distributed within the Early Treatment Diabetic Retinopathy Study grid. The program is designed for macular diseases. C, 68 points of examination is plotted symmetrically, similar to that of the Humphrey Field Analyzer 10-2 program. D, 45 points of examination plotted radially contrary to the symmetrical pattern of C. ETDRS indicates Early Treatment Diabetic Retinopathy Study.

The MAIA is an MP instrument that uses a scanning laser ophthalmoscope (SLO) to generate the retinal image for static perimetry. The device is nonmydriatic, requiring a minimum pupil size of 2.5 mm. The standard MAIA examination covers a 10-degree diameter area with 37 stimuli measurement points distributed in 3 concentric circles of 2 degrees, 6 degrees, and 10 degrees of diameter. The significant advantages of the MAIA microperimeter are its higher maximum target intensity and increased dynamic range of stimuli. However, the MAIA has limited customizable options, and only 2 fixation targets and 5 grid patterns are available. Light stimuli of the light emitting diode (LED) are Goldmann III size and projected directly onto the retinal surface at 200 milliseconds (ms). Background luminance ranges from 4 apostilbs (asb) (1.27 cd/m²) to a maximum luminance of 1000 asb (318.47 cd/m²). At the end of the examination, a sensitivity grid map and the average threshold are reported using a decibel scale distributed in attenuation ranges of 1 dB from 0 to 36 dB. Decibels are classified as normal (over 27 dB), suspect (between 26 and 27 dB), or anomalous (less than 26 dB).

The macular integrity index describes the likelihood that a patient’s responses are normal, suspect, or abnormal when compared with age-adjusted normative data. Greater numbers suggest a higher likelihood of abnormal findings, while lower numbers mean more likelihood of normal findings.

For evaluation of fixation, the MAIA microperimeter uses high-speed eye trackers (25 times/second) and plots the resulting distribution over the SLO image. The percentage of fixation points inside a circle of 1 degree and 2 degrees of diameter are used to calculate the fixation indexes P1 and P2, respectively. Fixation stability is classified as “stable” if more than 75% of the fixation points are located within P1, “relatively stable” if less than 75% of fixation points are located within P1 but more than 75% of the fixation points are located within P2, and “unstable” if less than 75% are located within P2. A more accurate estimation of the fixation pattern is provided by using the bivariate contour ellipse area (BCEA). The area of an ellipse, including 95% (BCEA95) and 63% (BCEA63) of points, is calculated, and the corresponding degree of the area is revealed. Lower BCEA values suggest greater fixation stability.

The MAIA has been shown to have a robust coefficient of repeatability, reliability, and intersession agreement.^11,12 Hence, MAIA has been used for investigating and monitoring paracentral scotoma in retinal clinical trials and in routine clinical care.^13–18 Currently, the MAIA result analysis printout shows the measured retinal sensitivity at each test locus, average sensitivity, macular integrity, and fixation stability. MAIA has a follow-up examination function that accurately remeasures the same anatomic locations and baseline test. This allows precise functional monitoring, even in cases where the retina morphology has changed due to pathology progression.

The characteristics MP-3 and MAIA are summarized in Table 1. In MP-3, the size of stimuli can be chosen from Goldman I to IV. It can possibly detect minimum light perception with relatively high maximum luminance (10000 asb) and covers broader posterior fundus (Φ 45 degree). MAIA has a scotopic test mode, which can distinguish rod or cone responses with cyan and red stimuli. Both instruments of the new generation have a good indication for assessment in any macular diseases. In addition, MP-3 is universally suited for examining posterior fundus diseases, while MAIA has specifically useful in rod and cone degeneration.

TABLE 1 - Characterictics of Retinal Microperimetry Instruments

Microperimetric Instruments	MP-3 (NIDEK)	MAIA (iCare)
Max. field of test	Φ 45 degrees (round)	36×36 degrees (square)
Min. pupil size	4 mm	2.5 mm
Size of stimuli	Goldmann I-V	Goldmann III
Scotopic test	none	cyan and red stimuli
Max. luminance	10000 asb	1000 asb
Threshold range	0-34 dB	0-36 dB
Follow-up measurement	Yes	Yes

Microperimetry Applications to Neovascular Age-related Macular Degeneration and Geographic Atrophy

Neovascular AMD (nAMD; also referred to as wet or exudative AMD) is a major cause of irreversible visual impairment in the elderly worldwide.^19,20 Geographic atrophy (GA) and macular neovascularization (MNV) are the late stages of nAMD.²¹ Anti-VEGF therapy has been conclusively shown to improve the vision in eyes with MNV. However, there is a subset of cases of nAMD that has a suboptimal response to anti-VEGF therapy. In addition, the administration of anti-VEGF therapy requires frequent follow-up visits to achieve the best outcomes. This then imposes a significant burden on the patients. In addition, patients who have a good anatomic response to the anti-VEGF therapy can still have a reduction of their vision in the long term due to the development of macular atrophy.²²

The visual acuity has been the traditional visual parameter used as the primary endpoint in many ophthalmic clinical trials, but visual acuity is primarily derived from a relatively small number of foveal cones and may not be adequate for characterizing the progression and functional impact of AMD, GA, and fovea-sparing GA.²³ Thus, MP, which examines the retinal sensitivity over a large area of the macula, has been proposed as a better technique for determining and quantifying the visual deficit in patients at the different stages of AMD. The MP technology has continued to advance with features such as faster eye tracking, which has improved the reliability of the MP measurements. The ability to monitor the fundus image by the tester during the examination has allowed for better placement of the stimuli on the retinal regions of interest. This localization is particularly important when studying the functional impact of specific anatomic biomarkers of interest, which are associated with a higher risk of progression to late AMD. These markers include intraretinal hyperreflective foci, drusen with hyporeflective cores, subretinal drusenoid deposits, and incomplete retinal pigment epithelial and outer retinal atrophy (iRORA).

MP has provided important information on the functional changes in the early and late stages of AMD and how they progress over time. As such, the MP can be a valuable tool to determine the risk of stratification and progression by monitoring the changes in eyes with AMD. In addition, the MP findings are commonly used as secondary endpoints in clinical trials in eyes with GA.²⁴ In the following sections, we shall summarize some of the major applications of MP in assessing non-neovascular and neovascular AMD.

Non-neovascular AMD and Geographic Atrophy (GA)

The enlargement of the GA in anatomic imaging studies has been considered an acceptable regulatory endpoint for therapeutic trials by the Food and Drug Administration of the United States. Monitoring the functional impact of this enlargement, especially by assessing the margins of the atrophy, has proven to be an important application of MP, especially because many regulatory bodies require corroborating functional assays in therapeutic trials. Meleth et al²⁵ demonstrated that the mean number of scotomatous points, which are defined as nonresponding loci to the highest stimulation, increased significantly (P=0.004) at a rate of 4.4 points/year in eyes with GA followed for over 24 months. The mean retinal sensitivities of all points decreased over time (P<0.003; Fig. 2).²⁵ In the Chroma and Spectri phase 3 clinical trials, the retinal sensitivities determined by MP were used as key secondary endpoints, with the number of absolute scotomatous points (no responses to the highest stimulus) and mean retinal sensitivities as the key outcome measures.^24,26 Among the various functional tests, MP had the highest correlation with the size of the GA area at both the baseline and at the 48-week endpoint.²⁴

FIGURE 2: Evolution of functional scotoma during the progression of geographic atrophy (GA) evaluated by microperimetry. Interpolated maps showing the functional scotoma progression in a representative GA eye at baseline (A), 12 months (B), and 24 months (C). Graph (D) represents the mean number of scotomatous points for all study eyes, which shows a steady increase over the 24 months of follow-up.²⁵ Copyright [The Association for Research in Vision and Ophthalmology, Inc, 2011 (Marryland, US)]. All permission requests for this image should be made to the copyright holder.

For MP to be an important tool in clinical trials and practice, its repeatability needs to be known. Alibhai et al²⁷ evaluated the test-retest variation of MP in patients with GA, and they proposed that a threshold change of±4 dB in the point-wise sensitivity measurements was a reasonable cut-off value to determine whether the true progression has occurred. Earlier studies have also shown a high degree of reproducibility in both clustered and full-field sensitivity measurements using the NIDEK MP-1.^28,29 The newer MP devices, such as the CenterVue MAIA and NIDEK MP-3, have a broader dynamic range of stimuli with faster and more accurate eye tracking, and the increased automation may further improve the repeatability.³⁰

MP has also been used to assess eyes with non-neovascular AMD before the development of GA. Thus, Corradetti et al³¹ demonstrated that AMD eyes at the intermediate stage were more likely to develop an iRORA over a 24-month period. They also reported that the scotopic mean sensitivity may have a greater impairment than the mesopic sensitivities at the earlier stages of AMD. These findings indicate that rod dysfunction can precede that of cones.³² Nassisi et al³³ found a significant correlation between the MP sensitivity and the severity of choriocapillaris flow deficits in eyes with early and intermediate AMD. This suggested that the relationship between these functional and anatomic biomarkers at these earlier stages of AMD is important.

Neovascular AMD

MP has also been used in studies of nAMD. An important predictor for the visual outcomes after anti-VEGF treatment for MNV is the interval between the onset of symptoms and the initiation of the treatment.^34,35 This has led to the assumption that there may be early changes before the development of exudation, that is, intraretinal fluid and/or subretinal fluid, that if identified could yield earlier treatments and better long-term visual outcomes.³⁶ Given that the retinal sensitivity may be a useful parameter for assessing the functional status of the retina in AMD,³⁷ 1 might predict that MP may be able to detect decreased retinal sensitivity in areas that would develop subretinal fluid and/or intraretinal fluid, and subsequent visual acuity reductions. Wightman et al³⁸ reported that MP could detect localized loss of retinal sensitivities in regions with fluid without any apparent change in the best-corrected visual acuity (BCVA) or the Amsler grid findings. This would then highlight the possibility of MP in detecting these early subtle functional alterations before the development of the exudations (Fig. 3).

FIGURE 3: Microperimetry retinal sensitivity before and after the development of subretinal fluid in a 70-year-old female with AMD. A, shows the NIR image (left), representative SDOCT B-scan (center), and MAIA pseudocolor MP sensitivity map (right) of an intermediate AMD eye without evidence of exudation. MP map demonstrates a localized region (in red) of reduced sensitivity inferonasally before the development of any exudation in this region. B, shows the NIR (left), SDOCT B-scan (center) and MAIA microperimetry sensitivity map (right) after the development of macular neovascularization associated with subretinal fluid. The previously detected inferonasal region of reduced sensitivity has further enlarged.³⁸ Copyright [The Authors, 2019 (California, US)]. All permission requests for this image should be made to the copyright holder. AMD indicates age-related macular neovascularization; MP, microperimetric; NIR, near infrared reflectance; SDOCT, spectral domain optical coherence tomography.

Wu et al³⁹ also explored the ability of MP to detect the onset of atrophy and exudative MNV in the Laser Intervention in Early Stages of AMD (LEAD) study. In this trial, MP was performed at the baseline and at intervals of 6 months in a cohort of eyes with bilateral large drusen. The results showed that MP was the best method of detecting the onset of nAMD and/or atrophic changes. However, the sensitivity was generally low, ranging between 11% and 35%, with a specificity of 95%. The low sensitivity made the authors conclude that MP on its own was not useful as a self-monitoring screening tool to detect the early onset of late AMD.

Other studies have evaluated the role of MP in assessing the response of eyes to anti-VEGF therapy by examining the improvement in sensitivity and resolution of exudation after the treatments.^40,41 Prager et al⁴¹ evaluated the changes in the retinal sensitivities in eyes with MNV treated with intravitreal bevacizumab (Avastin; Genentech, Inc, South San Francisco, CA) therapy, and they reported that the mean absolute scotoma size decreased from 33% to 22% at month 3 and to 23% at month 6. In addition, the mean differential light threshold significantly increased from 3.8 dB at the baseline to 5.5 dB at 6 months. Also, Asahi et al⁴² determined the retinal sensitivity by MP in a randomized phase 2 open-label study, in which they compared the pro re nata regimen to the treat-and-extend regimen or the monthly regimen with ranibizumab. For all patients, the average retinal sensitivity significantly improved by 1.7±0.5 dB (P<0.05).

In summary, MP has established itself as a useful tool in the monitoring of eyes with manifest GA. The role of MP in assessing the earlier stages of retinal disease, tracking the progression to the late stage of AMD, and assessing the function of eyes with MNV continues to evolve and should be better in the future, especially with large-cohort longitudinal studies.

Microperimetry in Stargardt Disease

MP is a key diagnostic tool in Stargardt disease (STGD) because it can detect signs of visual dysfunction before abnormalities become evident in retinal imaging.^43,44 STGD subtype 1 (STGD1) is the leading cause of inherited retinal diseases,⁴⁵ and it is caused by biallelic mutations in the ABCA4 gene. It has an autosomal recessive inheritance pattern⁴⁶ which causes an accumulation of lipofuscin in the retinal pigment epithelium⁴⁷ and consequent photoreceptor and retinal pigment epithelium atrophy.

A common quantitative metric derived from MP is the macular sensitivity, which is commonly evaluated within 3 retinal zones similar to those defined by the Early Treatment Diabetic Retinopathy Study grid: the foveal central subfield, inner ring, and outer ring. The ProgStar Study Group reported a mean baseline macular sensitivity of 10.73 dB in patients with a mean age of 33.7 at the time of the assessment,⁴⁸ and a mean decrease in the macular sensitivity of 0.87 dB/year.⁴⁹ Early in the disease course, the severity of the sensitivity loss tended to be the greatest in the foveal area, with less severe effects in the outer rings.⁴⁹ Investigations have shown that a longer duration of symptoms and older age were associated with poorer macular sensitivity,⁴⁹ with a 0.086 dB decrease per one additional year of age (P<0.001) and 0.21 dB/year of disease duration (P<0.001).⁴⁹

Another major parameter that can be assessed during MP evaluations is determining the site and steadiness of fixation. This site, also termed preferred retinal locus (PRL), is positioned at the center of the fovea in normal individuals, but in STGD, the PRL shifts more eccentrically to the superior part of the central atrophic lesion.⁵⁰ Over two-thirds of patients with more than 15 years of STGD1 develop an eccentric PRL.⁵¹ It has also been observed that for each year of a later onset of the symptoms, the PRL is located 0.14 degrees more centrally (P<0.0001).⁵⁰ The PRL moves 0.0014 degrees eccentrically from the anatomic fovea yearly (P=0.99).⁵²

Another parameter used to characterize the fixation is the stableness of the fixation. A PRL closer to the border of the macular atrophy will be more stable than a more eccentrically located PRL.⁵³ Despite the better stability of PRLs closer to the center, the macular sensitivity and the BCVA are poorer in these areas,⁵³ with a 1 degree of PRL eccentricity associated with a 2 to 3 letter loss (P<0.0001).⁵¹ The BCEA, which is an ellipsoidal region surrounding 1 to 3 SD of all fixation points,^54,55 is a quantitative metric of fixation stability.⁵⁴ The larger the BCEA, the more unstable the fixation. The ProgStar Study showed that the more eccentric PRLs were associated with larger BCEAs although the difference was not significant (P=0.14).⁵² This study also showed 1.21 square degrees of deterioration in the fixation stability after 1 year of follow-up (P=0.33).⁵² Some MP devices are equipped with a rehabilitation mode to perform PRL training, which helps patients achieve more stable fixation with consequent smaller BCEAs.^56–59

The Scotopic Microperimetric Assessment of Rod Function in Stargardt Disease (SMART) Study compared the MP findings under scotopic and mesopic conditions.^60,61 They found that the mean deviation (MD) from the normal macular sensitivity for the mesopic foveal, mesopic extrafoveal, and scotopic extrafoveal were 0.72, 0.86, and 1.12 dB/year, respectively.⁶¹ These findings suggested that while both rod and cone functions appeared to decline, the degeneration of the rods may be more prominent than the cone degeneration in some STGD1 patients.⁶¹

Different analysis algorithms have been applied to the MP data, and additional biomarkers to track the progression of STGD have been found. A higher number of deep or absolute scotomatous points where the brightest light stimuli were not perceived were associated with a longer duration of the disease,⁴⁸ with a mean increase of 1.56 points/year.⁶² The percentage of deep scotomatous points was highest in the fovea and lowest in the outer ring.⁴⁹ The number of deep scotomatous points is an indication of the extent of the functional impairment, and the rate of deep scotoma extension might be correlated with the rate of decrease in the fundus autofluorescence.^62,63 Instead of using a standardized grid, another more sensitive approach to detect disease progression is to evaluate the locations at the border of the atrophic lesions,⁶⁴ as it has been found that the rate of progression on the edge of the dense scotoma (2.9 dB/y) is faster than the decrease in the overall mean sensitivity (1.5 dB/y; P<0.001; Fig. 4).⁶⁴

FIGURE 4: Microperimetric assessment of a patient with Stargardt disease subtype 1 comparing the overall mean sensitivity with sensitivity at the edge of the deep scotoma over 12 months. The dense scotoma points are overlaid with red, and the yellow overlay includes the points at the edge of the deep scotomas. While mean sensitivity (left bottom values in green) did not progress through the follow-up period, the sensitivity at the edge of the deep scotoma (left bottom values in yellow) declined by 2.6 dB after 12 months.⁶⁴ Copyright [Elsevier Inc (Amsterdam Airport Schiphol, NL)]. All permission requests for this image should be made to the copyright holder.

Volumetric indexes were also found to reflect the retinal sensitivity more accurately⁶⁵ as they are not affected by confounding factors such as the MP instrument or test locations.⁶⁶ An advanced volumetric model termed the Visual Field Modeling and Analysis technique measures the total amount (VTOT) of sensitivity (hill of vision) within a visual field in the region of interest (Fig. 5).^48,65,66 In the ProgStar⁶ and Tanna et al⁶⁴ studies, the baseline VTOT in adults was 0.91 and 0.52 dB-sr with an annual decline of 0.077 to 0.06 dB-sr, respectively. In children, these values were a baseline VTOT of 0.87 dB and an annual decline of 0.21 dB-sr. These findings suggested a faster rate of disease progression in children.⁶⁵ A faster rate of progression was observed with a higher baseline VTOT and during the second year of follow-up.⁴⁸

FIGURE 5: Volumetric functional assessment and structural imaging in a 41-year-old patient with Stargardt disease subtype 1. A–E, shows the progression in total volume (VTOT) under the interpolated hill of vision model generated by Visual Function Modeling and Analysis over 2 years. While the patient had roughly stable visual acuity (40–48 letters) through 24 months of follow-up, the VTOT declined by 0.44 dB-sr. F, Fundus autofluorescence of the same eye at baseline demonstrates central decreased autofluorescence with surrounding spots of increased autofluorescence signal. G, Spectral domain optical coherence tomography of the fovea at baseline exhibits outer retinal atrophy and retinal thinning in the area of visual impairment.⁴⁸ Copyright [Elsevier Inc (Amsterdam Airport Schiphol, NL)]. All permission requests for this image should be made to the copyright holder.

FIGURE 6: Structure-Function correlation in the left eye of a 56-year-old man with diabetic macular ischemia (A–D, Angioplex, Zeiss; E, Spectralis Heidelberg, Germany; F, MAIA, CenterVue, Italy). This eye has been treated for proliferative diabetic retinopathy with panretinal photocoagulation. There was no history of macular edema. Visual acuity was 20/30. SCP (A) and DCP (B) showed enlarged and irregular foveal avascular zone with areas of perifoveal capillary loss. In (C), the SCP is shown in red and is superimposed over the DCP, which is represented in green. Areas with perifocal capillary loss are highlighted with *. In (D), the retinal thickness map is superimposed over the SCP and DCP flow map. Areas with perifoveal capillary loss highlighted in (C) show reduced retinal thickness and appeared as blue areas. The structural optical coherence tomography through the fovea is shown in (E). The microperimetry (F) showed stable fixation, but some areas with reduced retinal sensitivity (yellow and orange dots) were detected, which correlates with the areas of retinal thinning and ischemia. DCP indicates deep capillary plexus; SCP, superficial capillary plexus.

MP is not without its limitations. Testing can be time-consuming, and the test results can be impacted by fixation stability.⁶⁷ MP testing also requires good cooperation by the patient, which may be challenging in patients with reduced vision, unstable fixation, and in young children. However, it has been shown that reliable MP values can be obtained from children aged 8 years or above (Fig. 6).⁶⁵ Also, due to the slow progression of the functional impairments, MP might not detect disease progression in short-term studies, which is a limitation of clinical trials.⁶⁸

Microperimetry in Diabetic Retinopathy

Diabetic retinopathy (DR) is 1 of the leading causes of visual impairment and blindness. The pathogenesis of DR is complex, but key components include the loss of pericytes, microvasculopathy, neurodegeneration, and inflammation. The most frequent complications leading to vision loss include proliferative DR and DME. In addition, ischemia and damage to the retinal microstructures, such as the disruption of the ellipsoid zone and disorganization of the inner retinal layers, have been associated with different degrees of reduction of the visual acuity. However, the visual acuity tests may not be sensitive enough to localize areas of functional loss. Several functional tests, such as MP, multifocal electroretinography, perimetry, and low-luminance visual acuity, have been evaluated to complement the conventional visual acuity to characterize the quality of vision resulting from the changes.⁶⁹ In addition to determining the sensitivities, MP can also determine the site and stability of fixation. MP findings can be analyzed to determine if they are significantly correlated with structural changes determined by OCT and angiography.

Unoki et al⁷⁰ reported that the areas of capillary nonperfusion determined by fluorescein angiography were associated with the areas of loss of retinal sensitivity determined by MP. These areas also had thin and disorganized inner retinal layers and high-reflectivity deposits between the photoreceptor outer segments and the retinal pigment epithelium. Many subsequent studies confirmed these findings and further refined the structural and functional correlations in patients with DR.^71,72 Montesano and colleagues further evaluated diabetic individuals without DR and reported a thinning of the ganglion cell layer and the inner plexiform layer. They also found a significant correlation between the MP findings and the ganglion cell count.⁷³ These important findings suggested that early neuronal loss occurs in diabetes, even in individuals without DR. Similar structural and functional correlations were also reported in patients with various severities of DR.⁷⁴

The advantage of MP in correlating the fundus findings to retinal function can be further exploited by comparing the findings with the OCTA findings. In addition to assessing the global function, zonal and local correlations could also be evaluated in detail, thus offering point-to-point structural-functional correlations. A number of studies have reported reduced retinal sensitivity in eyes with diabetic macular ischemia (DMI), which is characterized by reduced vessel density and/or enlarged or disrupted foveal avascular zone.^75–78 Importantly, eyes with DMI may have normal visual acuity despite reduced retinal sensitivity suggesting that MP is more sensitive than the visual acuity, or the changes in the sensitivities may precede the reduction of the visual acuity in DMI.

There have also been inconsistencies in the correlations between the OCTA findings and retinal sensitivities. While some studies reported a significant correlation between the enlarged size of the foveal avascular zone and the reduced retinal sensitivity, others have only found a correlation between vessel density and retinal sensitivity. There is also a discrepancy in whether the vessel density in the superficial capillary plexus and deep capillary plexus are correlated more highly with the retinal sensitivity.⁷⁹ Differences in the study population and the instruments used may explain some of these inconsistent findings. There are also reports on the inconsistency of the spatial co-localization of areas of focal retinal sensitivity and capillary nonperfusion. Levine and colleagues performed a detailed point-by-point co-localization to evaluate the correlations between the focal retinal sensitivities and the microvascular changes. They reported that significant correlations were only present in eyes with proliferative diabetic retinopathy.⁸⁰ Interestingly, Mokrane et al⁸¹ reported that areas with retinal sensitivity less than 25 dB were always associated with capillary loss, but some areas of capillary loss had a preserved retinal sensitivity.

Altered fixation and reduced retinal sensitivity have been reported in eyes with DME, and the retinal sensitivity was inversely correlated with the retinal thickness. In addition, following treatment with anti-VEGF agents or intravitreal corticosteroids, improvements in retinal sensitivities have been reported.^82–84 The retinal sensitivity improvement was significantly correlated with the BCVA and inversely correlated with the OCT central subfield thickness.^85–88 In addition, Santos and colleagues reported that better retinal sensitivities before treatment were associated with better BCVA outcomes after therapy. This suggested that the baseline MP findings may have prognostic value for DME treatments.⁸⁹ Yohannan et al⁹⁰ further demonstrated that disruption of the ellipsoid zone is correlated with a significant decrease in the point sensitivity in eyes with DME.

While MP offers many advantages which complement other functional assessments, its limitations should also be considered. Interdevice differences should be considered since different instruments and settings are used by different groups,^91,92 The reported retinal sensitivities are further affected by the luminance and contrast sensitivity. There is no standardized testing protocol, for example, the number of points tested, the area covered, and the lighting condition used when conducting the test. In addition, normative data for retinal sensitivities have not been established.

Even though the value of the MP findings in understanding DR has been widely accepted, and it is less complex and easier to conduct than multifocal electroretinograms and its use in the real clinic as a routine examination has not occurred. However, MP takes more than 5 minutes to determine the sensitivities of 40 points within the central 10-degree field and requires some concentration by the patients. To overcome these problems, authors (S.H. and K.O.M.) have sought for a more practical protocol for using MP-3 in the clinic. The number of examining points and areas can also be limited as examining points can be designed and conducted freely by MP-3. We have used a self-designed program of only 9 points within 5 degrees of the central macular field in eyes with DME for an easy routine examination. It required less than 3 minutes for an examination of each eye with good compliance by the patients. Although in cases of poor visual acuity in relatively advanced cases (Fig. 7), the sensitivities of the macular area were not uniform even in the 5-degree field. The results indicated that more detailed functional assessments of the macular area other than visual acuity are possible for routine assessments with this easy method. Similarly, in Tsai et al’s study,^75,76 the MP protocol was customed to capture the central 1500×1500 µm where the perifoveal capillary loss occurs with 13-Goldman III size stimuli.

FIGURE 7: A, The left eye of a 62-year-old female of type II diabetes. Retinal sensitivity obtained by MP-3 (NIDEK) is displayed on the color fundus image. The central sensitivity of fovea is 0 dB and 8 points of 5 degrees from the central fovea are 6 dB, 18 dB, 9 dB, 5 dB, 5 dB, 9 dB, 20 dB, or 2 dB clockwise. LogMAR is 0.7. The net time of the examination is 67 seconds. B, MP-3 of 4 weeks after anti-vascular endothelial growth factors injection (A). The central sensitivity of fovea is 0 dB and 8 points of 5degrees from the central fovea are 17 dB, 0 dB, 13 dB, 25 dB, 23 dB, 27 dB, 15 dB, or 9 dB clockwise. LogMAR is 0.4. The net time of the examination is 48 seconds. C, Corresponding color scanning laser ophthalmoscope image of A (Mirante, NIDEK). D, Corresponding optical coherence tomography image of A with intraretinal fluid, hard exudates, and serous retinal detachment.

Microperimetry in Surgical Macular Diseases

Macular diseases, such as epiretinal membrane, macular hole, vitreomacular traction, and retinal detachment, are major causes of irreversible blindness which require surgical management. For many of these diseases, the foveal center is typically spared in the initial stages, but as these diseases progress, the lesion spreads into the fovea to cause central vision loss. Treatment at this early stage can prevent permanent loss of the high-quality foveal vision that is critical for daily tasks. Patients with permanent damage at the foveal center develop a PRL by using a peripheral retinal area for eccentric viewing. As a result, visual acuity served by these PRLs may not be an acute representation of disease progression. Yet, most clinicians rely on visual acuity when monitoring these macular diseases to determine visual function, as common presenting symptoms such as visual distortion or metamorphopsia cannot be objectively quantified. In these cases, MP can be used as an additional investigational tool in identifying perifoveal retinal dysfunction, given that it can quantitatively examine macula function in high detail.

Epiretinal Membrane/Internal Limiting Membrane

Macular epiretinal membrane (ERM) is a relatively common disorder, affecting up to 12% of individuals older than 70 years.⁹³ Patients with ERM may report several visual symptoms, including metamorphopsia, decreased visual acuity, and monocular diplopia. Pars plana vitrectomy (PPV) is the gold standard procedure used to surgically remove ERM in symptomatic patients.^94,95

The typical clinical assessment of patients with ERM involves functional exams such as visual acuity and quality of life measures alongside OCT. MP may complement functional measures to support surgeons in patient selection and evaluating prognosis, especially in cases when a patient’s BCVA is preserved. Moreover, its auto-tracking system and ability to perform follow-up examinations using the same anatomic reference points allow greater reliability and reproducibility of test results than automated perimetry.^96,97 An example of the MAIA parameter of a patient with ERM is demonstrated in Figure 8.

FIGURE 8: Standard sensitivity output generated from the MAIA system of an eye with epiretinal membrane. The standard MAIA examination covers a 10-degree diameter area with 37 measurement points. Traction lines are visible, and a relatively diffuse sensitivity reduction is demonstrated in the sensitivity map and sensitivity plot (top right). Macular integrity and average threshold values are also reported. Threshold frequencies are compared with a histogram of reference database values on a decibel scale that is color-coded according to MAIA normative studies, where “green” represents normal values, “yellow” suspect, “red” abnormal, and “black” represents scotoma. The fixation plot demonstrated centrally located preferred retinal locus with relatively unstable fixation.

A recent study including 30 eyes with ERM investigated the potential recovery of retinal function by measuring retinal sensitivity with MP-1 microperimetry.⁹⁸ All patients went under 25-gage PPV with ERM and internal limiting membrane (ILM) peel and phacoemulsification of the lens with intraocular lens insertion. Consistent with previous studies,^99–104 there was a significant early reduction in central retinal thickness and improvement in BCVA, which peaked at 1 year postoperatively. Meanwhile, mean retinal sensitivity and retinal defects in the central 10-degree area continuously improved from baseline to 1 and 4 years, indicating that underlying functional improvements can occur into the longer term. A similar example of MAIA parameters of pre-ERM and post-ERM surgery is demonstrated in Figure 9.

FIGURE 9: A, Preoperative and postoperative OCT and MAIA in a patient with epiretinal membrane. B, A preoperative and postoperative OCT and MAIA in a patient with myopic schisis. C, Preocriplasmin and postocriplasmin OCT and MAIA in a patient with vitreomacular traction. OCT indicates optical coherence tomography. OCT indicates optical coherence tomography.

The recurrence rate of ERM after peeling is estimated to be between 5% and 10%.¹⁰⁵ This may be due to remnants of cells and collagen fibers over the ILM for incomplete release of the tangential tractions over the retina. As a result, some surgeons opt to also peel the ILM during PPV. However, the value of ILM peeling in ERM surgery remains a controversial procedure because its effect on visual improvement has not been proven, and its potential to result in mechanical and functional damage to the Muller cells and inner retinal layers.^106–108 In a randomized clinical trial study, Ripandelli et al¹⁰⁹ compared microperimetric responses in 30 eyes that underwent ILM peeling against 30 eyes that did not undergo ILM peeling and found that by 12-month follow-up, eyes that did not receive ILM peeling showed a greater and faster recovery of mean retinal sensitivity within the central 4-degree area. Conversely, eyes that had ILM peel had a significantly larger number of absolute microscotomas (0 dB) within the 12-degree central retinal area. In a second randomized clinical trial study, Russo et al¹¹⁰ investigated the functional and anatomic results of complete ILM peel versus foveal-sparing ILM peel. The mean retinal sensitivity was significantly improved in foveal-spared eyes at 12 months after surgery, while a nonsignificant reduction in mean retinal sensitivity was observed in the complete ILM peel group. It is important to note that 15% of the foveal-sparing group developed ERM recurrence with significant metamorphopsia and reduction in visual acuity requiring revision surgery. The potential that retinal function may be compromised should be considered against the risk of ERM occurrence when deciding whether to peel the ILM.

Overall, MP appears to be an effective additional tool in evaluating subtle changes in retinal function, undetectable with a visual acuity exam, even in long-term follow-up. However, the literature remains limited at this point in time. Assessment of a large prospective cohort of eyes that undergo ERM peeling surgery will characterize the longitudinal microperimetric response of the macular. Larger-scale studies may contribute to building a grading or classification system for ERM and macular impairment for use in clinical practice.

Macular Hole

A full-thickness macular hole (MH) is a retinal break commonly involving the fovea. Surgery involving PPV with a gas tamponade agent, with or without ERM/ILM peel, is often required to close the hole. Traditionally, clinical examination and investigations, including BCVA and OCT, have been used in the diagnosis and management of this condition. The increasing availability of microperimetry in clinical practice has allowed for more reliable and objective examination of functional retinal sensitivity in patients with MH. Significant improvements in microperimetric macular sensitivity have been observed postoperatively in patients with full-thickness macular holes, who can present with centralized scotomas.¹¹¹ The absence of absolute scotomas in the fovea postoperatively demonstrates the persistence of foveal function and encourages vitreoretinal surgery.

Wang and colleagues have examined the predictive value of preoperative macular sensitivity testing in predicting visual prognosis in a study of 44 eyes using the MP-3 microperimeter. A positive linear correlation was found between preoperative macular sensitivity (mean value of inner and outer 0.5 degrees of MH margin) and postoperative BCVA at 4 months. Furthermore, Richter-Mueksch et al¹¹² showed that a significantly higher number of patients demonstrated improvement in retinal sensitivity, measured with MP-1, than improvement in BCVA at 12 weeks after macular surgery in patients with MH and macular pucker. MP also has a role in macular rehabilitation training. For instance, MAIA’s biofeedback training employs auditory and visual biofeedback signals to train low-vision patients with central scotoma and unstable fixation. The training has been found to improve visual acuity in patients with the insufficient recovery of BCVA after successful MH surgery.¹¹³

MP in the MH is useful in demonstrating absolute scotoma within the hole. Preoperative MH sensitivity might be a reliable and sensitive predictor factor for visual prognosis after successful surgery. MP is also a valuable functional monitoring tool, as BCVA may underestimate the functional benefit of surgical intervention for focal vitreomacular pathology.

Vitreomacular Traction

Vitreomacular traction (VMT) is an incomplete posterior vitreous detachment with macular adhesion. VMT can cause various anatomic distortions¹¹⁴ resulting from pseudocysts, macular schisis, cystoid macular edema, subretinal fluid,¹¹⁵ and ultimately MH formation. These can result in decreased vision, central vision loss, and metamorphopsia. Treatment options include observation, surgery, and pharmacologic vitreolysis.

Intravitreal injection of ocriplasmin, a recombinant, truncated form of human plasmin, has demonstrated efficacy in resolving issues from vitreomacular adhesion (VMA) in patients with VMT.^116–118Figure 9 demonstrates postocriplasmin anatomic and functional improvement in a patient with VMT. While clinically significant improvements in BCVA have been observed in patients with VMA resolution after ocriplasmin treatment,^119,120 visual improvement may be difficult to quantify in patients with metamorphopsia and relatively preserved BCVA due to highly focal foveal disruption in VMT. In these cases, associations between retinal function and treatment outcomes have been found, suggesting that MP can be useful in assessing baseline retinal sensitivity and functional change over time. A longitudinal cohort study by Cacciamani et al¹²¹ examined 16 eyes with symptomatic VMT using the MP-1 microperimeter to investigate whether improvements in macular function were associated with anatomic resolution after ocriplasmin treatment (125 µg intravitreal injection). Significant increases in mean retinal sensitivity (within central 4 degrees) of ~3.7 dB were seen in those with resolved VMA following treatment, whereas no measurable changes in retinal function were seen in those with unresolved VMA. No significant changes in BCVA were seen in all subjects, including those with complete VMA. In a similar interventional study, which examined 21 eyes with VMT after being treated with ocriplasmin injection, found that significant early improvements in retinal sensitivity could be detected using the MP-1 microperimeter at 3 months postoperative, but not using BCVA.¹²²

Although ocriplasmin has been shown to be effective in achieving good anatomic outcomes, the decision to treat with vitrectomy, enzymatic vitreolysis, or observation remains difficult. Apart from a small case series, no detailed studies have evaluated the functional response and predictive role of MP in a surgical cohort in eyes with VMT. However, given that retinal and macular sensitivity seems to better quantify visual dysfunction in patients with VMA/VMT compared with BCVA after ocriplasmin treatment, MP is also likely to be useful in measuring functional outcomes in patients undergoing PPV. MP may also play a predictive role in guiding clinical decision making, such as whether to perform an early vitrectomy; however, this needs to be examined in a future study.

Retinal Detachment

Visual recovery after rhegmatogenous retinal detachment (RRD) repair is often incomplete. MP can be a valuable tool in assessing patients with impaired vision recovery after retinal detachment repair. Rossetti et al¹²³ reported microperimetric outcomes using the MP-1 in 6 patients at 1, 3, 6, 12, 18, and 60+ months after scleral buckle for macula-off RRD. They found an increase in macular sensitivity over the first 6 months, followed by gradual improvement up to 18 months. But by 5 years, there was a nonsignificant reduction in overall mean sensitivity. Several studies have linked these functional changes to anatomic features using MP and OCT imaging to determine the mechanism of incomplete retinal function recovery following the repair of retinal detachment. Smith et al¹²⁴ assessed functional outcomes using the MP-1 microperimeter in 7 eyes after macula-off RRD repair. Although 2 eyes had retinal sensitivity within the normal range, 5 had reduced sensitivity in regions of the macula that correlated with disruption of the ellipsoid zone or persistent subretinal fluid. Lai et al¹²⁵ reported the MP-1 outcome of 12 patients (5 macula-on and 7 macula-off RRD) after a successful repair at least 6 months prior. Eighty-six percent of eyes with macula-on detachment had a mean macular sensitivity of greater than 15 dB with no inner or outer retinal abnormalities on OCT and normal fundus autofluorescence signal in the macula. In contrast, 71% of eyes with macula-off detachment had lower sensitivity values ranging from 0.9 to 10.8 dB with abnormal OCT and autofluorescence features. Delolme et al¹²⁶ examined the correlation between OCT and retinal sensitivity using the OptosOCT SLO microperimeter in 30 patients with successful macula-off RRD repair 6 months prior. They reported eyes with more than 1 lesion on OCT had significantly lower foveal and macular retinal sensitivity. Sensitivity in the central 12 degrees was correlated to the presence of ellipsoid zone defect, while sensitivity in the central 4 degrees was associated with thinner photoreceptor outer thickness.

Overall, the correlation of functional mapping with OCT features has identified photoreceptor cell damage to be the main reason for the incomplete recovery of retinal sensitivity. Assessment of visual distortion and retinal sensitivity in the central visual field using a fundus-tracking microperimeter can provide additional invaluable information to clarify the diagnoses and allow detailed monitoring of disease progression and treatment responses in patients with surgical macular diseases. However, further longitudinal studies are required to correlate retinal sublayer thickness and reflectivity maps with MP to identify predictive imaging biomarkers for poor-vision recovery with the aim of directing patients toward other appropriate treatments.

Retinoschisis and Myopic Traction Maculopathy

Myopic traction maculopathy (MTM) is a macular complication associated with pathologic myopia. MTM is estimated to affect between 9%¹²⁷ to 34%¹²⁸ of highly myopic eyes with a posterior staphyloma. Eyes with MTM include those with vitreomacular traction, retinal thickening, macular retinoschisis (MRS), lamellar macular hole, and foveal retinal detachment (Fig. 10).^129,130 The diagnosis of MTM is based on the OCT findings.¹³¹ Patients may complain of a worsening of visual function and increased metamorphopsia with decreased visual acuity. The inflexibility of the ILM is considered to be an important cause of maculopathy and ILM peeling is the surgical procedure for treating MTM.

FIGURE 10: Optical coherence tomography images of 4 types of myopic reaction maculopathy. A, Macular retinoschisis. B, Foveal retinal detachment. C, Macula hole. D, Macula hole retinal detachment.

Studies have reported MP as a useful tool in evaluating patients with retinoschisis and MTM to guide the decision regarding surgical intervention. Ripandelli et al¹³² observed 214 eyes with high myopia and posterior staphyloma. The study investigated conditions such as macular retinal schisis and partial/full-thickness MH. Decreased fixation stability and foveal sensitivity correlated to the need for surgery, while baseline foveal sensitivity and fixation did not. Baptista and colleagues reported MP outcomes using MP-3 in 50 eyes with high myopes and MTM; MP sensitivity was significantly reduced in eyes with atrophic areas compared with the eyes with schisis alone. The study concluded MP analysis as a valuable part of the multimodal anatomic and functional assessment for more precise characterization in patients with myopic maculopathy.¹³³

The BCVA and OCT findings are mainly used to assess the changes in visual function and structural outcomes after surgery of MTM (Figs. 9, 11). However, paracentral retinal sensitivity is also important for the quality of vision. Shinohara et al¹³⁴ reported on the retinal sensitivity of eyes with a fovea-sparing ILM peeling for MRS using MP. The results showed that the BCVA and retinal sensitivity at 2 degrees were significantly better than the preoperative values. However, some cases developed postoperative microscotomas at the paracentral 2 degrees and/or 6 degrees despite the absence of foveal retinal detachment or MRS.

FIGURE 11: Preoperative and postoperative OCT images and MP-3s in a patient with myopic traction maculopathy. A, Preoperative OCT. B, Preoperative MP-3. C, Postoperative OCT. D, Postoperative MP-3. OCT indicates optical coherence tomography.

Earlier studies reported that ILM peeling can cause a decrease in retinal sensitivity.^135,136 This may be true, especially in highly myopic eyes, because the retina is more vulnerable than that of non-highly myopic eyes.

The value of MP in the assessment of visual function before and after surgery is that it can detect the changes in the paracentral retinal function, including postoperative microscotomas, which do not affect the BCVA.

CONCLUSIONS

The importance of retinal MP has been widely accepted. All serious macular disorders need more precise functional assessments other than visual acuity testing. MP together with modern imaging techniques, can provide many physiological insights into pathologic macular conditions. The demand for careful assessments and observations has emerged with the introduction of advanced treatment of macular diseases, such as anti-VEGF treatment or recent modern macular surgical techniques. Recent commercially available MP instruments have provided a solution for the requirement. However, there are still some practical problems even though MP can be performed easier with an automated eye-tracking system. Unlike visual acuity, MP is still not routinely performed due to the time and concentration required from the examinee. To overcome this issue, the optimal program for each macular condition should be investigated in the future.