Assessment of Reading Behavior with an Infrared Eye Tracker after 360° Macular Translocation for Age-Related Macular Degeneration

Abstract

Purpose.: Macular translocation (MT360) is complex surgery used to restore reading in exudative age-related macular degeneration (AMD). MT360 involves retinal rotation and subsequent oculomotor globe counterrotation and is not without significant surgical risk. This study attempts to gauge the optimal potential of MT360 in restoring reading ability and describe the quality and extent of recovery.

Methods.: The six best outcomes were examined from a consecutive series of 23 MT360 cases. Reading behavior and fixation characteristics were examined with an infrared eye tracker. Results were compared to age-matched normal subjects and patients with untreated exudative and nonexudative AMD. Retinal sensitivity was examined with microperimetry to establish threshold visual function.

Results.: MT360 produced significant improvements in visual function over untreated disease and approximated normal function for reading speed and fixation quality. Relative to the comparative groups, eye tracking revealed the MT360 cohort generated a greater number of horizontal and vertical saccades, of longer latency and reduced velocity. In contrast, saccadic behavior when reading (forward and regressive saccades) closely matched normal function. Microperimetry revealed a reduction in the central scotoma with three patients recovering normal foveal sensitivity.

Conclusions.: Near normal reading function is recovered despite profound surgical disruption to the anatomy (retinal/oculomotor). MT360 restores foveal function sufficient to produce a single stable locus of fixation, with marked reduction of the central scotoma. Despite the limitations on saccadic function, the quality of reading saccadic behavior is maintained with good reading ability. Oculomotor surgery appears not to limit reading ability, and the results of retinal surgery approximate normal macular function.

Difficulty in reading is the commonest disability reported secondary to central visual loss from macular disease. The ability to read is a highly valued task, and its loss is linked to considerable psychological and social distress and a marked reduction in quality of life. ^{1 –6}  

Exudative age-related macular degeneration (wet AMD) is a common, rapid cause of central vision loss. Recent advances in antivascular endothelial growth factor (VEGF) treatments have offered the possibility of reversal of vision loss. In severe lesions, notably where there is significant disruption of the retinal anatomy (e.g., large subretinal hemorrhages, a large pigment epithelial detachment, or retinal pigment epithelial tears with “scrolling” of the epithelium), anti-VEGFs are ineffective, and treatments options are limited to surgical methods designed to restore normal anatomy (macular translocation or retinal pigment epithelial transplantation) or visual rehabilitation. In terms of visual recovery, a primary goal of such treatments is the restoration of reading capability. 

Conventional rehabilitation with low visual aids tries to restore reading capability by use of magnification. Despite magnification, reading speeds are often limited to <50 words per minute, well below the level required for fluent reading (>80 words per minute). ^7,8 Macular translocation surgery (MT360) has been used to restore reading vision in severe wet AMD. MT360 attempts to rescue foveal photoreceptors before irreversible retinal atrophy/damage has occurred, re-establish normal subfoveal anatomy, and provide stable, improved central vision. 

Good visual outcomes have been reported when MT360 is performed during the window of opportunity for treatment that corresponds to the period of viability of neurosensory cells. ^{9 –13} In most studies, outcome measures have concentrated on distance acuity rather than near visual function or reading speed. Of note, Eckardt et al. ⁹ and Lai et al. ¹⁰ reported improvements in near acuity after translocation; Fujikado et al. ¹⁴ specifically examined reading ability after MT360 and reported an improvement in critical print size (CPS), determined with a Japanese version of the Minnesota Acuity Chart (MN Read—J Chart); and Toth et al. ¹⁵ investigated various parameters of near visual function after translocation and reported significant improvements in near acuity, contrast sensitivity, and reading speed. 

This study was conducted to demonstrate the optimal potential of MT360 in the restoration of normal visual function by describing the quality and extent of this recovery. We observed the best-outcome cases from a series of 23 consecutive patients with subfoveal choroidal neovascular membrane (CNV) treated with MT360. MT360 is an invasive and involved procedure (new foveal location, extraocular muscle surgery, and residual torsion), and documenting the potential for restoration of close to normal function is valuable. We documented not only the well-recognized components of near visual function (reading acuity, contrast sensitivity reading speeds, and CPS) but with the use of a an infrared eye tracker and microperimetry, we examined saccadic behavior when reading (number, duration, velocity, and direction of saccades), fixation characteristics (preferred retinal loci and stability of fixation), and retinal sensitivity (size, location, and density of any postoperative scotoma), and compared these parameters to both untreated disease and normal function. 

Methods

Patients and Ethical Approval

Six patients (3 women, 3 men; age range, 61–74) recruited from a series of 23 consecutive patients (15 women, 8 men; age range, 61–94 years), with second-eye subfoveal CNV secondary to severe AMD, underwent uncomplicated MT360 with silicone oil tamponade (surgeon LDC). Six weeks later, the procedure was completed with combined counterrotation surgery (surgeon, JL) and removal of the oil (surgeon, LDC). All surgeries were performed at Moorfields Eye Hospital, London between May 2003 and January 2006. 

Central to this study was the assessment of reading behavior and fixation with an infrared eye tracker (see below). The eye tracker assessment involved performing tasks that requiring good reading ability. Recruitment from the main surgical series was based on patients who achieved improved or stable distance and reading vision, and a distance and reading acuity of logarithm of minimal angle of resolution (logMAR) of 0.70 or better. Ten patients met both these criteria; however, four were excluded because they were not able to cooperate with the testing requirements of the study because of the onset of multi-infarct dementia, ¹ disabling persistent postoperative torsion, ¹ CNV recurrence, ¹ or difficulties related to transport to the testing center. ¹ The six remaining represented the best outcomes from the surgical series, and all patients were able to perform the reading and fixation tasks described later. 

No patient had diagnosed neurologic or psychiatric disease or other ocular comorbidity. All patients spoke English as a first language and were able to give informed consent. The study conformed to the Declaration of Helsinki, and the ethics committee of Moorfields Eye Hospital gave their approval for the study to take place. 

Preoperative Assessment and Follow-up

Patients underwent a complete ophthalmic assessment before surgery at day 1 and after surgery at 12 months. The assessment consisted of best corrected visual acuity (BCVA), reading acuity and speed, contrast sensitivity, slit lamp biomicroscopy, Goldmann applanation tonometry, ophthalmoscopy, slit lamp assessment of foveal fixation, color fundus photography, scanning laser ophthalmoscope (SLO) autofluorescence, indocyanine green and fluorescein angiography, and optical coherence tomography (Table 1). 

Table 1.

 

Eye Tracker Cohort: Patient Demographics and Post-MT360 Outcomes

Patient	Age (y)	Sex	Follow-up after MT360 (mo)	Residual Torsion (deg)	ETDRS Distance Acuity (logMAR)			MN Read Reading Acuity (logMAR)			MN Read CPS (logMAR)			MN Read Reading Speed (wpm)			Pelli-R obson Contrast Sensitivity (PR Log CS)
					Pre-Op	Post-Op	Δ	Pre-Op	Post-Op	Δ	Pre-Op	Post-Op	Δ	Pre-Op	Post-Op	Δ	Pre-Op	Post-Op	Δ
1	74	F	29	05 (in)	0.44	0.10	0.34	0.60	0.22	0.38	1.10	0.50	0.60	150	231	81	1.05	1.65	0.60
2	63	F	20	10 (in)	0.60	0.30	0.30	0.85	0.60	0.25	1.10	0.90	0.20	146	133	−13	0.90	1.20	0.30
3	69	F	16	05 (in)	1.80	0.12	1.68	1.42	0.22	1.20	1.10	0.70	0.40	70	177	107	0.00	1.65	1.65
4	71	M	12	03 (in)	0.80	0.20	0.60	1.42	0.52	0.91	1.20	0.50	0.70	52	177	125	0.60	1.20	0.60
5	74	M	05	08 (in)	0.88	0.52	0.36	1.02	0.52	0.50	1.10	1.10	0	73	55	−18	0.30	1.55	1.25
6	61	M	06	00	0.50	0.54	−0.04	1.02	0.66	0.36	1.10	1.10	0	140	139	−1	1.05	1.30	0.25
Median	70		14	05	0.70	0.25	P = 0.09	1.02	0.52	P = 0.01	1.10	0.80	P = 0.05	107	158	P = 0.14	0.75	1.43	P = 0.02

 in, incyclotorsion.

Distance Acuity.

The BCVA was measured with a back-illuminated standard Early Treatment of Diabetic Retinopathy Study (ETDRS) chart (Lighthouse Low Vision Products, Long Island City, NY) at 4 m and recorded as a logMAR score. 

Reading Acuity.

The Minnesota Reading Acuity Chart (MN Read; Precision Vision, La Salle, IL) was used to assess reading performance: reading acuity (smallest print that can be read without significant errors), CPS (smallest print that can be read with maximum speed), and reading speed (maximum reading speed when not limited by print size). Reading acuity and CPS were recorded as logMAR scores and reading speed as words per minute (wpm). 

Contrast Sensitivity.

Contrast sensitivity was measured using the Pelli-Robson Contrast Sensitivity Chart (Clement Clarke Inc., Columbus, OH) to assess contrast sensitivity at different spatial frequencies. 

Color Fundus Photography and Angiography.

All color fundus photographic and angiographic images were taken and processed with a retinal camera (TRC-50IX; Topcon Medical Systems, Inc., Paramus, NJ). 

SLO Autofluorescence.

An autofluorescence scan of the posterior pole was performed on all patients with a confocal scanning laser ophthalmoscope (Retina Angiograph 2 HRA2; Heidelberg Engineering, Dossenheim, Germany). 

Slit-Lamp Assessment of Fixation.

A simple, five-step, slit lamp–based system was used to assess the quality of foveal fixation (1). ¹⁶ Fixation was classified on the basis of the results as foveal, unstable, or no foveal fixation. 

Infrared Eye Tracker Assessment of Reading and Fixation

Eye Movement Recording.

Eye movements were measured with a commercial system (Eyelink Gazetracker; SensoMotoric Instruments, Teltow, Germany, running EyeLink Software, ver. 2.04). The system is a headband-mounted apparatus equipped with two adjustable infrared cameras that record eye position by using the “bright pupil” technique. A third camera on the headband tracks head motion by monitoring infrared emitters positioned at the corners of a high-resolution display monitor (21-in. Trinton GDM-F500R; Sony Corp., Tokyo, Japan), with a resolution of 800 × 600 pixels and a refresh frequency of 85 Hz. The luminance of the white background of the screen was 125 cd/m². The patients were seated 50 cm from the monitor and wore suitable refractive correction for the screen distance. With this arrangement, compensation for head motion was made so that a real position of gaze could be calculated. Eye position was measured at a temporal resolution of 250 Hz, with a gaze position accuracy of <0.5°, according to the manufacturer. 

Before each assessment, eye tracker calibration, validation, and drift correction were performed using the algorithms provided for this purpose. Only trials where calibration was categorized as good by the software were included. Saccades were defined as being eye velocity movements with velocity >30 deg/s⁻¹ or acceleration >8000 deg/s⁻² and described in terms of their number, average latency, average velocity, and direction (horizontal, vertical, forward, and regressive). 

Reading Assessment.

A full description of the eye tracker assessment of reading is given in the literature. ^17,18 Briefly, a reading task was performed by using 10 randomly selected sentences consisting of 10 words and a 150-word paragraph displayed on the computer monitor. The sentences were displayed in Times New Roman font, at a Flesch-Kincaid reading grade of 8.0, presented at the optimal text size for the patient according to their CPS, as determined by their performance on the MN Read Chart. 

Assessment of Preferred Retinal Locus/Loci and Fixation Stability.

A full description of the eye tracker assessment of fixation stability is given in the literature. ^{17 –20} Briefly, fixation stability was recorded while patients fixated a target presented for 10 seconds in each of five locations, in a randomized order. The target was a round black disc of 3° diameter with an 18-minute white central detail displayed against a white background. 

The data collected were used to describe fixation by calculating a bivariate contour ellipse, as described in the literature. ^{17 –21} A bivariate contour ellipse describes the locus of fixation in normal observers. The area of this ellipse—the bivariate contour ellipse area (BCEA)—indicates fixation stability. The BCEA of normal individuals is approximately 80 to 1200 min arc² and, with macular disease, it ranges from near normal to over 13,000 min arc². ²² A nonparametric modeling technique (the kernel density estimator) was used to assess whether individual BCEAs provided a better fit to the data than a global BCEA. This analysis was used to determine the number of preferred retinal loci (PRL). 

Assessment of Horizontal and Vertical Saccades.

A full description of the eye tracker assessment of horizontal and vertical saccades is given in the literature. ^17,18 Briefly, horizontal and vertical saccades were assessed while patients followed a simple fixation target (a round black disc of 3° diameter with an 18-minute white central detail displayed against a white background), from left to right for horizontal saccades and from top to bottom for vertical saccades. 

Eye Tracker Control/Comparative Data Groups.

Data collected from eye tracker assessment after MT360 were compared with similar data collected from normal controls (normal), patients with dry AMD, and patients with wet AMD. Each group contained 10 age-matched patients who underwent the aforementioned eye tracker assessment of reading and fixation (Table 2). 

Table 2.

 

Control/Comparative Group Demographics

Group	n	Mean Age, y (range)	Diagnosis	n
Normal	10	71.0 (62–79)	No pathology
Dry AMD	10	76.2 (62–84)	Geographic atrophy	9
			Macular drusen	1
Wet AMD	10	79.2 (74–85)	CNV	7
			PED	3
MT360	6	68.7 (61–74)	CNV	6

 PED, pigment epithelial detachment.

Microperimetry Assessment of Retinal Sensitivity and Fixation

The size of any postoperative scotoma was measured with a microperimeter (MP-1 Microperimeter; Nidek Co., Ltd., Gamagori, Japan), a full description of which is given in the literature. ²³ Briefly, the patients were asked to fixate a central cross while responding, by means of a button press, when they saw a point target presented at a retinal location. The stimuli used were Goldmann III size targets of 200-cd/m intensity ² and were presented for 200 ms. Identification of a retinal landmark enabled a scotoma map to be superimposed onto a retinal image. 

Before each assessment, patients also underwent an objective fixation analysis with the microperimeter. Using a 1° single white cross target of 200-cd/m² intensity patients were asked to fixate the target for 30 seconds. Identification of a retinal landmark enabled the fixation analysis to be superimposed onto a retinal image. 

Surgical Technique

All surgeries were performed in patients under general anesthesia by a single surgeon: macular translocation and removal of silicone oil (LDC) and counterrotation surgery (JL). 

Macular Translocation.

Macular translocation (MT360) was performed as described by Eckardt et al. ⁹  

Globe Counterrotation.

Counterrotation of the translocated eye was performed with removal of silicone oil approximately 6 to 8 weeks after retinal rotation surgery. In all patients, the retina was rotated superiorly, requiring excyclotorsion of the globe. This was achieved with surgery on the oblique muscles combined with transposition of the medial and lateral recti. Counterrotation surgery produced a postoperative median incyclotorsion of 5° (0°–10°), with no patients reporting symptoms of diplopia or tilt (Table 1). 

Statistical Methods

Descriptive statistics (n, median, range, minimum, maximum, and interquartile range) were obtained for each data set (MT360 cohort and the control/comparative groups). The data sets were examined for significant differences in ETDRS acuity, MN Read reading acuity, MN Read CPS, MN Read critical print reading speed, Pelli-Robson contrast sensitivity, eye tracker reading speed, horizontal and vertical saccades, saccadic speed and latency, global fixation, and number of preferred PRLs. Statistical analysis was performed with the Mann-Whitney test (SPSS statistical software; SPSS, Inc., Chicago, IL). All outcome measures quoted herein represent median values, unless otherwise stated. 

Results

MT360 Acuity Outcomes

Surgical Series.

The ETDRS distance and MNRead acuity outcomes for the surgical series of 23 patients (15 women, 8 men; age range, 61–94 years) treated with MT360 are illustrated in Figure 1. 

Figure 1.

 

Surgical group: acuity outcomes. ETDRS distance and MNRead reading acuity outcomes after MT360 in a series of 23 consecutive patients. The patients selected for assessment with the eye tracker are numbered and highlighted with solid data points (Table 1).

Distance Acuity.

At 12 (04–36) months' follow-up, the mean postoperative distance acuity (range) improved from logMAR 0.90 (0.18–2.1) to 0.67 (0.10–1.80; P < 0.05). After surgery, the distance vision improved in 14 (60.9%) patients by greater than 0.1 logMAR, remained unchanged within 0.05 logMAR of the preoperative acuity in 5 (21.7%) patients, and deteriorated by greater than logMAR 0.1 in 4 (17.4%) patients. Before surgery, 48% of patients had a distance acuity of logMAR ≤0.7 (20/100 or better), and 4% a distance acuity of logMAR ≤0.3 (20/40 or better). After surgery this improved to 69% with a distance acuity of logMAR ≤0.7 (20/100 or better), 26% with a distance acuity of logMAR ≤0.3 (20/40 or better), and 39% with a 3-line gain in visual acuity. 

Reading Acuity.

For the same period, the mean postoperative reading acuity (range) improved from logMAR 1.21 (0.60–2.00) to 0.91 (0.22–2.00; P < 0.05). After surgery, the reading acuity improved in 15 (65.2%) patients by greater than 0.1 logMAR, remained within 0.05 logMAR of the preoperative acuity in 4 (17.4%) patients, and deteriorated by greater than 0.1 logMAR in 4 (17.4%) patients. Before surgery, 13% of patients had a reading acuity of logMAR ≤0.7 (N10 or better), no patients had an acuity of logMAR ≤0.3 (N5), with the best preoperative reading acuity recorded at logMAR 0.6 (N8) in a single patient. Before surgery, 52% of patients had a reading acuity of logMAR ≤0.7 (N10), 17% a near acuity of logMAR ≤0.3 (N5), and 61% a gain of 0.3 logMAR. 

Eye Tracker MT360 Cohort

The six best available outcomes (three women, three men; age range, 61–74) from the aforementioned series were recruited to investigate post-MT360 reading behavior (Table 1). At 14 (5–29) months of follow-up, in this cohort (MT360 cohort), surgery improved the postoperative distance acuity from logMAR 0.70 (0.44–1.80) to 0.25 (0.10–0.65; P > 0.05), and the contrast sensitivity (range) from logMAR 0.75 (0–1.05) to 1.43 (1.20–1.65; P < 0.02). In terms of reading function, the postoperative reading acuity improved from logMAR 1.02 (0.60–1.43) to 0.52 (0.22–0.66; P < 0.01), the reading speed from 107 (52–150) to 158 (55–231) wpm (P > 0.05), and the CPS from logMAR 1.10 (1.10–1.20) to 0.80 (0.50–1.10; P < 0.05). 

The relationship between distance and reading acuity, before and after MT360, is summarized in Figure 2. Preoperative distance acuity correlated well with reading acuity (r = +0.71); however, no such correlation was found with CPS (r = −0.04). After surgery both the reading acuity and CPS significantly improved, and both parameters correlated well with the postoperative distance acuity (Fig. 2). 

Figure 2.

 

Relationship between distance acuity and reading function. The charts illustrate the relationship between distance acuity and reading function (acuity and CPS). A comparison of the preoperative (A) and postoperative (B) data reveals the restoration of reading function after MT360 and the reestablishment of the association between distance and near function.

Comparison of Visual Function: MT360 versus Normal versus AMD

Distance Acuity, Reading Function, and Contrast Sensitivity.

Outcome measures (distance and reading acuity and reading speed) for the MT360 cohort and the comparative groups are summarized in Figure 3. The trend for each measure was for postoperative function for the MT360 cohort to be greater than that of both untreated AMD groups and less than the age-matched normal group. This difference in function between the MT360 cohort and the AMD groups reached significance (P < 0.05) for all parameters except CPS and contrast sensitivity. Similarly, the difference in function between MT360 cohort and age-matched normal subjects reached significance (P < 0.05) for all parameters except reading speed (Fig. 3). 

Figure 3.

 

Distance acuity and reading MT360 cohort versus comparative groups. The charts compare the visual parameters of distance acuity (A), reading acuity (B), and reading speed (C) between the MT360 cohort and each of the control/comparative groups. The box plots display the median value, the range, and the SD for each group. P values are highlighted to show results reaching a statistical difference (P < 0.05).

Eye Tracker Assessment of Reading Function.

Horizontal Saccade Task.

Assessment of horizontal saccades in response to a fixation target demonstrated that the MT360 cohort performed a greater number of saccades (MT360 2.44) than the comparative groups (normal 2.00, P < 0.05; dry AMD 1.84, P < 0.05; and wet AMD 2.35, P > 0.05). Examining saccade duration, the latency was longer for the MT360 cohort (269 ms) than the comparative groups (normal 225 ms, P > 0.05; dry AMD 238 ms, P > 0.05; and wet AMD 209 ms, P > 0.05). Finally, saccade velocity was slower for the MT360 cohort (364 ms⁻¹) compared with the other groups (normal 479 ms⁻¹, P < 0.05; dry AMD 459 ms⁻¹, P < 0.05; and wet AMD 470 ms⁻¹, P < 0.05; Table 3). 

Table 3.

 

Saccadic Function: MT360 versus Comparative Groups

Groups	MT360	Normal	Dry AMD	Wet AMD
Horizontal Saccades
Number of saccades	2.44	2.00	1.84	2.35
		P = 0.009	P = 0.008
Latency/ms	269	225	238	209
Velocity/ms−1	364	479	459	470
		P = 0.050	P = 0.013	P = 0.023
Vertical Saccades
Number of saccades	2.32	1.76	2.20	2.09
		P = 0.044
Latency/ms	249	197	217	234
		P = 0.039
Velocity/ms−1	378	446	404	378
Reading Saccades
Forward saccades	13.5	8.7	17.3	29.1
		P = 0.002		P = 0.002
Regressive saccades	29.4	31.5	12.0	33.0

 Data are a comparison of parameters of saccadic function between the MT360 cohort and each of the control/comparative groups. The values represent the median result.

Vertical Saccade Task.

A similar pattern of results was observed for vertical saccades, with the MT360 cohort performing a greater number of saccades of increased latency and slower velocity (Table 3). 

Reading Task.

Saccadic assessment when reading revealed that the MT360 cohort displayed a greater number of forward saccades (13.5) than the age-matched normal group (8.7, P < 0.05) but less than the dry (17.3, P > 0.05) and wet (29.1, P < 0.05) AMD groups. There was no significant difference between the groups for the number of regressive saccades performed (MT360 29.4; normal 31.5, P > 0.05; dry AMD 12.0, P > 0.05; and wet AMD 33.0, P > 0.05; Table 3). 

Reading Speeds.

Assessment of reading speeds revealed that the MT360 cohort read at a speed (101 wpm) that approximated that of the normal group (139 wpm, P < 0.05) and at a greater speed than that of the AMD groups (dry 79 wpm, P > 0.05; wet 39 wpm P < 0.05). Reading speeds, as calculated from MN Read Reading Acuity Charts revealed a similar pattern (MT360 158 wpm; normal 151 wpm, P > 0.05; dry AMD 94 wpm, P < 0.05; and wet AMD 51 wpm, P < 0.05). 

Eye Tracker Assessment of Fixation.

The global BCEA values for the MT360 cohort (310 min arc²) were closer to the normal group (116 min arc²; P > 0.05) than the AMD groups (dry AMD 1426 min arc²; P > 0.05; wet AMD 651 min arc²; P > 0.05), no statistically significant difference was found between the MT360 cohort and any comparative group. Analysis of fixation stability revealed a single preferred retinal loci for the MT360 cohort and the age-matched normal subjects (P > 0.05) with multiple loci for both AMD groups (P < 0.05). 

The aforementioned eye tracker data revealed several significant results and patterns; however, it must be emphasized that nonsignificant results suggest only a trend. Furthermore, the sample sizes limit the ability to regard one test or group as considerably different from another. 

Microperimetry Assessment of Retinal Sensitivity and Fixation.

Retinal Sensitivity.

Microperimetry assessment (MP-1; Nidek) of macular sensitivity demonstrated that in three (patients 1, 3, and 4) of the six patients in the MT360 cohort, surgery restored near normal levels of foveal and parafoveal retinal sensitivity (Fig. 4). Only patient 6 demonstrated a significant residual absolute central scotoma. 

Figure 4.

 

Microperimeter (MP-1; Nidek, Gamagori, Japan) assessment of post-MT360 retinal sensitivity and fixation analysis. Shown for each patient, the recovery of retinal sensitivity and fixation after MT360. For the macular sensitivity images, the areas of peripheral scotoma in red represent nonfunctional retinal areas overlying the original site of subfoveal pathology. Foveal sensitivity maps show the extent of recovery of central function and both sets of sensitivity maps are overlaid with the fixation analysis taken at the time. A separate fixation analysis is also illustrated and the percentage of time the target is placed within 2° and 4° of the fovea during the 30 seconds of the test is illustrated along with the perimetric classification of fixation. All patients demonstrated high quality fixation, classified as greater than 50% of the preferred fixation points located within a 2° diameter circle centered on the fovea. Finally, for each patient the pre- and postoperative BCEA values generated from the perimeter are also illustrated demonstrating either significant improvement in fixation or maintenance of good preoperative fixation.

Fixation.

Microperimetry fixation analysis demonstrated high-quality, stable fixation for all patients in the MT360 cohort. The mean percentage of time that the target was placed within 2° of the fovea over the 30 seconds of the test period ranged from 78% to 100% (Fig. 4). Pre- and post-MT360 BCEA values from the perimeter are also shown for comparison. 

Discussion

MT360 differs from current treatments for severe AMD that are intended to modify the disease process. Instead, translocation offers a unique rescue procedure for foveal photoreceptors in patients whose fellow eye has already lost central vision. Previous reports of translocation have reported this rescue in terms of acuity alone. ^9,12,24 More recently, several groups have examined the impact of MT360 on near visual function and reading ability. ^{10,13 –15,25} Although these studies provide good evidence for the recovery of foveal photoreceptor function by midperipheral retinal pigment epithelium (RPE), none has described the quality and extent of this recovery compared with normal visual function. 

MT360 is a complex procedure encompassing both retinal and strabismus surgery. It is not without significant surgical risk, including endophthalmitis, proliferative vitreoretinopathy (PVR) and associated retinal detachment, choroidal hemorrhage, and anterior segment ischemia. Although improvements in the surgical technique and vitreoretinal instrumentation have considerably reduced some of these risks, particularly PVR detachment, total loss of sight remains a possibility secondary to these complications. Nevertheless, MT360 remains the only treatment modality to offer retinal reconstruction for severe neovascular AMD, in which there is gross mechanical disruption of retinal architecture not amenable to the current generation of medical AMD therapies. 

The results of the first 23 consecutive patients treated with MT360, a series that includes the surgical learning curve, illustrated a significant restoration of both distance and reading acuity (Fig. 1). MT360 led to 39% of patients gaining three lines of distance acuity and 61% gaining 0.3 logMAR of reading acuity, with 17% achieving a near acuity of logMAR ≤0.3 (N5) compared with no patients reading at this level before surgery. These results compare favorably with those in previous investigations into MT360. ^{9,10,12 –15} From this series, we examined the best-outcome cases, to gauge the optimal potential of this complex surgery to restore normal function. Best-outcome cases represented patients who achieved improved or stable distance and reading vision and a distance and reading acuity of logMAR 0.7 or better. These criteria were set to recruit only those patients capable of meeting the demands of the eye tracker and microperimetry assessments. Ten (43.5%) patients met both these criteria from the surgical series in which 69% achieved a distance and 52% a reading acuity of logMAR 0.7 or better, with four patients excluded for reasons outlined in the Methods section. 

At the 14-month follow-up, the measured parameters of visual function all improved or remained relatively stable in this cohort (Table 1). As with the main surgical series, the observed trend was for a greater improvement in near than in distance function. Several studies of MT360 have reported similar findings. ^9,10,13,14 Other than sample size, the discrepancy in near compared with distance vision is difficult to explain and may reflect the complex interplay of location, density, and size of a residual scotoma on reading function. A possible explanation is that MT360 results in a reduction in the foveal scotomatous area, without an improvement in overall retinal sensitivity, thus resulting in a greater improvement in near function. In addition, alterations in microsaccadic behavior after counterrotation surgery (discussed below) may have more of a deleterious effect on distance acuity than near reading ability. 

The failure of reading speed (which is not limited by print size) to improve significantly reflects the fact that a reasonable speed may be attained before surgery (at a larger print size) and only marked improvements in this parameter will reach significance. In this respect, the CPS (smallest print read with maximum speed), which significantly improved after MT360, better reflects improvements in reading ability (Table 1). 

Before surgery distance acuity correlated well (r = +0.71) with reading acuity as expected; however, there was no such correlation (r = −0.04) with CPS (Fig. 2). Patients have a profound central scotoma in their previously better second eye, and it is predominately the loss of reading ability and not distance acuity that commonly instigates presentation. The poor reading capability stems from the scotoma forcing eccentric viewing, ²⁵ an unstable locus for fixation, ^17,21,26,27 and compromised eye movement control. ^{28 –30} In the long term, patients may develop adaptive mechanisms to partially overcome these problems; however, a requirement of successful MT360 is a short duration between acute visual loss and surgery. ¹⁶ Thus, before surgery, although patients may still resolve smaller print sizes, the CPS remains relatively poor and independent of distance and reading acuity. This result again demonstrates that both distance and reading acuity are inadequate measures of reading capability. 

After surgery, distance acuity was well correlated with both the reading acuity (r = +0.79) and CPS (r = +0.91; P < 0.001; Fig. 2). Surgery dramatically reduced or reversed the absolute central scotoma, recovering levels of macular sensitivity approaching normal with stable central fixation (Fig. 4). MT360 resulted in foveal anatomy that closely approximates a normal physiological state, and this is reflected in significant improvements in function (reading acuity, CPS, and contrast sensitivity) and hence reading ability. These results support the premise that restoration of the choriocapillaris–Bruch's–RPE–photoreceptor axis rescues foveal function, and this can be achieved with the midperipheral RPE substituting for the diseased subfoveal RPE. 

To gauge the extent of the recovery detailed above, outcome measures (distance acuity, reading acuity, reading speed, CPS, and contrast sensitivity), eye tracker assessment of reading behavior, and microperimetry were examined, comparing, where possible, the MT360 cohort with age-matched normal subjects and patients with untreated wet and dry AMD. The comparison revealed a pattern indicating that the recovery of function after MT360 was greater than either of the untreated AMD groups and less than the age-matched normal group (Fig. 3, Table 3). The results were consistent with an improvement in the absolute central scotoma with restoration of the macular threshold function and resolution of the size and/or density of the preoperative scotoma. As microperimetry was not available before surgery, changes in retinal threshold function could not be quantified directly; however, all patients in the surgical cohort presented with large a subfoveal hemorrhage and a recent loss of reading capability. Although microperimetry revealed that surgery returns macular sensitivity toward normal levels, a relative scotoma persists (Fig. 4). This residual scotoma is likely to account for reduced reading ability and would also explain the failure of the surgical cohort to match outcomes comparable to normal, except for reading speed. As discussed earlier, although CPS better reflects reading ability, it is of interest to note that, after translocation and counterrotation, patients were able to recover sufficient function to achieve a normal level of maximum reading speed (Fig. 3). 

In addition to near normal levels of macular sensitivity, patients 1, 3, and 4 also displayed the best three postoperative outcomes (Table 1). Of note, patient 1 had the best preoperative vision, whereas patients 3 and 4 had the worst. This result suggests that although good preoperative vision will, as expected, result in good postoperative outcomes, the major determinant for any rescue procedure is likely to be the number and quality of the foveal photoreceptor pool before surgery. Thus, in the present surgical series, case selection was based primarily on the degree of residual foveal function and not on acuity. ¹⁶  

MT360 is intended to restore function by reestablishing retinal anatomy, and so the health of the underlying choriocapillaris–Bruch's–RPE bed that receives the fovea is equally critical to the success of the procedure. In MT360, the fovea is customarily superiorly rotated 20° to 30° and after surgery lies at the margins of the original macular RPE bed. It is known that loss of function is matched by changes in the underlying choriocapillaris–Bruch's–RPE complex. ^{31 –33} As such, a detailed assessment of foveal, parafoveal, and paramacular RPE function may further improve case selection and surgical outcomes. In the present surgical series autofluorescence was performed to give a measure of macular and paramacular RPE function; however, multifocal ERG or microperimetry may be additional investigations that assist in determining the degree of subfoveal support likely to be available at the recipient site. ^{34 –37} Furthermore, the macula consists of a cone-dominated fovea surrounded by a rod-dominated parafovea and measures of scotopic, photopic, and chromatic sensitivities may further determine the degree of residual function and support before surgery. To date, a wealth of studies have described the loss of central macular structure and function, particularly early cone damage, as well as the involvement of wider retinal areas in AMD than are evident from funduscopy (Hageman GS, et al. IOVS 2002;43:ARVO E-Abstract 1988. ^{31,34 –49} Ideally, these factors need to be considered to determine the degree and extent of preexisting disease and establish the baseline retinal health from which recovery is attempted. 

Practically, when patients presented with disease second eyes, case selection for the MT360 surgical series was based on a rapid clinical assessment of residual foveal function. This was determined from the duration of symptoms and performance on a foveal fixation task (1) and not the presenting acuity. ¹⁶ Patients with symptoms beyond 12 weeks were excluded, as were those who did not demonstrate good foveal fixation. It could be argued that duration of visual loss could be used without assessing fixation behavior, as intuitively the two parameters are related. However, simply using duration of visual loss to predict outcomes would not identify which patients were still within the critical window of opportunity to recover residual photoreceptor function. Although this critical window was estimated to be in the region of 12 weeks, the exact duration depended on the number of remaining viable photoreceptors at presentation and the severity and course of the disease process. Thus, patients with a slow, insidious disease course would be expected to have a smaller pool of viable foveal photoreceptors at presentation than would patients presenting with a short overall disease onset. As such, fixation behavior was an important determinant of the critical period of photoreceptor viability. When used together with time since acute visual loss, both parameters provided a powerful method of assessing suitability for MT360 and good outcomes. Ideally, to indicate the contribution of the fixation task versus the duration of visual loss, an examination of cases where there was a marked discordance between fixation behavior and duration of visual loss was required. As expected by the natural history of the disease, no cases of good fixation and a markedly extended period of visual loss were encountered, nor were there any cases of poor fixation and immediate presentation. 

With regard to the procedure itself, ultrastructural and electrophysiological studies have found a partial loss of photoreceptor outer segments after translocation, thought to be torn away during separation of the retina from the RPE. ^{50 –52} Therefore considering that surgical recovery is undertaken on the background of an ageing and diseased retina and that the physical processes involved in MT360 will result in a further loss of photoreceptors, recovery of truly normal function is less likely. 

Eye tracker assessment of saccadic behavior revealed that translocation resulted in a greater number of horizontal saccades of reduced velocity and increased latency compared with that in with the other groups (Table 3). Examination of vertical saccades, which are implicated in finding a new line when reading, revealed a similar pattern. A detailed examination of horizontal saccades when reading revealed that the MT360 cohort displayed a greater number of forward saccades than the age-matched normal subjects but fewer than the AMD groups, with no difference in the number of regressive saccades between groups (Table 3). Eye tracking and microperimetry also established that MT360 restored a single stable central locus of fixation with BCEA values approaching normal subjects (Fig. 4). Thus, in the presence of high-quality stable fixation, these results reflect the direct effects of counterrotation surgery on gaze control during reading rather than deficits in macular function. Counterrotation surgery (which involves transposition of two to four recti muscles) is necessary to correct posttranslocation torsion to less than 10° (when it is considered that symptoms of diplopia and tilt are negligible). Despite the suboptimal oculomotor dynamics highlighted above, the MT360 cohort still achieved near normal reading capability. 

The purpose of our study was to gauge the optimal potential of MT360 in rescuing function and to describe the quality and extent of the recovery. MT360 is a complex two-stage procedure, after which patients have a reduced peripheral field (secondary to 360° retinectomy performed during retinal translocation) and degree of residual torsion (secondary to counterrotation surgery). Despite the profound anatomic (retinal and oculomotor) disruption that MT360 induces, the results indicate that the procedure's results approximate normal macular threshold levels with stable foveal fixation. Although eye tracking identified that surgery quantitatively limited saccadic behavior (number, velocity, and latency), these limitations on oculomotor control appear not to adversely affect qualitative saccadic behavior (saccade direction) when reading, which again approximated that of normal subjects. It is acknowledged that the study findings are clearly limited by the small sample size and the variability of the data set. Nevertheless, the present study demonstrates a clear rescue of photoreceptors and identifies trends in saccadic activity when reading that are an improvement over untreated disease. 

Footnotes

 Supported by the Special Trustees of Moorfields Eye Hospital.

Footnotes

 Disclosure: G. Uppal, None; M. Feely, None; M. Crossland, None; L. Membrey, None; J. Lee, None; L. da Cruz, None; G.S. Rubin, None

Appendix

Table A1.

 

Slit Lamp Assessment of Quality of Fixation

 

 Failed task: defined as failure to follow instructions or maintain fixation for 1 to 3 seconds as outlined.

Table A2.

 

Classification of Fixation Based on Slit Lamp Assessment

Slit Lamp Tasks	Classification of Fixation
All tasks successfully completed	Foveal (good) fixation
Tasks performed, but poorly	Unstable (poor) fixation
More than a single saccade + fixates object for less than 1 second
Unable to complete any tasks	No foveal fixation