|Home | About | Journals | Submit | Contact Us | Français|
To study the long-term visual and anatomic outcomes of anti-vascular endothelial growth factor (VEGF) monotherapy for the treatment of patients with retinal angiomatous proliferation (RAP).
Retrospective review of patients who were diagnosed with AMD and RAP lesions, and who received anti-VEGF injections as the only mode of therapy.
20 eyes (15 patients; 9 women, 6 men) with RAP lesions treated by anti- VEGF were encountered. Mean patient age was 85.8 years (SD ± 4.54). Nine eyes were treated with intravitreal ranibizumab alone, 8 eyes with bevacizumab alone, and 3 eyes received both drugs. At the 1, 3 and 6 month follow-up the median VA had improved from baseline (20/72) to 20/52, (range: 20/25 to 20/400), 20/45 (range: 20/20 to 20/400), and 20/56 (range: 20/20 to 20/400), respectively, (P> 0.001, P= 0.001, and P= 0.05, respectively). At 24 month follow-up, the improvement of VA, defined as halving of the visual angle, occurred in 37.5% of the cases.
Anti-VEGF monotherapy represents a useful treatment option for RAP, with stable or improved visual acuity in 62.5% of patients at 2 years. 25% of eyes only required a single injection, however, in most cases (75%) repeated treatments were required, highlighting the need for long term follow up. Although in this small study, the results for visual improvement were not statistically significant beyond 3 months; our findings warrant further large-scale investigation.
Yannuzzi and associates1 coined the term retinal angiomatous proliferation (RAP) to describe a variant of neovascular age-related macular degeneration (AMD) characterized by neovascularization with a presumed retinal origin. They hypothesized that RAP begins as intraretinal neovascularization (stage I), progresses to subretinal neovascularization with or without serous pigment epithelial detachment (PED) (stage II), and ultimately results in choroidal neovascularization (CNV) (stage III).1 A recent histopathologic study has confirmed the intraretinal origin of the neovascular process in RAP lesions. 2
RAP is estimated to occur in 10%–15% of newly diagnosed cases of neovascular AMD, RAP lesions are frequently bilateral, and within a three-year period, all patients with unilateral RAP lesions developed bilateral disease.3 The natural history and response to therapy in RAP appears distinct from other neovascular AMD subtypes.4,5 Indeed, a number of treatments for neovascular AMD including photodynamic therapy (PDT) have not proven effective in the treatment of RAP. 4,6
The purpose of the current study was to determine longer-term visual and anatomic outcomes of anti-VEGF monotherapy in patients with RAP.
We performed a retrospective review of medical records of all patients who underwent anti-VEGF treatment for AMD at the Doheny Eye Institute between August 1, 2005, and March 31, 2009‥ Patients were excluded if they had CNV secondary to conditions other than AMD if they received treatment other than anti-VEGF agents or had less than 6 months of follow-up.
Patients were included if they had AMD with typical manifestations of RAP, including preretinal, intraretinal, and/or subretinal hemorrhages, intraretinal edema, and the presence of neovascularization in one of the three stages of the disorder as confirmed by fluorescein or indocyanine green angiography and optical coherence tomography (OCT). At baseline, and at each monthly follow-up visit, best-corrected Snellen visual acuity (VA), and OCT were evaluated. Fluorescein angiography was performed at baseline and at physician’s discretion at subsequent visits. Intravitreal injection of either bevacizumab (1.25 mg) or ranibizumab (0.5 mg) was also performed at the discretion of the treating physician. No initial loading doses were used.
The decision for retreatment was made at each monthly follow-up visit as determined by OCT-based criteria. Specifically, patients were re-treated for persistence or recurrence of intraretinal, subretinal or sub-RPE fluid, or for any increase in retinal thickness.
Before 2007, all OCT images were obtained using Stratus OCT (Carl Zeiss Meditec Inc., Dublin, CA). With the availability of Fourier domain OCT (FDOCT), all OCT images were obtained using 3D-OCT 1000 (Topcon, Tokyo, Japan). The OCT images were qualitatively analyzed by two of the authors (TSH, PAK). The assessed parameters were subretinal fluid (SRF), intraretinal fluid (IRF), intraretinal cystoid spaces (ICS), and pigment epithelial detachment (PED). Intraretinal fluid included diffuse thickening of the retina as well as intraretinal microcystic edema. Intraretinal cystoids spaces were defined as distinct, round areas with well-demarcated borders and central hyporeflectivity.
Statistical analysis was performed using SAS V9 programming language (SAS Inst. Cary, NC). Snellen VA was converted to logMAR for the purposes of statistical analysis. Because VA was not obtained by protocol refraction, a high threshold was used to identify a change in vision. Accordingly, a decrease in VA was defined as doubling of the visual angle (≥0.3 change on logMAR). An increase in VA was defined as halving of the visual angle (≤ −0.3 on logMAR). If the VA changed by a smaller amount, it was classified as being stable. Paired t-tests were used to evaluate the mean changes in VAL from baseline. Decline in VA was further analysed using a Kaplan- Meier graph.
Twenty eyes of 15 patients were included in this study. Each eye had a diagnosis of AMD with a RAP lesion, treated with anti-VEGF monotherapy, with a documented follow- up of at least 6 months. Of the 15 patients, 9 were female and 6 were male, with a mean age of 85.8 years (SD ± 4.54). Five eyes had stage II & 15 eyes had stage III lesions. The mean baseline VA was 20/72 (range 20/30 to 30/2000).
Subretinal fluid was present in 13 eyes (65%), IRF in 18 eyes (90%), ICS in 15 eyes (75%) and PED in 17 eyes (85%). The PED was classified as a fibrovascular in 14 eyes (70%) and as a serous in 3 eyes (15%). In the fellow eyes, RAP was present in 11 eyes, occult CNV without RAP in 1 eye, dry AMD with atrophy in 4 eyes, and advanced dry AMD with geographic atrophy in 3 eyes.
Mean patient follow-up was 19.8 months (range 9–42 months). Nine eyes were treated exclusively with intravitreal ranibizumab; 8 eyes exclusively with bevacizumab; and 3 eyes received both ranibizumab and bevacizumab during the course of their follow-up. Fifteen of the 20 eyes (75%) required repeated injections throughout the follow-up period. For those 15 patients, the average frequency of reinjection was 3.45 months (range: 2 injections to 35 injections; median 3). 5/20 eyes (25%) required a single intravitreal injection. In the eyes that required reinjections, the range was 2–35 injections, with a mean and median of 6 and 4 injections respectively. At the last follow up visit, eleven of the 20 eyes (55%) required reinjection.
There were no documented systemic or local adverse events during the follow- up period.
At the 1-, 3- and 6-month follow-up, the median VA had improved from baseline (20/72) to 20/52, (range: 20/25 to 20/400), 20/45 (range: 20/20 to 20/400), and 20/56 (range: 20/20 to 20/400), respectively. The degree of improvement was statistically significant for 1 and 3 months, but not for 6 months (P> 0.001, P= 0.001, and P= 0.05, respectively, paired t- test).
At the 12-, 18- and 24-month follow-up, available for 15, 9 and 8 eyes, respectively, the median acuity was 20/85 (range: 20/40 to 20/400), 20/94 (range: 20/40 to 20/400), and 20/109 (range: 20/60 to 20/400). Compared to baseline, these VA were not statistically significant (P= 0.7, P= 0.81, P= 0.91; paired t- test).
At 12 and 24 months, improvement of VA was seen in 40% and 37.5%, stable VA in 33.3% and 25% & decrease in VA in 26.7%, 37.5% of eyes, respectively.
At the 30-, 36- and 42-month follow-up, available for 4, 2 and 1 eyes, respectively, the median acuity was 20/100 (range: 20/40 to 20/200) and 20/265 (range: 20/200 to 20/400). The change in vision was not statistically significant (P= 0.26 and P= 0.09, respectively, paired t- test). At 30-month follow-up, 1 eye (25%) showed improvement of vision, , and 3 eyes (75%) showed decreased vision. 2 eyes that completed 36 months follow-up showed decrease in vision. At the 42-month follow-up, available for only a single eye, VA was stable from baseline (20/100).
Kaplan-Meier cumulative survival curve (Figure 2) was used to display the change in vision over time. Five patients experienced a 3-line decrease in vision (deemed to be a “treatment failure”) during the follow-up period‥ The cumulative survival at 6, 9, 12, and 30 months was 90%, 85%, 78% and 38%, respectively.
At the last follow up, decreased VA was attributable to disciform scar in 1, geographic atrophy in 3 and persistent SRF and IRF in 1 eye (receiving treatment).
Table 1 shows changes in OCT features between baseline and the last follow-up visit. At the last visit SRF was present in 4 eyes (20%), IRF in 12 eyes (60%), ICS in 11 eyes (55%) and PED in 15 eyes (75%).
An 81-year-old female presented with RAP in her left eye. She had a history of wet AMD with PED in the other eye. She received a single intra-vitreal bevacizumab injection. Her VA improved from 20/200 to 20/70, and she maintained that improvement for the entire follow-up period of 27 months. Preinjection fundus photographs, fluorescein angiogram, and OCT show a type III RAP with a serous PED, SRF, and cystoid macular edema (Figure 3, A–C). One month after injection, the PED and cystoids changes resolved, which was maintained throughout the follow-up period. (Figure 3,D)
An 89-year-old man presented with RAP lesion in his right eye. He had a history of RAP lesion in the left eye treated with both PDT and intravitreal ranibizumab. He received 6 intravitreal ranibizumab injections over 42 months of follow-up. His VA was 20/60 at baseline and 20/100 at his last follow-up visit His preinjection fundus photograph, fluorescein angiogram, and OCT showed IRF with PED (Figure 4, A–B). At his last visit, OCT showed resolution of IRF with retinal thinning and RPE atrophy, consistent with progression to geographic atrophy, which explains the decreased vision. (Figure 4,C).
In this retrospective review, eyes treated with intravitreal anti- VEGF for RAP had anatomic and visual improvement, which was maintained up to 2 years of follow up. All eyes (100%) had improved or stable VA at months 1 and 3; this figure declined to 62.5% at 2 years of follow-up. Similar to the results reported by others,12–16 we found that eyes treated with either intravitreal bevacizumab 12–15 or ranibizumab 16 for RAP had significant short term functional benefit. Meyerle et al 12 showed that only 5.9% of eyes experienced decreased vision during the first 3 months of follow-up. Contrary to their results, our patients did not manifest any decrease in VA during the first 3 months.
We sought to compare our results to the analysis performed by the American Academy of Ophthalmology17 of different anti-VEGF treatment trials for AMD, none of which, to our knowledge, excluded RAP lesions. Table 2 summarizes the comparison, which shows similar visual results in our series. This comparison has to be taken in light of the differences in study design between our retrospective review where treatment decisions based on OCT, compared to MARINA and PIER, prospective studies, where monthly injections were performed regardless of OCT or VA. With these limitations in mind, our outcomes suggest that treatment based on OCT criteria (without initial loading dose), as performed in this study, can achieve acceptable visual and anatomic outcomes in RAP.
An interesting finding in our study was that 25% of eyes required a single treatment (see case 1). We were able to demonstrate this finding because our approach did not necessitate 3 initial loading injections, and as shown in that case, the patient developed geographic atrophy without recurrence over the course of subsequent 42 months. With the exception of the initial loading dose, our approach is similar to PrONTO. 18 Over 24 months, median frequency of injections in our RAP cohort was 3 months, compared to 2 months in PrONTO patients with RAP (n=10). Reassuringly, the decreased frequency did not correlate with a decrease in outcome, with similar VAL outcome in the total AMD population in PrONTO at 24 months to our population (3 lines or more VA improvement in 43% compared to 37.5 % in our population).
Our functional results are similar to those of Rouvas et al, 19 who compared intravitreal ranibizumab, intravitreal ranibizumab with PDT, and intravitreal triamcinolone with PDT for the treatment of RAP. Anti-VEGF monotherapy was associated with decrease in VA in 38.46% of their patients (26.7% in our group) at 12 months of follow-up
Since our study included patients who underwent FDOCT and time-domain OCT during their follow-up, we decided to study qualitative OCT parameters at baseline and last visit. Forooghian et al, 20 showed a significant difference between FDOCT and time-domain OCT in calculating macular thickness and volume. At baseline, 90% of our cohort had IRF, 75% ICS; and 65% had SRF. At the last visit, 60% of eyes had IRF, 55% ICS and only 20% SRF. Minimal improvement of PED was noted, with a decrease from 85% at baseline to 75% at the last visit. These results suggest that the fluid did not completely resolve, with the least effect on the PED. Our findings were generally biased toward detection of more fluid at the final visit since the last visit OCT was FDOCT with higher resolution and sensitivity. Keane et al 21 performed a quantitative sub-analysis of OCT after treatment with ranibizumab for neovascular AMD using software for manual segmentation and grading (OCTOR, available at www.diesel.la). They similarly showed that, although neurosensory retinal edema and SRF showed an early reduction to nadir after the initiation of ranibizumab therapy, the effect was attenuated over time.
RAP lesions are considered “high-flow,”.22 with histopathologic evidence of multiple growth factors, suggesting an ischemic and inflammatory component to the development and progression of RAP.23 Anti-VEGF monotherapy targets only one component of RAP lesions with early improvement of vision. Long term results suggest that anti-VEGF alone may not be sufficient to maintain anatomic and functional improvement.
This review has several limitations related to the retrospective nature, small sample size, non standard protocol of VA measurements, use of both FD and time-domain OCT, and qualitative OCT assessment. Our OCT results were biased toward better detection of fluid at the final visit, with residual SRF in 20%, IRF in 60% and PED in 75% of eyes. The presence of fluid on follow up is an expected corollary of our treatment approach as compared to other approaches, where patients are continued on treatment while OCT is dry (treat and extend).24
Another limitation is the lack of eyes with stage 1 RAP lesions, which may be more responsive to anti-VEGF therapy. This may have biased our results towards a worse outcome, and may explain the long term decreased vision in 37.5% of patients followed to 2 years. In contrast, the long term functional outcome of untreated eyes with RAP appears dismal. Viola et al reported that most patients presenting with Stage I (6 of 9) and all of Stage II rapidly progressed to visual loss within 6 months. In 13 eyes (81%), VA deteriorated by 2 or more lines at 6 months.25
To our knowledge, this is the longest follow- up study for RAP lesions treated with anti-VEGF monotherapy. Overall, our findings suggest that OCT-based anti- VEGF monotherapy for RAP represents an acceptable treatment approach and questions the need for 3 loading doses of anti-VEGF. Prospective studies with larger sample size, exploring various combination therapies and perhaps more aggressive anti-VEGF treatment approaches may help resolve the controversies surrounding RAP.
Dr. Fawzi receives grant support from Optovue, Inc and Reichert, Inc, unrelated to the articles subject matter. Dr. Sadda is co-inventor of Doheny intellectual property related to optical coherence tomography that has been licensed by Topcon Medical Systems and has served as a consultant for Heidelberg Engineering. However, it is not related to the article’s subject matter.
Supported in part by National Institutes of Health Grant EY03040 and National Eye Institute Grant R01 EY014375
Competing Interest: None Declared.
The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, an exclusive licence (or non exclusive for government employees) on a worldwide basis to the BMJ Publishing Group Ltd to permit this article (if accepted) to be published in BJO and any other BMJPGL products and sublicences such use and exploit all subsidiary rights, as set out in our licence (http://group.bmj.com/products/journals/instructions-for-authors/licence-forms).