- Email: [email protected]
Stereotactic Radiotherapy for Neovascular Age-Related Macular Degeneration
Stereotactic Radiotherapy for Neovascular Age-Related Macular Degeneration Year 2 Results of the INTREPID Study Timothy L. Jackson, PhD, FRCOphth,1 Usha Chakravarthy, MD, PhD,2 Jason S. Slakter, MD,3 Alyson Muldrew, PhD,4 E. Mark Shusterman, MD,5 Denis O’Shaughnessy, PhD,5 Mark Arnoldussen, PhD,5 Michael E. Gertner, MD,6 Linda Danielson, MSc,7 Darius M. Moshfeghi, MD,8 on behalf of the INTREPID Study Group* Purpose: To determine the safety and efficacy of low-voltage, external-beam, stereotactic radiotherapy (SRT) for patients with neovascular age-related macular degeneration (AMD). Design: Randomized, double-masked, sham-controlled, multicenter, clinical trial. Participants: A total of 230 participants with neovascular AMD who received 3 ranibizumab or bevacizumab injections within the preceding year and requiring treatment at enrollment. Methods: Participants received 16 Gray, 24 Gray, or sham SRT. All arms received pro re nata (PRN) ranibizumab for 12 months, with PRN bevacizumab or ranibizumab thereafter. Main Outcome Measures: Mean number of PRN injections; best-corrected visual acuity (BCVA); loss of <15 Early Treatment of Diabetic Retinopathy Study letters; change in optical coherence tomography central subfield thickness; and change in angiographic total lesion area and choroidal neovascularization (CNV) area. Results: At year 2, the 16 and 24 Gray arms received fewer PRN treatments compared with sham (mean 4.5, P ¼ 0.008; mean 5.4, P ¼ 0.09; and mean 6.6, respectively). Change in mean BCVA was 10.0, 7.5, and 6.7 letters for the 16 Gray, 24 Gray, and sham arms, respectively, with 46 (68%), 51 (75%), and 58 participants (79%), respectively, losing <15 letters. Mean central subfield thickness decreased by 67.0 mm, 55.4 mm, and 33.3 mm, respectively. Mean total active lesion area increased by 1.0, 4.2, and 2.7 mm2, respectively. Mean CNV area decreased by 0.1 mm2 in all groups. An independent reading center detected microvascular abnormalities in 6 control eyes and 29 SRT eyes, of which 18 were attributed to radiation; however, only 2 of these possibly affected vision. An exploratory subgroup analysis found that lesions with a greatest linear dimension 4 mm (the size of the treatment zone) and a macular volume greater than the median (7.4 mm3) were more responsive to SRT, with 3.9 PRN injections versus 7.1 in comparable sham-treated participants (P ¼ 0.001) and mean BCVA 4.4 letters superior to sham (P ¼ 0.24). Conclusions: A single dose of SRT significantly reduces intravitreal injections over 2 years. Radiation can induce microvascular change, but in only 1% of eyes does this possibly affect vision. The best response occurs when AMD lesions fit within the treatment zone and they are actively leaking. Ophthalmology 2015;122:138-145 ª 2015 by the American Academy of Ophthalmology. *Supplemental material is available at www.aaojournal.org.
Neovascular age-related macular degeneration (AMD) is a leading cause of blindness in the developed world, with some studies finding that dry and neovascular AMD together account for more blind registrations than all other eye diseases combined.1e4 The standard of care for neovascular AMD involves intravitreal injections of drugs targeting vascular endothelial growth factor (VEGF), but for most patients this necessitates ongoing hospital review and repeated intravitreal injections.5,6 Radiation has been investigated as an alternative treatment option,7 but recent studies present apparently conflicting results.8e13 The Choroidal Neovascularization
138
2015 by the American Academy of Ophthalmology Published by Elsevier Inc.
Secondary to AMD Treated with Beta Radiation Epiretinal Therapy (CABERNET) trial was a pivotal randomized controlled trial (RCT) of epimacular brachytherapy (EMB) used for treatment-naïve neovascular AMD.9,12 Those in the radiation arm received 24 Gray of EMB delivered via a pars plana vitrectomy using an endoscopic probe held over the macula for approximately 3 to 4 minutes. The trial failed to meet either of its co-primary 2-year end points.12 By contrast, the IRay in Conjunction with Anti-VEGF Treatment for Patients with Wet AMD (INTREPID) RCT met its primary end point, showing a statistically significant, one-third reduction in anti-VEGF injections at 1 year http://dx.doi.org/10.1016/j.ophtha.2014.07.043 ISSN 0161-6420/14
Jackson et al
Two-Year Results of the INTREPID Study
after stereotactic radiotherapy (SRT).13 The trial recruited 230 participants with previously treated neovascular AMD who had already received at least 3 anti-VEGF injections in the preceding year and had an ongoing need for antiVEGF therapy at enrollment. Stereotactic radiotherapy is a nonsurgical procedure undertaken in an office setting, using a robotically controlled device that delivers 3 beams of radiation through the inferior sclera to overlap at the macula (Fig 1, available at www.aaojournal.org).14 The eye was held in position using a suction-coupled contact lens; eye tracking software suspended treatment if the eye moved out of alignment. Typical total procedure time was less than 20 minutes, with less than 4 minutes of X-ray delivery. In participants whose AMD lesions were 4 mm (the size of the treatment zone), and who had notable fluid leakage at the time of treatment, there was a highly significant 55% reduction in anti-VEGF retreatment and visual acuity (VA) that was significantly superior (5.3-letter difference) to sham.15 No significant safety concerns were identified.13 Although the year 1 results of INTREPID were encouraging, they were not sufficient to establish the safety of SRT. In particular, radiation retinopathy may occur after 1 year, thus necessitating longer surveillance. In the CABERNET study of EMB, 10 cases (3%) developed radiation retinopathy, and 8 of these occurred in the second year.9,12 In another study of 53 participants treated with EMB, no cases of radiation retinopathy were observed in the first year, but 1 case (2%) emerged in the second year. 8,10,11 Therefore, the second year results of INTREPID provide valuable information regarding the longer-term effects of SRT, because there is a higher likelihood of detecting microvascular abnormalities (MVAs) than in year 1. Furthermore, year 2 results help establish the durability of the treatment effect. This article details the 2-year outcome of the INTREPID study, presented on behalf of the INTREPID study group (Appendix 1, available at www.aaojournal.org).
Methods Study Design Details of the INTREPID study have been reported.13 Briefly, 230 participants with neovascular AMD already receiving anti-VEGF therapy were recruited into a randomized, double-masked, shamcontrolled clinical trial, across 21 European sites (Clinical Trial Registration at www.clinicaltrials.gov, identifier: NCT01016873; accessed October 17, 2013). To be eligible, participants had to have neovascular AMD that was treated with at least 3 intravitreal anti-VEGF injections in the year before enrollment and to have required additional anti-VEGF treatment at the time of enrollment. The choroidal neovascularization (CNV) complex was limited to <12 disc areas, with the greatest linear dimension 6 mm, and the distance from the center of the fovea to the farthest point on the CNV lesion perimeter 3 mm. Full inclusion and exclusion criteria are listed in Appendix 2 (available at www.aaojournal.org). Institutional review board/research ethics committee approval was received for all sites, all participants provided written informed consent, and the trial complied with the tenets of the Declaration of Helsinki.
Study Treatment Participants were randomized to 16 Gray (n ¼ 75) or 24 Gray (n ¼ 75) SRT, or sham SRT (n ¼ 80), using a CE-marked, low-voltage, X-rayebased system (Oraya Therapeutics, Newark, CA). They received a baseline injection of ranibizumab alongside SRT, and thereafter attended every 4 weeks for 1 year for Early Treatment of Diabetic Retinopathy Study (ETDRS) refraction and determination of best-corrected VA (BCVA), slit-lamp biomicroscopy, dilated fundus examination, and time-domain optical coherence tomography (Stratus OCT, Carl Zeiss Meditec, Dublin, CA). Participants were retreated with monthly pro re nata (PRN) intravitreal 0.5 mg ranibizumab (Lucentis, Genentech, South San Francisco, CA) if they met predefined retreatment criteria, including any of the following: a 100-mm increase in central subfield thickness from the lowest previous OCT measurement; new or increased macular hemorrhage documented by fundus photographs; or a >5 letter decrease in BCVA since the last visit or the baseline BCVA, with disease activity such as persistent or increased fluid on OCT or leakage on fluorescein angiography (FA). After the first year of protocol-mandated follow-up and treatment, participants reverted to their standard care, with the criteria for retreatment determined by the attending clinician. Participants returned for mandated study safety visits at 18 and 24 months (with a further safety visit planned for 36 months). At the month 24 visit, participants underwent full ophthalmic examination, ETDRS BCVA, OCT, fundus photography, and FA using trial-certified equipment and staff. Images were assessed by the same independent reading center.
Outcome Measures The primary outcome was the number of PRN ranibizumab injections administered over 52 weeks. Year 1 secondary outcomes included the change in mean BCVA; the proportion of participants losing <15 letters, gaining 0 letters, and gaining 15 letters; and the change in mean total lesion area and mean CNV area based on fundus photographs and FA, assessed by a masked, independent reading center.13 These outcomes were also assessed at year 2. In addition, the reading center assessed the change in OCT central subfield thickness at 1 and 2 years. Safety outcomes included adverse events (AEs) and serious AEs. The reading center specifically examined for any MVAs that might be due to radiation. To increase the likelihood of detecting radiationinduced changes, the graders reported all MVAs even if these were within the area occupied by the neovascular lesion and therefore might be due to AMD rather than radiation. Graders were masked to study arm and timing (before or after baseline treatment/sham).
Subgroup Analysis An exploratory subgroup analysis was conducted to determine which baseline variables influence the response to SRT. A prior analysis of the year 1 data showed that lesions that could fit within the 4-mm treatment zone responded better than lesions extending beyond the zone, as were lesions with significant macular fluid at the time of treatment.15 These analyses were repeated at year 2. Group definitions were lesions 4 mm in greatest linear dimension versus those >4 mm, and lesions with an OCT macular volume greater than the median value of 7.4 mm3 versus lesions 7.4 mm3. The greatest linear dimension and macular volume were determined by the reading center. Several other subgroup analyses were undertaken to determine which variables influence the response to SRT, as undertaken at year 1.15
139
Ophthalmology
Volume 122, Number 1, January 2015
Table 1. Number of Pro Re Nata AntieVascular Endothelial Growth Factor Injections 16 Gray (N [ 75)
24 Gray (N [ 75)
16 Gray and 24 Gray Combined (N [ 150)
Sham (N [ 80)
Follow-up
Mean
SD
P
Mean
SD
P
Mean
SD
P
Mean
SD
Week 52 Week 104
2.6 4.5
2.5 3.9
0.009 0.008
2.4 5.4
2.4 4.6
0.003 0.09
2.5 4.9
2.4 4.3
0.0008 0.009
3.7 6.6
2.6 4.3
SD ¼ standard deviation.
Statistical Analysis
Visual Acuity
The statistical approach to the year 1 end points has been described.13 The primary group for the efficacy analysis was the intention-to-treat population, which consisted of all randomized patients. All efficacy analyses were performed by randomized arm assignment. The analysis population for safety consisted of all treated patients. All safety analyses were addressed by actual treatment received. The year 2 analyses were primarily exploratory, but were planned to mirror those of year 1. Specifically, each treated cohort was compared with sham with respect to the number of ranibizumab injections up to and including month 24 by comparing the means calculated by an analysis of variance adjusted for the number (3 vs. >3) and type of previous injections (ranibizumab vs. bevacizumab vs. other), whether the patient exhibited a dry macula after previous anti-VEGF therapy, time since diagnosis of wet AMD (<6 vs. 6 months), and baseline VA score (<54 vs. 54 letters). In addition, the analysis was adjusted for the minimization algorithm (protocol version B vs. CþDþE vs. F). The change in mean BCVA in each SRT arm compared with that in the pooled sham arms from baseline to 52 weeks was analyzed using a restricted maximum likelihoodebased repeatedmeasures approach. Data from all scheduled post-baseline visits, up to and including the outcome visit (24 months), were used in the analysis. The model included the same stratification factors as for the number of injections.
As previously reported, at the end of year 1, the mean change from baseline in VA was 0.28 (SD, 8.77) letters in the 16 Gy arm, þ0.40 (SD, 10.33) letters in the 24 Gy arm, and 1.57 (SD, 11.90) letters in the sham arm (P ¼ 0.73 and P ¼ 0.19, 16 Gy and 24 Gy compared with sham, respectively).13 At year 2, the mean BCVA was 48.3 letters (SD, 18.1) in the 16 Gy group, 52.5 letters (SD, 17.7) in the 24 Gy group, and 51.1 letters (SD, 20.1) in the sham group. The mean change in BCVA from baseline to 2 years was 10.0 letters (SD, 16.5) in the 16 Gy group, 7.5 letters (SD, 16.5) in the 24 Gy group, and 6.7 letters (SD, 17.1) in the sham group. If the 16 Gy and 24 Gy arms were combined, the mean VA change was 8.8 letters (SD, 16.5). This difference was not statistically significant. The proportions of participants losing <15 letters in the 16 Gy, 24 Gy, and sham arms were 68% (N ¼ 46), 75% (N ¼ 51), and 79% (N ¼ 58), respectively. The proportions of participants gaining 0 letters were 32% (N ¼ 22), 43% (N ¼ 29), and 38% (N ¼ 28), respectively. The proportions of participants gaining 15 letters were 3% (N ¼ 2), 1% (N ¼ 1), and 3% (N ¼ 2).
Results
Fluorescein Angiography, Fundus Photography, and Optical Coherence Tomography Mean total active lesion area increased by 1.0 mm2 (SD, 5.7), 4.2 mm2 (SD, 8.1), and 2.7 mm2 (SD, 8.6) for the 16 Gy, 24 Gy, and sham arms, respectively, at 104 weeks. The mean CNV area decreased by 0.1 mm2 in all groups (SD, 0.3, 0.4, and 0.2 for the 16 Gy, 24 Gy, and sham arms, respectively). Mean OCT central subfield thickness at year 2 decreased by 66.1 mm (SD, 101.5), 55.4 mm (SD, 130.3), and 33.3 mm (SD, 112.7) in the 16 Gy, 24
Number of Anti-Vascular Endothelial Growth Factor Retreatments The arms were balanced in terms of baseline demographics.13 Of the 230 participants who enrolled, 212 (92%) remained in the study through the week 104 visit (Fig 2, available at www.aaojourna l.org). At enrollment, the mean time from diagnosis of wet AMD was 12.8 months (standard deviation [SD], 7.3) in the 16 Gy arm, 16.7 months (SD, 10.1) in the 24 Gy arm, and 16.2 months (SD, 13.6) in the sham arm. The mean number of anti-VEGF injections the patients had already received before enrollment was 4.99 (SD, 3.87) in the 16 Gy arm, 6.23 (SD, 4.76) in the 24 Gy arm, and 5.48 (SD, 4.19) in the sham arm. At week 104 inclusive, those in the combined 16 and 24 Gy arms had a highly significant 26% reduction in the number of PRN injections compared with the sham arm (1.6; 95% confidence interval, 2.7 to 0.4, P ¼ 0.009). The mean number of PRN anti-VEGF injections is shown in Table 1 and Figure 3. Both the 16 Gy and 24 Gy SRT arms received significantly fewer PRN ranibizumab treatments at 12 months compared with the sham arm.13 This difference remained significant at year 2 in the 16 Gray arm but was not significant in the 24 Gray arm (Table 1).
140
Figure 3. The proportion of participants receiving different numbers of pro re nata (PRN) intravitreal injections over 2 years. The combined stereotactic radiotherapy (SRT) arm is compared with the sham arm.
Jackson et al
Two-Year Results of the INTREPID Study
Table 2. Summary of Year 2 Adverse Events and Serious Adverse Events in the Study Eye No. of Patients No. of evaluable participants in year 2* Participants with AE AE in study eye Radiation-related AE Procedure-related AE Participants with SAE SAE in study eye Total number of AEs Total number of SAEs
16 Gray (N [ 74) 13 4 1 2
71 (18%) (6%) (1%) 0 (3%) 0 17 2
24 Gray (N [ 73) 14 6 1 2
69 (20%) (9%) (1%) 0 (3%) 0 22 2
Sham (N [ 79) 12 4
3 1
76 (16%) (5%) 0 0 (4%) (1%) 14 3
Total (N [ 226) 216 39 (18%) 14 (6%) 2 (<1%) 0 7 (3%) 1 (<1%) 53 7
AE ¼ adverse event; SAE ¼ serious adverse event. Note: All AEs with a start date after the week 52 visit date were included in the year 2 analysis. *All safety participants who had a visit after week 52.
Gy, and sham arms, respectively. The combined SRT group had a 61.1 mm (SD, 116.7) reduction in central subfield thickness.
Subgroup Analysis Lesion size and macular fluid volume both influenced the response to SRT. In the 60 eyes with lesions 4 mm and a macular volume greater than the median (26% of the trial population at week 104), the SRT-treated eyes (N ¼ 40) had 45% fewer anti-VEGF injections than comparable sham-treated eyes (N ¼ 20) (3.9 [SD, 3.5] for combined SRT arms vs. 7.1 [SD, 3.3]; P ¼ 0.001). The BCVA in these SRT-treated eyes was also 4.4 letters superior to comparable sham-treated eyes (1.7 [SD, 12.5] vs. 6.2 [SD, 14.6], P ¼ 0.24). In this subgroup, mean total active lesion area increased by 1.6 mm2 (SD, 4.0) and 2.7 mm2 (SD, 2.3) for the SRT and sham arms, respectively, at 104 weeks. The mean CNV area decreased by 0.5 mm2 (SD, 0.5) and 0.0 mm2 (SD, 0) in the SRT and sham arms, respectively. Mean decreases in OCT central subfield thickness at year 2 were 120.7 mm (SD, 112.7) for the SRT-treated eyes compared with 40.1 mm (SD, 96.0) in the sham group. Other baseline variables analyzed for their influence on treatment response are shown in Figure 4 (available at www.aao journal.org). The absence of fibrosis at the time of SRT treatment was associated with a significantly greater decrease in the number of PRN injections (mean 4.9 [SD, 3.9] in the SRT group vs. 6.9 [SD, 4.2] in the sham arm; P ¼ 0.004) and a significantly greater decrease in OCT central subfield thickness (mean 79.2 mm [SD, 118.9] for the SRT treated eyes compared with 15.9 mm [SD, 115.1] in the sham group; P ¼ 0.005).
Safety The numbers of AEs and serious AEs were similar across arms (Table 2). Table 3 (available at www.aaojournal.org) lists AEs occurring in the study eye during year 2. Per protocol, only “related” AEs were required to be reported after the first year of follow-up. Any FA or fundus photographs that the reading center identified as having possible MVAs were referred for review by expert reading center clinicians (U.C. and J.S.S.) who were masked to both study arm and timing (before or after SRT). The reading center identified 35 cases with MVAs. These were reviewed by the experts, and 18 of these were thought likely to be related to radiation. When subsequently unmasked, all 18 of these cases were found to be in the radiation arms (9 per arm). Of the remaining 17 cases that were not deemed to be radiation-related, 6 were in the sham group. Only 2 cases (1 from each SRT arm) were thought to
have radiation-related MVAs that were likely to affect vision, because all other cases were extrafoveal, with 10 of 18 located outside the 4-mm treatment zone (Fig 5, available at www.aao journal.org). Of the 2 participants with foveal MVAs, 1 lost 46 letters and required 13 PRN injections over 2 years, and the other lost 41 letters and required 8 PRN injections over 2 years. As with other cases of MVAs, the main feature was capillary nonperfusion. The experts agreed that the main cause of visual impairment in these 2 eyes was the underlying AMD, but that the foveal MVAs may have been contributory. Representative images are shown and described in Figure 6. At 2 years, the mean change in BCVA in the 18 SRT eyes with MVAs attributed to radiation was similar to SRT-treated eyes without MVA, at 9.0 letters (SD, 21.1) and 8.7 letters (SD, 15.7), respectively.
Discussion At 1 year, the INTREPID trial met its primary end point, showing that SRT significantly reduced the frequency of anti-VEGF injections that patients required to treat their neovascular AMD.13 This result was encouraging, but it did not indicate whether the benefit of SRT persisted in the longer term. In addition, ongoing safety review was needed, because radiation can induce MVAs more than 1 year after exposure.8e12 Our results confirmed that SRT produced a significant reduction in anti-VEGF requirement over 2 years. In the second year, some participants developed MVAs in response to radiation, but these seldom affected vision, because the majority occurred outside the fovea. Overall, SRT produced a reduction in anti-VEGF retreatment by one quarter over 2 years. In particular, 15% of SRT recipients required no anti-VEGF injections in the 2 years after SRT and 45% required 3 injections or fewer. For these patients, there is the possibility of reduced clinic attendance and reduced cumulative risk of repeated injections, such as endophthalmitis, increased intraocular pressure, macular atrophy, and arteriothrombotic events. Our subgroup analysis suggested that patients are best treated with SRT when they have appreciable macular fluid. Likewise, lesions that could fit within the 4-mm treatment zone responded better than larger lesions. For eyes with both characteristics (26% of participants), there was a highly significant 45% reduction in anti-VEGF retreatment and a
141
Ophthalmology
Volume 122, Number 1, January 2015
Figure 6. A1e4, A 70-year-old woman treated with 16 Gray, comparing baseline with year 2. The patient gained 1 letter of visual acuity (VA) (74 letters at baseline vs. 75 letters at year 2) and required 1 pro re nata [PRN]) injection over 2 years (at week 20). Her medical history included osteoporosis and hyperlipidemia. The arrow shows an area of telangiectasia and capillary nonperfusion due to radiation exposure (A4). B1e4, A 69-year-old woman treated with 16 Gray. This patient lost 25 letters over 2 years (67 letters at baseline vs. 42 letters at year 2) and required 8 PRN injections. Her medical history included hypothyroidism and hyperlipidemia. The fundus photograph at year 2 shows a retinal hemorrhage (B2), and the fluorescein angiogram shows an area of capillary nonperfusion (circled; B4). The expert panel concluded that the vision loss was due to disease progression.
trend for better VA compared with sham-treated eyes with the same characteristics. It seems logical that lesions within the treatment zone would do better than those beyond it, because the latter may be undertreated at their outer margin. Radiation is known to preferentially damage proliferating
142
cells, and it is possible that actively proliferating endothelial cells are more likely to leak fluid.14,16 Therefore, the findings of our subgroup analysis are biologically plausible and consistent with our findings at year 1.15 In addition, a prospectively defined analysis comparing eyes with baseline
Jackson et al
Two-Year Results of the INTREPID Study
Figure 6. (continued). C1e6, A 65-year-old woman, 1 of the 2 patients thought to have lost vision in part because of microvascular abnormalities. She was treated with 24 Gray and experienced a 46-letter loss over 2 years (67 letters at baseline vs. 21 letters at year 2). This patient required 13 PRN ranibizumab injections. Her medical history included hypertension and hyperlipidemia. Early (C3e4) and late-phase (C5e6) fluorescein angiograms are provided. Fundus photographs (C1e2) show disciform scarring secondary to wet age-related macular degeneration. The fluorescein angiogram shows telangiectasia and capillary dropout centered on the fovea, visible in the early phase (C4). A late-phase angiogram shows extensive staining due to the underlying disciform lesion (C6). The expert panel concluded that the vision loss was primarily due to disease progression but may have been affected by radiation.
fibrosis with those without fibrosis found the latter had significantly fewer anti-VEGF injections and a greater decrease in central subfield thickness. This finding suggests that SRT may be best used before the formation of macular scarring. Masked analysis by the reading center found MVAs in 19% of the SRT participants and 8% of controls. A subsequent expert analysis of the SRT arm found that 10% of this group had MVAs attributable to radiation. As shown in Figure 5 (available at www.aaojournal.org), all but 2 MVAs occurred outside the fovea and the majority occurred outside the 4-mm treatment zone. Microvascular abnormalities were generally not reported by investigators. This is because MVAs are better seen by careful scrutiny of color photographs and fluorescein angiograms by trained graders. Only 2 cases of MVA were detected in the clinical setting. All other cases were detected by the reading center. It remains possible that at least some of these relatively nonspecific
changes were due to other causes. For example, an analysis of fundus photographs and FAs from a large RCT of pegaptanib found that 14% of untreated eyes with neovascular AMD have MVAs.17 The CABERNET study of EMB found that 3% of participants developed MVAs over 2 years, but paradoxically these participants had better outcomes than those without MVAs.9,12 Our patients with MVAs had outcomes similar to those without MVAs. The retinal changes seen in response to low-voltage SRT for treatment of wet AMD therefore appear different than the changes associated with high-dose, mega-voltage treatment for ocular malignancy. Ongoing evaluation of images obtained at the 36-month follow-up has indicated regression of previously noted MVAs in a number of SRT-treated eyes (Moshfeghi DM. INTREPID study of stereotactic radiotherapy for wet age-related macular degeneration: efficacy and safety at 24 and 36 months. Presented at: Retina Specialty Day, American Academy of Ophthalmology Annual
143
Ophthalmology
Volume 122, Number 1, January 2015
Meeting, November 15, 2013, New Orleans, LA). Nonetheless, patients considering SRT should be warned that there is an approximately 1% risk of foveal MVAs over 2 years, and that long-term follow-up is required. Although the CABERNET study of EMB and the INTREPID study of SRT had similar safety outcomes, the efficacy results differed substantially.9,12,13 This apparent discrepancy may be due to a number of factors.14 The EMB device used a strontium source that delivered an exponentially reduced dose with increasing distance from the probe.18 Therefore, probe placement by the surgeon was critical, not only in positioning on the surface of the retina but also topographically in relation to the area of greatest disease activity, which for occult lesions can sometimes be difficult to define. By contrast, the 90% isodose of the SRT device extends for 4 mm before showing a rapid decline because of the highly collimated and tightly targeted X-ray beams (Fig 1, available at www.aaojournal.org). Consequently, if a lesion fits within the 4-mm treatment zone, then the entire lesion receives an appropriate dose. Also, the SRT device uses eye tracking software to ensure correct positioning, whereas the EMB device relies on surgical manipulation. Of note, SRT is nonsurgical and avoids vitrectomy. Animal studies suggest that vitrectomy reduces intravitreal anti-VEGF half-life up to 10-fold, and therefore any residual disease activity may be hard to control.19,20 Strengths of this study include its multicenter, randomized, double-masked, sham-controlled design, with independent reading of fundus photographs, FAs, and OCTs. Year 2 data are more useful than year 1 data in terms of safety, but radiation retinopathy can occur beyond 2 years, and a year 3 safety analysis is planned. The fact that participants returned to standard care between year 1 and year 2 study visits is both a strength and weakness, because retreatment and monitoring are likely to be less regimented and verifiable outside of a trial, but conversely, pragmatic data may be more generalizable to routine care. The retreatment criteria in year 1 were typical of those used at the time the INTREPID study was designed, but they were less aggressive than current “treat until dry” regimens.21 The year 2 local care retreatment criteria varied by site. These less stringent retreatment criteria may have reduced the magnitude of the benefit demonstrated by SRT if all arms received fewer injections than would be expected with current practice. Also, the time-domain OCT used in this study was less sensitive than current spectral-domain OCT, and this may have reduced the likelihood of detecting macular fluid and further reduced the number of injections in both arms. When this study commenced, most clinical trials used time-domain OCT because the protocols for spectral-domain image acquisition and grading were not yet well validated and established. However, clinicians remained masked to participants’ SRT status throughout the 2-year study period addressed in this article, and therefore bias could not have been introduced into anti-VEGF retreatment decision making. These facts strengthen the study’s findings, because significantly fewer injections were seen despite low injection rates in both arms. They also suggest that the proportional reduction in injection frequency may be more relevant than the absolute reduction. In
144
retrospect, it would have been logical to exclude lesions beyond the 4-mm treatment zone, because subgroup analysis suggests that better case selection would have strengthened the result. Although subgroup analyses are important in terms of future trial design and may help refine case selection, they are exploratory in nature and should be interpreted accordingly. The VA and OCT results of the radiation-treated arms did not differ significantly from the control group, but the study was not designed to establish if they were superior or noninferior. The VA results cannot be compared with trials of treatment-naïve patients, which have generally shown substantial visual gains at the start of antiVEGF therapy. By comparison, VA typically decreases as patients enter the chronic treatment phase, as seen in 3 of 4 arms transitioning from the first to second year of the Comparison of Age-Related Macular Degeneration Treatments Trials (CATT).21,22 In conclusion, this study found that SRT was associated with significantly reduced anti-VEGF retreatment over 2 years. Microvascular changes may have occurred in response to radiation, but they seldom affected vision.
References 1. Owen CG, Jarrar Z, Wormald R, et al. The estimated prevalence and incidence of late stage age-related macular degeneration in the UK. Br J Ophthalmol 2012;96:752–6. 2. Munoz B, West SK, Rubin GS, et al. SEE Study Team. Causes of blindness and visual impairment in a population of older Americans: the Salisbury Eye Evaluation Study. Arch Ophthalmol 2000;118:819–25. 3. Finger RP, Fimmers R, Holz FG, Scholl HP. Prevalence and causes of registered blindness in the largest federal state of Germany. Br J Ophthalmol 2011;95:1061–7. 4. Bunce C, Xing W, Wormald R. Causes of blind and partial sight certifications in England and Wales: April 2007-March 2008. Eye (Lond) 2010;24:1692–9. 5. Heier JS, Brown DM, Chong V, et al; VIEW 1 and VIEW 2 Study Groups. Intravitreal aflibercept (VEGF Trap-Eye) in wet age-related macular degeneration. Ophthalmology 2012;119: 2537–48. 6. Rosenfeld PJ, Brown DM, Heier JS, et al; MARINA Study Group. Ranibizumab for neovascular age-related macular degeneration. N Engl J Med 2006;355:1419–31. 7. Chakravarthy U, Houston RF, Archer DB. Treatment of agerelated subfoveal neovascular membranes by teletherapy: a pilot study. Br J Ophthalmol 1993;77:265–73. 8. Petrarca R, Dugel PU, Nau J, et al. Macular epiretinal brachytherapy in treated age-related macular degeneration (MERITAGE): month 12 optical coherence tomography and fluorescein angiography. Ophthalmology 2013;120:328–33. 9. Jackson TL, Dugel PU, Bebchuk JD, et al; CABERNET Study Group. Epimacular brachytherapy for neovascular age-related macular degeneration (CABERNET): fluorescein angiography and optical coherence tomography. Ophthalmology 2013;120: 1597–603. 10. Petrarca R, Dugel PU, Bennett M, et al. Macular epiretinal brachytherapy in treated age-related macular degeneration (MERITAGE): month 24 safety and efficacy results. Retina 2014;34:874–9. 11. Dugel PU, Petrarca R, Bennett M, et al. Macular epiretinal brachytherapy in treated age-related macular degeneration:
Jackson et al
12.
13.
14. 15.
16.
Two-Year Results of the INTREPID Study
MERITAGE Study: twelve-month safety and efficacy results. Ophthalmology 2012;119:1425–31. Dugel PU, Bebchuk JD, Nau J, et al; CABERNET Study Group. Epimacular brachytherapy for neovascular age-related macular degeneration: a randomized, controlled trial (CABERNET). Ophthalmology 2013;120:317–27. Jackson TL, Chakravarthy U, Kaiser PK, et al; INTREPID Study Group. Stereotactic radiotherapy for neovascular agerelated macular degeneration: 52-week safety and efficacy results of the INTREPID Study. Ophthalmology 2013;120: 1893–900. Petrarca R, Jackson TL. Radiation therapy for neovascular age-related macular degeneration. Clin Ophthalmol 2011;5: 57–63. Jackson TL, Shusterman EM, Anoldussen, et al. Stereotactic radiotherapy for wet age-related macular degeneration (INTREPID): influence of baseline characteristics on clinical response. Retina August 6, 2014 [Epub ahead of print]. Kirwan JF, Constable PH, Murdoch IE, Khaw PT. Beta irradiation: new uses for an old treatment: a review. Eye (Lond) 2003;17:207–15.
17. Jackson TL, Danis RP, Goldbaum M, et al. Retinal vascular abnormalities in neovascular age-related macular degeneration. Retina 2014;34:568–75. 18. Petrarca R, Richardson M, Douiri A, et al. Safety testing of epimacular brachytherapy with microperimetry and indocyanine green angiography: 12-month results. Retina 2013;33: 1232–40. 19. Christoforidis JB, Xie Z, Jiang A, et al. Serum levels of intravitreal bevacizumab after vitrectomy, lensectomy and non-surgical controls. Curr Eye Res 2013;38:761–6. 20. Kakinoki M, Sawada O, Sawada T, et al. Effect of vitrectomy on aqueous VEGF concentration and pharmacokinetics of bevacizumab in macaque monkeys. Invest Ophthalmol Vis Sci 2012;53:5877–80. 21. CATT Research Group. Ranibizumab and bevacizumab for neovascular age-related macular degeneration. N Engl J Med 2011;364:1897–908. 22. Comparison of Age-related Macular Degeneration Treatments Trials (CATT) Research Group, Martin DF, Maguire MG, Fine SL, et al. Ranibizumab and bevacizumab for treatment of neovascular age-related macular degeneration: two-year results. Ophthalmology 2012;119:1388–98.
Footnotes and Financial Disclosures Originally received: February 8, 2014. Final revision: March 11, 2014. Accepted: July 29, 2014. Available online: September 7, 2014. 1
J.S.S.: Research Grant Support from Oraya. D.M.M.: stock options. Manuscript no. 2014-224.
School of Medicine, King’s College London, London, United Kingdom.
2
Queen’s University of Belfast, Consultant in Ophthalmology, The Belfast Health and Social Care Trust, Belfast, Northern Ireland.
3
Digital Angiography Reading Center, Great Neck, New York.
4
Centre for Vision and Vascular Science, Queen’s University of Belfast, Belfast, Northern Ireland. 5
Oraya Therapeutics, Inc, Newark, California.
6
Department of Surgery, Stanford University, Stanford, California.
7
The International Drug Development Institute, Louvain-la-Neuve, Belgium.
A.M.: Employee of the Centre for Vision and Vascular Science, which was retained by Oraya Therapeutics as a reading center for the INTREPID trial. M.A.: Employee of Oraya Therapeutics. E.M.S. and D.O.: Former employees and current consultants to Oraya Therapeutics and have stock options. M.E.G.: Founder and board member of Oraya Therapeutics and holds significant equity. L.D.: Employee of the International Drug Development Institute, which was retained by Oraya Therapeutics to conduct data management and statistical analysis of the INTREPID trial. Oraya Therapeutics participated in the design of the study, conducting the study, data collection, data management, data analysis, and review of the manuscript.
8
Byers Eye Institute, Horngren Family Vitreoretinal Center, Department of Ophthalmology, Stanford University School of Medicine, Palo Alto, California. *A full listing of the INTREPID Study Group is available in Appendix 1 (available at www.aaojournal.org).
Presented at: EURETINA, September 26e27, 2013, Hamburg, Germany. Financial Disclosure(s): The author(s) have made the following disclosure(s): T.L.J.: Employer received research funding from Oraya, and received a single advisory board payment and travel support from Oraya. U.C., J.S.S., A.M., D.M.M.: Consultants to Oraya Therapeutics.
Abbreviations and Acronyms: AE ¼ adverse event; AMD ¼ age-related macular degeneration; BCVA ¼ best-corrected visual acuity; CNV ¼ choroidal neovascularization; EMB ¼ epimacular brachytherapy; FA ¼ fluorescein angiography; MVA ¼ microvascular abnormality; OCT ¼ optical coherence tomography; PRN ¼ pro re nata; RCT ¼ randomized controlled trial; SD ¼ standard deviation; SRT ¼ stereotactic radiotherapy; VA ¼ visual acuity; VEGF ¼ vascular endothelial growth factor. Correspondence: Timothy L. Jackson, PhD, FRCOphth, School of Medicine, King’s College London, Department of Ophthalmology, King’s College Hospital, London SE5 9RS, UK. E-mail: [email protected].
145
Neovascular age-related macular degeneration (AMD) is a leading cause of blindness in the developed world, with some studies finding that dry and neovascular AMD together account for more blind registrations than all other eye diseases combined.1e4 The standard of care for neovascular AMD involves intravitreal injections of drugs targeting vascular endothelial growth factor (VEGF), but for most patients this necessitates ongoing hospital review and repeated intravitreal injections.5,6 Radiation has been investigated as an alternative treatment option,7 but recent studies present apparently conflicting results.8e13 The Choroidal Neovascularization
138
2015 by the American Academy of Ophthalmology Published by Elsevier Inc.
Secondary to AMD Treated with Beta Radiation Epiretinal Therapy (CABERNET) trial was a pivotal randomized controlled trial (RCT) of epimacular brachytherapy (EMB) used for treatment-naïve neovascular AMD.9,12 Those in the radiation arm received 24 Gray of EMB delivered via a pars plana vitrectomy using an endoscopic probe held over the macula for approximately 3 to 4 minutes. The trial failed to meet either of its co-primary 2-year end points.12 By contrast, the IRay in Conjunction with Anti-VEGF Treatment for Patients with Wet AMD (INTREPID) RCT met its primary end point, showing a statistically significant, one-third reduction in anti-VEGF injections at 1 year http://dx.doi.org/10.1016/j.ophtha.2014.07.043 ISSN 0161-6420/14
Jackson et al
Two-Year Results of the INTREPID Study
after stereotactic radiotherapy (SRT).13 The trial recruited 230 participants with previously treated neovascular AMD who had already received at least 3 anti-VEGF injections in the preceding year and had an ongoing need for antiVEGF therapy at enrollment. Stereotactic radiotherapy is a nonsurgical procedure undertaken in an office setting, using a robotically controlled device that delivers 3 beams of radiation through the inferior sclera to overlap at the macula (Fig 1, available at www.aaojournal.org).14 The eye was held in position using a suction-coupled contact lens; eye tracking software suspended treatment if the eye moved out of alignment. Typical total procedure time was less than 20 minutes, with less than 4 minutes of X-ray delivery. In participants whose AMD lesions were 4 mm (the size of the treatment zone), and who had notable fluid leakage at the time of treatment, there was a highly significant 55% reduction in anti-VEGF retreatment and visual acuity (VA) that was significantly superior (5.3-letter difference) to sham.15 No significant safety concerns were identified.13 Although the year 1 results of INTREPID were encouraging, they were not sufficient to establish the safety of SRT. In particular, radiation retinopathy may occur after 1 year, thus necessitating longer surveillance. In the CABERNET study of EMB, 10 cases (3%) developed radiation retinopathy, and 8 of these occurred in the second year.9,12 In another study of 53 participants treated with EMB, no cases of radiation retinopathy were observed in the first year, but 1 case (2%) emerged in the second year. 8,10,11 Therefore, the second year results of INTREPID provide valuable information regarding the longer-term effects of SRT, because there is a higher likelihood of detecting microvascular abnormalities (MVAs) than in year 1. Furthermore, year 2 results help establish the durability of the treatment effect. This article details the 2-year outcome of the INTREPID study, presented on behalf of the INTREPID study group (Appendix 1, available at www.aaojournal.org).
Methods Study Design Details of the INTREPID study have been reported.13 Briefly, 230 participants with neovascular AMD already receiving anti-VEGF therapy were recruited into a randomized, double-masked, shamcontrolled clinical trial, across 21 European sites (Clinical Trial Registration at www.clinicaltrials.gov, identifier: NCT01016873; accessed October 17, 2013). To be eligible, participants had to have neovascular AMD that was treated with at least 3 intravitreal anti-VEGF injections in the year before enrollment and to have required additional anti-VEGF treatment at the time of enrollment. The choroidal neovascularization (CNV) complex was limited to <12 disc areas, with the greatest linear dimension 6 mm, and the distance from the center of the fovea to the farthest point on the CNV lesion perimeter 3 mm. Full inclusion and exclusion criteria are listed in Appendix 2 (available at www.aaojournal.org). Institutional review board/research ethics committee approval was received for all sites, all participants provided written informed consent, and the trial complied with the tenets of the Declaration of Helsinki.
Study Treatment Participants were randomized to 16 Gray (n ¼ 75) or 24 Gray (n ¼ 75) SRT, or sham SRT (n ¼ 80), using a CE-marked, low-voltage, X-rayebased system (Oraya Therapeutics, Newark, CA). They received a baseline injection of ranibizumab alongside SRT, and thereafter attended every 4 weeks for 1 year for Early Treatment of Diabetic Retinopathy Study (ETDRS) refraction and determination of best-corrected VA (BCVA), slit-lamp biomicroscopy, dilated fundus examination, and time-domain optical coherence tomography (Stratus OCT, Carl Zeiss Meditec, Dublin, CA). Participants were retreated with monthly pro re nata (PRN) intravitreal 0.5 mg ranibizumab (Lucentis, Genentech, South San Francisco, CA) if they met predefined retreatment criteria, including any of the following: a 100-mm increase in central subfield thickness from the lowest previous OCT measurement; new or increased macular hemorrhage documented by fundus photographs; or a >5 letter decrease in BCVA since the last visit or the baseline BCVA, with disease activity such as persistent or increased fluid on OCT or leakage on fluorescein angiography (FA). After the first year of protocol-mandated follow-up and treatment, participants reverted to their standard care, with the criteria for retreatment determined by the attending clinician. Participants returned for mandated study safety visits at 18 and 24 months (with a further safety visit planned for 36 months). At the month 24 visit, participants underwent full ophthalmic examination, ETDRS BCVA, OCT, fundus photography, and FA using trial-certified equipment and staff. Images were assessed by the same independent reading center.
Outcome Measures The primary outcome was the number of PRN ranibizumab injections administered over 52 weeks. Year 1 secondary outcomes included the change in mean BCVA; the proportion of participants losing <15 letters, gaining 0 letters, and gaining 15 letters; and the change in mean total lesion area and mean CNV area based on fundus photographs and FA, assessed by a masked, independent reading center.13 These outcomes were also assessed at year 2. In addition, the reading center assessed the change in OCT central subfield thickness at 1 and 2 years. Safety outcomes included adverse events (AEs) and serious AEs. The reading center specifically examined for any MVAs that might be due to radiation. To increase the likelihood of detecting radiationinduced changes, the graders reported all MVAs even if these were within the area occupied by the neovascular lesion and therefore might be due to AMD rather than radiation. Graders were masked to study arm and timing (before or after baseline treatment/sham).
Subgroup Analysis An exploratory subgroup analysis was conducted to determine which baseline variables influence the response to SRT. A prior analysis of the year 1 data showed that lesions that could fit within the 4-mm treatment zone responded better than lesions extending beyond the zone, as were lesions with significant macular fluid at the time of treatment.15 These analyses were repeated at year 2. Group definitions were lesions 4 mm in greatest linear dimension versus those >4 mm, and lesions with an OCT macular volume greater than the median value of 7.4 mm3 versus lesions 7.4 mm3. The greatest linear dimension and macular volume were determined by the reading center. Several other subgroup analyses were undertaken to determine which variables influence the response to SRT, as undertaken at year 1.15
139
Ophthalmology
Volume 122, Number 1, January 2015
Table 1. Number of Pro Re Nata AntieVascular Endothelial Growth Factor Injections 16 Gray (N [ 75)
24 Gray (N [ 75)
16 Gray and 24 Gray Combined (N [ 150)
Sham (N [ 80)
Follow-up
Mean
SD
P
Mean
SD
P
Mean
SD
P
Mean
SD
Week 52 Week 104
2.6 4.5
2.5 3.9
0.009 0.008
2.4 5.4
2.4 4.6
0.003 0.09
2.5 4.9
2.4 4.3
0.0008 0.009
3.7 6.6
2.6 4.3
SD ¼ standard deviation.
Statistical Analysis
Visual Acuity
The statistical approach to the year 1 end points has been described.13 The primary group for the efficacy analysis was the intention-to-treat population, which consisted of all randomized patients. All efficacy analyses were performed by randomized arm assignment. The analysis population for safety consisted of all treated patients. All safety analyses were addressed by actual treatment received. The year 2 analyses were primarily exploratory, but were planned to mirror those of year 1. Specifically, each treated cohort was compared with sham with respect to the number of ranibizumab injections up to and including month 24 by comparing the means calculated by an analysis of variance adjusted for the number (3 vs. >3) and type of previous injections (ranibizumab vs. bevacizumab vs. other), whether the patient exhibited a dry macula after previous anti-VEGF therapy, time since diagnosis of wet AMD (<6 vs. 6 months), and baseline VA score (<54 vs. 54 letters). In addition, the analysis was adjusted for the minimization algorithm (protocol version B vs. CþDþE vs. F). The change in mean BCVA in each SRT arm compared with that in the pooled sham arms from baseline to 52 weeks was analyzed using a restricted maximum likelihoodebased repeatedmeasures approach. Data from all scheduled post-baseline visits, up to and including the outcome visit (24 months), were used in the analysis. The model included the same stratification factors as for the number of injections.
As previously reported, at the end of year 1, the mean change from baseline in VA was 0.28 (SD, 8.77) letters in the 16 Gy arm, þ0.40 (SD, 10.33) letters in the 24 Gy arm, and 1.57 (SD, 11.90) letters in the sham arm (P ¼ 0.73 and P ¼ 0.19, 16 Gy and 24 Gy compared with sham, respectively).13 At year 2, the mean BCVA was 48.3 letters (SD, 18.1) in the 16 Gy group, 52.5 letters (SD, 17.7) in the 24 Gy group, and 51.1 letters (SD, 20.1) in the sham group. The mean change in BCVA from baseline to 2 years was 10.0 letters (SD, 16.5) in the 16 Gy group, 7.5 letters (SD, 16.5) in the 24 Gy group, and 6.7 letters (SD, 17.1) in the sham group. If the 16 Gy and 24 Gy arms were combined, the mean VA change was 8.8 letters (SD, 16.5). This difference was not statistically significant. The proportions of participants losing <15 letters in the 16 Gy, 24 Gy, and sham arms were 68% (N ¼ 46), 75% (N ¼ 51), and 79% (N ¼ 58), respectively. The proportions of participants gaining 0 letters were 32% (N ¼ 22), 43% (N ¼ 29), and 38% (N ¼ 28), respectively. The proportions of participants gaining 15 letters were 3% (N ¼ 2), 1% (N ¼ 1), and 3% (N ¼ 2).
Results
Fluorescein Angiography, Fundus Photography, and Optical Coherence Tomography Mean total active lesion area increased by 1.0 mm2 (SD, 5.7), 4.2 mm2 (SD, 8.1), and 2.7 mm2 (SD, 8.6) for the 16 Gy, 24 Gy, and sham arms, respectively, at 104 weeks. The mean CNV area decreased by 0.1 mm2 in all groups (SD, 0.3, 0.4, and 0.2 for the 16 Gy, 24 Gy, and sham arms, respectively). Mean OCT central subfield thickness at year 2 decreased by 66.1 mm (SD, 101.5), 55.4 mm (SD, 130.3), and 33.3 mm (SD, 112.7) in the 16 Gy, 24
Number of Anti-Vascular Endothelial Growth Factor Retreatments The arms were balanced in terms of baseline demographics.13 Of the 230 participants who enrolled, 212 (92%) remained in the study through the week 104 visit (Fig 2, available at www.aaojourna l.org). At enrollment, the mean time from diagnosis of wet AMD was 12.8 months (standard deviation [SD], 7.3) in the 16 Gy arm, 16.7 months (SD, 10.1) in the 24 Gy arm, and 16.2 months (SD, 13.6) in the sham arm. The mean number of anti-VEGF injections the patients had already received before enrollment was 4.99 (SD, 3.87) in the 16 Gy arm, 6.23 (SD, 4.76) in the 24 Gy arm, and 5.48 (SD, 4.19) in the sham arm. At week 104 inclusive, those in the combined 16 and 24 Gy arms had a highly significant 26% reduction in the number of PRN injections compared with the sham arm (1.6; 95% confidence interval, 2.7 to 0.4, P ¼ 0.009). The mean number of PRN anti-VEGF injections is shown in Table 1 and Figure 3. Both the 16 Gy and 24 Gy SRT arms received significantly fewer PRN ranibizumab treatments at 12 months compared with the sham arm.13 This difference remained significant at year 2 in the 16 Gray arm but was not significant in the 24 Gray arm (Table 1).
140
Figure 3. The proportion of participants receiving different numbers of pro re nata (PRN) intravitreal injections over 2 years. The combined stereotactic radiotherapy (SRT) arm is compared with the sham arm.
Jackson et al
Two-Year Results of the INTREPID Study
Table 2. Summary of Year 2 Adverse Events and Serious Adverse Events in the Study Eye No. of Patients No. of evaluable participants in year 2* Participants with AE AE in study eye Radiation-related AE Procedure-related AE Participants with SAE SAE in study eye Total number of AEs Total number of SAEs
16 Gray (N [ 74) 13 4 1 2
71 (18%) (6%) (1%) 0 (3%) 0 17 2
24 Gray (N [ 73) 14 6 1 2
69 (20%) (9%) (1%) 0 (3%) 0 22 2
Sham (N [ 79) 12 4
3 1
76 (16%) (5%) 0 0 (4%) (1%) 14 3
Total (N [ 226) 216 39 (18%) 14 (6%) 2 (<1%) 0 7 (3%) 1 (<1%) 53 7
AE ¼ adverse event; SAE ¼ serious adverse event. Note: All AEs with a start date after the week 52 visit date were included in the year 2 analysis. *All safety participants who had a visit after week 52.
Gy, and sham arms, respectively. The combined SRT group had a 61.1 mm (SD, 116.7) reduction in central subfield thickness.
Subgroup Analysis Lesion size and macular fluid volume both influenced the response to SRT. In the 60 eyes with lesions 4 mm and a macular volume greater than the median (26% of the trial population at week 104), the SRT-treated eyes (N ¼ 40) had 45% fewer anti-VEGF injections than comparable sham-treated eyes (N ¼ 20) (3.9 [SD, 3.5] for combined SRT arms vs. 7.1 [SD, 3.3]; P ¼ 0.001). The BCVA in these SRT-treated eyes was also 4.4 letters superior to comparable sham-treated eyes (1.7 [SD, 12.5] vs. 6.2 [SD, 14.6], P ¼ 0.24). In this subgroup, mean total active lesion area increased by 1.6 mm2 (SD, 4.0) and 2.7 mm2 (SD, 2.3) for the SRT and sham arms, respectively, at 104 weeks. The mean CNV area decreased by 0.5 mm2 (SD, 0.5) and 0.0 mm2 (SD, 0) in the SRT and sham arms, respectively. Mean decreases in OCT central subfield thickness at year 2 were 120.7 mm (SD, 112.7) for the SRT-treated eyes compared with 40.1 mm (SD, 96.0) in the sham group. Other baseline variables analyzed for their influence on treatment response are shown in Figure 4 (available at www.aao journal.org). The absence of fibrosis at the time of SRT treatment was associated with a significantly greater decrease in the number of PRN injections (mean 4.9 [SD, 3.9] in the SRT group vs. 6.9 [SD, 4.2] in the sham arm; P ¼ 0.004) and a significantly greater decrease in OCT central subfield thickness (mean 79.2 mm [SD, 118.9] for the SRT treated eyes compared with 15.9 mm [SD, 115.1] in the sham group; P ¼ 0.005).
Safety The numbers of AEs and serious AEs were similar across arms (Table 2). Table 3 (available at www.aaojournal.org) lists AEs occurring in the study eye during year 2. Per protocol, only “related” AEs were required to be reported after the first year of follow-up. Any FA or fundus photographs that the reading center identified as having possible MVAs were referred for review by expert reading center clinicians (U.C. and J.S.S.) who were masked to both study arm and timing (before or after SRT). The reading center identified 35 cases with MVAs. These were reviewed by the experts, and 18 of these were thought likely to be related to radiation. When subsequently unmasked, all 18 of these cases were found to be in the radiation arms (9 per arm). Of the remaining 17 cases that were not deemed to be radiation-related, 6 were in the sham group. Only 2 cases (1 from each SRT arm) were thought to
have radiation-related MVAs that were likely to affect vision, because all other cases were extrafoveal, with 10 of 18 located outside the 4-mm treatment zone (Fig 5, available at www.aao journal.org). Of the 2 participants with foveal MVAs, 1 lost 46 letters and required 13 PRN injections over 2 years, and the other lost 41 letters and required 8 PRN injections over 2 years. As with other cases of MVAs, the main feature was capillary nonperfusion. The experts agreed that the main cause of visual impairment in these 2 eyes was the underlying AMD, but that the foveal MVAs may have been contributory. Representative images are shown and described in Figure 6. At 2 years, the mean change in BCVA in the 18 SRT eyes with MVAs attributed to radiation was similar to SRT-treated eyes without MVA, at 9.0 letters (SD, 21.1) and 8.7 letters (SD, 15.7), respectively.
Discussion At 1 year, the INTREPID trial met its primary end point, showing that SRT significantly reduced the frequency of anti-VEGF injections that patients required to treat their neovascular AMD.13 This result was encouraging, but it did not indicate whether the benefit of SRT persisted in the longer term. In addition, ongoing safety review was needed, because radiation can induce MVAs more than 1 year after exposure.8e12 Our results confirmed that SRT produced a significant reduction in anti-VEGF requirement over 2 years. In the second year, some participants developed MVAs in response to radiation, but these seldom affected vision, because the majority occurred outside the fovea. Overall, SRT produced a reduction in anti-VEGF retreatment by one quarter over 2 years. In particular, 15% of SRT recipients required no anti-VEGF injections in the 2 years after SRT and 45% required 3 injections or fewer. For these patients, there is the possibility of reduced clinic attendance and reduced cumulative risk of repeated injections, such as endophthalmitis, increased intraocular pressure, macular atrophy, and arteriothrombotic events. Our subgroup analysis suggested that patients are best treated with SRT when they have appreciable macular fluid. Likewise, lesions that could fit within the 4-mm treatment zone responded better than larger lesions. For eyes with both characteristics (26% of participants), there was a highly significant 45% reduction in anti-VEGF retreatment and a
141
Ophthalmology
Volume 122, Number 1, January 2015
Figure 6. A1e4, A 70-year-old woman treated with 16 Gray, comparing baseline with year 2. The patient gained 1 letter of visual acuity (VA) (74 letters at baseline vs. 75 letters at year 2) and required 1 pro re nata [PRN]) injection over 2 years (at week 20). Her medical history included osteoporosis and hyperlipidemia. The arrow shows an area of telangiectasia and capillary nonperfusion due to radiation exposure (A4). B1e4, A 69-year-old woman treated with 16 Gray. This patient lost 25 letters over 2 years (67 letters at baseline vs. 42 letters at year 2) and required 8 PRN injections. Her medical history included hypothyroidism and hyperlipidemia. The fundus photograph at year 2 shows a retinal hemorrhage (B2), and the fluorescein angiogram shows an area of capillary nonperfusion (circled; B4). The expert panel concluded that the vision loss was due to disease progression.
trend for better VA compared with sham-treated eyes with the same characteristics. It seems logical that lesions within the treatment zone would do better than those beyond it, because the latter may be undertreated at their outer margin. Radiation is known to preferentially damage proliferating
142
cells, and it is possible that actively proliferating endothelial cells are more likely to leak fluid.14,16 Therefore, the findings of our subgroup analysis are biologically plausible and consistent with our findings at year 1.15 In addition, a prospectively defined analysis comparing eyes with baseline
Jackson et al
Two-Year Results of the INTREPID Study
Figure 6. (continued). C1e6, A 65-year-old woman, 1 of the 2 patients thought to have lost vision in part because of microvascular abnormalities. She was treated with 24 Gray and experienced a 46-letter loss over 2 years (67 letters at baseline vs. 21 letters at year 2). This patient required 13 PRN ranibizumab injections. Her medical history included hypertension and hyperlipidemia. Early (C3e4) and late-phase (C5e6) fluorescein angiograms are provided. Fundus photographs (C1e2) show disciform scarring secondary to wet age-related macular degeneration. The fluorescein angiogram shows telangiectasia and capillary dropout centered on the fovea, visible in the early phase (C4). A late-phase angiogram shows extensive staining due to the underlying disciform lesion (C6). The expert panel concluded that the vision loss was primarily due to disease progression but may have been affected by radiation.
fibrosis with those without fibrosis found the latter had significantly fewer anti-VEGF injections and a greater decrease in central subfield thickness. This finding suggests that SRT may be best used before the formation of macular scarring. Masked analysis by the reading center found MVAs in 19% of the SRT participants and 8% of controls. A subsequent expert analysis of the SRT arm found that 10% of this group had MVAs attributable to radiation. As shown in Figure 5 (available at www.aaojournal.org), all but 2 MVAs occurred outside the fovea and the majority occurred outside the 4-mm treatment zone. Microvascular abnormalities were generally not reported by investigators. This is because MVAs are better seen by careful scrutiny of color photographs and fluorescein angiograms by trained graders. Only 2 cases of MVA were detected in the clinical setting. All other cases were detected by the reading center. It remains possible that at least some of these relatively nonspecific
changes were due to other causes. For example, an analysis of fundus photographs and FAs from a large RCT of pegaptanib found that 14% of untreated eyes with neovascular AMD have MVAs.17 The CABERNET study of EMB found that 3% of participants developed MVAs over 2 years, but paradoxically these participants had better outcomes than those without MVAs.9,12 Our patients with MVAs had outcomes similar to those without MVAs. The retinal changes seen in response to low-voltage SRT for treatment of wet AMD therefore appear different than the changes associated with high-dose, mega-voltage treatment for ocular malignancy. Ongoing evaluation of images obtained at the 36-month follow-up has indicated regression of previously noted MVAs in a number of SRT-treated eyes (Moshfeghi DM. INTREPID study of stereotactic radiotherapy for wet age-related macular degeneration: efficacy and safety at 24 and 36 months. Presented at: Retina Specialty Day, American Academy of Ophthalmology Annual
143
Ophthalmology
Volume 122, Number 1, January 2015
Meeting, November 15, 2013, New Orleans, LA). Nonetheless, patients considering SRT should be warned that there is an approximately 1% risk of foveal MVAs over 2 years, and that long-term follow-up is required. Although the CABERNET study of EMB and the INTREPID study of SRT had similar safety outcomes, the efficacy results differed substantially.9,12,13 This apparent discrepancy may be due to a number of factors.14 The EMB device used a strontium source that delivered an exponentially reduced dose with increasing distance from the probe.18 Therefore, probe placement by the surgeon was critical, not only in positioning on the surface of the retina but also topographically in relation to the area of greatest disease activity, which for occult lesions can sometimes be difficult to define. By contrast, the 90% isodose of the SRT device extends for 4 mm before showing a rapid decline because of the highly collimated and tightly targeted X-ray beams (Fig 1, available at www.aaojournal.org). Consequently, if a lesion fits within the 4-mm treatment zone, then the entire lesion receives an appropriate dose. Also, the SRT device uses eye tracking software to ensure correct positioning, whereas the EMB device relies on surgical manipulation. Of note, SRT is nonsurgical and avoids vitrectomy. Animal studies suggest that vitrectomy reduces intravitreal anti-VEGF half-life up to 10-fold, and therefore any residual disease activity may be hard to control.19,20 Strengths of this study include its multicenter, randomized, double-masked, sham-controlled design, with independent reading of fundus photographs, FAs, and OCTs. Year 2 data are more useful than year 1 data in terms of safety, but radiation retinopathy can occur beyond 2 years, and a year 3 safety analysis is planned. The fact that participants returned to standard care between year 1 and year 2 study visits is both a strength and weakness, because retreatment and monitoring are likely to be less regimented and verifiable outside of a trial, but conversely, pragmatic data may be more generalizable to routine care. The retreatment criteria in year 1 were typical of those used at the time the INTREPID study was designed, but they were less aggressive than current “treat until dry” regimens.21 The year 2 local care retreatment criteria varied by site. These less stringent retreatment criteria may have reduced the magnitude of the benefit demonstrated by SRT if all arms received fewer injections than would be expected with current practice. Also, the time-domain OCT used in this study was less sensitive than current spectral-domain OCT, and this may have reduced the likelihood of detecting macular fluid and further reduced the number of injections in both arms. When this study commenced, most clinical trials used time-domain OCT because the protocols for spectral-domain image acquisition and grading were not yet well validated and established. However, clinicians remained masked to participants’ SRT status throughout the 2-year study period addressed in this article, and therefore bias could not have been introduced into anti-VEGF retreatment decision making. These facts strengthen the study’s findings, because significantly fewer injections were seen despite low injection rates in both arms. They also suggest that the proportional reduction in injection frequency may be more relevant than the absolute reduction. In
144
retrospect, it would have been logical to exclude lesions beyond the 4-mm treatment zone, because subgroup analysis suggests that better case selection would have strengthened the result. Although subgroup analyses are important in terms of future trial design and may help refine case selection, they are exploratory in nature and should be interpreted accordingly. The VA and OCT results of the radiation-treated arms did not differ significantly from the control group, but the study was not designed to establish if they were superior or noninferior. The VA results cannot be compared with trials of treatment-naïve patients, which have generally shown substantial visual gains at the start of antiVEGF therapy. By comparison, VA typically decreases as patients enter the chronic treatment phase, as seen in 3 of 4 arms transitioning from the first to second year of the Comparison of Age-Related Macular Degeneration Treatments Trials (CATT).21,22 In conclusion, this study found that SRT was associated with significantly reduced anti-VEGF retreatment over 2 years. Microvascular changes may have occurred in response to radiation, but they seldom affected vision.
References 1. Owen CG, Jarrar Z, Wormald R, et al. The estimated prevalence and incidence of late stage age-related macular degeneration in the UK. Br J Ophthalmol 2012;96:752–6. 2. Munoz B, West SK, Rubin GS, et al. SEE Study Team. Causes of blindness and visual impairment in a population of older Americans: the Salisbury Eye Evaluation Study. Arch Ophthalmol 2000;118:819–25. 3. Finger RP, Fimmers R, Holz FG, Scholl HP. Prevalence and causes of registered blindness in the largest federal state of Germany. Br J Ophthalmol 2011;95:1061–7. 4. Bunce C, Xing W, Wormald R. Causes of blind and partial sight certifications in England and Wales: April 2007-March 2008. Eye (Lond) 2010;24:1692–9. 5. Heier JS, Brown DM, Chong V, et al; VIEW 1 and VIEW 2 Study Groups. Intravitreal aflibercept (VEGF Trap-Eye) in wet age-related macular degeneration. Ophthalmology 2012;119: 2537–48. 6. Rosenfeld PJ, Brown DM, Heier JS, et al; MARINA Study Group. Ranibizumab for neovascular age-related macular degeneration. N Engl J Med 2006;355:1419–31. 7. Chakravarthy U, Houston RF, Archer DB. Treatment of agerelated subfoveal neovascular membranes by teletherapy: a pilot study. Br J Ophthalmol 1993;77:265–73. 8. Petrarca R, Dugel PU, Nau J, et al. Macular epiretinal brachytherapy in treated age-related macular degeneration (MERITAGE): month 12 optical coherence tomography and fluorescein angiography. Ophthalmology 2013;120:328–33. 9. Jackson TL, Dugel PU, Bebchuk JD, et al; CABERNET Study Group. Epimacular brachytherapy for neovascular age-related macular degeneration (CABERNET): fluorescein angiography and optical coherence tomography. Ophthalmology 2013;120: 1597–603. 10. Petrarca R, Dugel PU, Bennett M, et al. Macular epiretinal brachytherapy in treated age-related macular degeneration (MERITAGE): month 24 safety and efficacy results. Retina 2014;34:874–9. 11. Dugel PU, Petrarca R, Bennett M, et al. Macular epiretinal brachytherapy in treated age-related macular degeneration:
Jackson et al
12.
13.
14. 15.
16.
Two-Year Results of the INTREPID Study
MERITAGE Study: twelve-month safety and efficacy results. Ophthalmology 2012;119:1425–31. Dugel PU, Bebchuk JD, Nau J, et al; CABERNET Study Group. Epimacular brachytherapy for neovascular age-related macular degeneration: a randomized, controlled trial (CABERNET). Ophthalmology 2013;120:317–27. Jackson TL, Chakravarthy U, Kaiser PK, et al; INTREPID Study Group. Stereotactic radiotherapy for neovascular agerelated macular degeneration: 52-week safety and efficacy results of the INTREPID Study. Ophthalmology 2013;120: 1893–900. Petrarca R, Jackson TL. Radiation therapy for neovascular age-related macular degeneration. Clin Ophthalmol 2011;5: 57–63. Jackson TL, Shusterman EM, Anoldussen, et al. Stereotactic radiotherapy for wet age-related macular degeneration (INTREPID): influence of baseline characteristics on clinical response. Retina August 6, 2014 [Epub ahead of print]. Kirwan JF, Constable PH, Murdoch IE, Khaw PT. Beta irradiation: new uses for an old treatment: a review. Eye (Lond) 2003;17:207–15.
17. Jackson TL, Danis RP, Goldbaum M, et al. Retinal vascular abnormalities in neovascular age-related macular degeneration. Retina 2014;34:568–75. 18. Petrarca R, Richardson M, Douiri A, et al. Safety testing of epimacular brachytherapy with microperimetry and indocyanine green angiography: 12-month results. Retina 2013;33: 1232–40. 19. Christoforidis JB, Xie Z, Jiang A, et al. Serum levels of intravitreal bevacizumab after vitrectomy, lensectomy and non-surgical controls. Curr Eye Res 2013;38:761–6. 20. Kakinoki M, Sawada O, Sawada T, et al. Effect of vitrectomy on aqueous VEGF concentration and pharmacokinetics of bevacizumab in macaque monkeys. Invest Ophthalmol Vis Sci 2012;53:5877–80. 21. CATT Research Group. Ranibizumab and bevacizumab for neovascular age-related macular degeneration. N Engl J Med 2011;364:1897–908. 22. Comparison of Age-related Macular Degeneration Treatments Trials (CATT) Research Group, Martin DF, Maguire MG, Fine SL, et al. Ranibizumab and bevacizumab for treatment of neovascular age-related macular degeneration: two-year results. Ophthalmology 2012;119:1388–98.
Footnotes and Financial Disclosures Originally received: February 8, 2014. Final revision: March 11, 2014. Accepted: July 29, 2014. Available online: September 7, 2014. 1
J.S.S.: Research Grant Support from Oraya. D.M.M.: stock options. Manuscript no. 2014-224.
School of Medicine, King’s College London, London, United Kingdom.
2
Queen’s University of Belfast, Consultant in Ophthalmology, The Belfast Health and Social Care Trust, Belfast, Northern Ireland.
3
Digital Angiography Reading Center, Great Neck, New York.
4
Centre for Vision and Vascular Science, Queen’s University of Belfast, Belfast, Northern Ireland. 5
Oraya Therapeutics, Inc, Newark, California.
6
Department of Surgery, Stanford University, Stanford, California.
7
The International Drug Development Institute, Louvain-la-Neuve, Belgium.
A.M.: Employee of the Centre for Vision and Vascular Science, which was retained by Oraya Therapeutics as a reading center for the INTREPID trial. M.A.: Employee of Oraya Therapeutics. E.M.S. and D.O.: Former employees and current consultants to Oraya Therapeutics and have stock options. M.E.G.: Founder and board member of Oraya Therapeutics and holds significant equity. L.D.: Employee of the International Drug Development Institute, which was retained by Oraya Therapeutics to conduct data management and statistical analysis of the INTREPID trial. Oraya Therapeutics participated in the design of the study, conducting the study, data collection, data management, data analysis, and review of the manuscript.
8
Byers Eye Institute, Horngren Family Vitreoretinal Center, Department of Ophthalmology, Stanford University School of Medicine, Palo Alto, California. *A full listing of the INTREPID Study Group is available in Appendix 1 (available at www.aaojournal.org).
Presented at: EURETINA, September 26e27, 2013, Hamburg, Germany. Financial Disclosure(s): The author(s) have made the following disclosure(s): T.L.J.: Employer received research funding from Oraya, and received a single advisory board payment and travel support from Oraya. U.C., J.S.S., A.M., D.M.M.: Consultants to Oraya Therapeutics.
Abbreviations and Acronyms: AE ¼ adverse event; AMD ¼ age-related macular degeneration; BCVA ¼ best-corrected visual acuity; CNV ¼ choroidal neovascularization; EMB ¼ epimacular brachytherapy; FA ¼ fluorescein angiography; MVA ¼ microvascular abnormality; OCT ¼ optical coherence tomography; PRN ¼ pro re nata; RCT ¼ randomized controlled trial; SD ¼ standard deviation; SRT ¼ stereotactic radiotherapy; VA ¼ visual acuity; VEGF ¼ vascular endothelial growth factor. Correspondence: Timothy L. Jackson, PhD, FRCOphth, School of Medicine, King’s College London, Department of Ophthalmology, King’s College Hospital, London SE5 9RS, UK. E-mail: [email protected].
145