Combination Therapy for AMD: Approaching a Solution From Different Angles
SHANTAN REDDY, MD, MPH. HOWARD F. FINE, MD, MHSc
Age-related macular degeneration (AMD) is the most common cause of blindness in developed countries.1 It is a complex disease, the pathogenesis of which is multifactorial, with demographic risk factors such as age and race, environmental risk factors such as diet and smoking, and genetic risk factors such as complement factor H affecting the risk of advanced disease. Our understanding of AMD at the cellular level is rapidly increasing. Mechanisms include chronic inflammatory insults, formation of an abnormal extracellular matrix with lipid accumulation, and oxidative stress to the retinal pigment epithelium (RPE) and choriocapillaris. Vascular endothelial growth factor (VEGF), a potent agonist of vascular permeability and angiogenesis, plays a major role.2-8 Our improved understanding is leading to more effective and more targeted therapies.
The first study to have a major impact on the management of AMD was the Macular Photocoagulation Study (MPS), which described the devastating natural course of the disease and offered thermal laser photocoagulation as a treatment option through the destruction of the neovascular complex along with surrounding structures.9 Patients treated with thermal laser lost a mean of 5.2 lines over 5 years, and untreated patients lost significantly more (7.1 lines). While the MPS demonstrated an effective therapy for extrafoveal choroidal neovascularization (CNV), visual results for subfoveal disease were not acceptable to patients or to their physicians.
|Shantan Reddy, MD, MPH, practices ophthalmology at the New York University Medical Center. Howard F. Fine, MD, MHSc practices ophthalmology with Vitreous-Retina-Macula Consultants of New York and the LuEsther T. Mertz Retina Research Center of the Manhattan Eye, Ear, and Throat Hospital. Neither author has any financial interest in any product mentioned in this article. Dr. Fine may be reached at email@example.com.|
After its approval by the Food and Drug Administration (FDA) in 2000, photodynamic therapy (PDT) with verteporfin (Visudyne, QLT/Novartis) emerged as a welcome alternative to thermal laser for the treatment of CNV. The Treatment of Age-related macular degeneration with Photodynamic therapy (TAP) study reported that 62% of treated eyes vs 46.4% of controls lost fewer than 3 lines after 12 months.10 The Verteporfin in Photodynamic Therapy (VIP) study found that 55% of the treatment group, compared to 68% of the control group, lost more than 3 lines after 2 years.11 Patients required, on average, more than 3 treatments during the first year and more than 2 treatments in the second year. Side effects of PDT include self-limited back or chest pain, acute severe visual decrease in 0.8% to 5% of patients in the first 7 days after treatment (which resolved roughly half the time), rip of the RPE, and even choroidal infarction.12-14
Even with PDT, however, a significant proportion of patients lose vision from AMD. Clearly PDT alone will not cure AMD. One explanation for the relatively high CNV recurrence rate may be that inflammatory cells, which are identifiable components of existing neovascular membranes, are not targeted by PDT alone and may even be induced by such therapy through collateral damage to the adjacent choriocapillaris and RPE. Though there is an initial regression of the neovascular lesion after PDT, inflammatory cells can induce an upregulation of VEGF.4-8 In short, PDT targets the vascular complex but fails to address the inflammatory components of CNV. In addition, nonselective damage to surrounding structures, such as choroid and pigment epithelium, may cause an upregulation of VEGF.15,16
A new approach has emerged in the battle against AMD, as retina specialists, much like oncologists fighting cancer, have begun to use combinations of therapies, with the hope of blocking different steps in disease pathogenesis in an additive or synergistic fashion.
PDT + IVTA
Triamcinolone acetonide is a relatively insoluble steroid with a duration of visible intravitreal crystals of approximately 2 months and with a number of anti-inflammatory, antiangiogenic, and antipermeability properties.17 Intravitreal triamcinolone acetonide (IVTA) alone has caused a regression of CNV membranes in primate and rat models, has been used in pilot studies for the treatment of exudative AMD, and does not seem to cause ocular damage. Well-reported side effects include cataractogenesis and an increase in intraocular pressure (IOP), which occurs in roughly a third of patients.18-27
Several investigators began to independently investigate whether combining PDT with IVTA would offer a synergistic approach by causing an acute regression of the CNV while suppressing the associated inflammatory components (Table 1). These investigators administered PDT prior to IVTA. The rationale was to avoid clouding of the vitreous body, which would make PDT more difficult to perform. Independent groups led by Spaide and colleagues24 injected approximately 4 mg of triamcinolone, Augustin and Schmidt-Erfurth26 injected 25 mg, and Ruiz-Moreno et al.28 approximately 20 mg of triamcinolone.24-28
Spaide and associates24 described 26 patients with CNV who were treated with a combination of PDT and IVTA. Thirteen patients had been previously treated with PDT and 13 had not. IVTA was administered immediately after PDT in all cases. In the group of patients naïve to PDT, Spaide and associates noted a visual acuity (VA) improvement of 2.5 lines above baseline at 12 months, with patients receiving a mean of 1.24 treatments. In the previously treated PDT group, the VA improved by 0.44 lines on average, with a retreatment rate of 1.2 over the first 12 months. Rechtman and colleagues,25 in their series of 14 eyes, found that 7% gained ≥30 Early Treatment Diabetic Retinopathy Study (ETDRS) letters, 50% maintained stable vision, 14% lost 15 to 29 letters, and 29% lost ≥30 letters. The mean number of treatments during the first year was 2.57.
Augustin and Schmidt-Erfurth26 examined 184 consecutive patients with CNV secondary to AMD who were treated with PDT and IVTA. The median follow-up was 38.8 weeks, and 84% of their patients had subfoveal CNV, 10.3% had juxtafoveal CNV, and 9.2% had extrafoveal CNV. VA improved by a mean of 1.22 lines, and the mean number of required treatments was 1.21. Of their patients, 12.5% required 2 treatments, 3.26% required 3 treatments, and only 0.5% of patients required 4 treatments over the course of their study. Fackler and colleagues27 reviewed 24 patients treated with combination therapy who were followed for up to 12 months. They found that the change in VA was not statistically significant at any follow-up visit. Overall, 62% of their patients required only 1 PDT retreatment over the 12 months of the study. Riuz-Moreno and associates28 recently evaluated the efficacy of combination therapy in a prospective, nonrandomized study that included 30 consecutive patients. Fifteen patients had not been treated with PDT previously, and 15 patients had been unsuccessfully treated with PDT alone. The newly treated group improved an average of 0.7 lines; however, the group that was previously treated lost 0.7 lines over the first year. Their control group lost an average of 2.2 lines. The average number of retreatments during the first year for the newly treated, previously treated, and control groups were 1.6, 1.2, and 2.8, respectively.
Steroid-induced glaucoma was controllable in the majority of patients in these early pilot studies. Spaide et al24 reported that 38.5% of patients experienced a rise in IOP, which was controlled by medication. This was similar to that reported by Rechtman and associates25 (28.5%) and Augustin and Schmidt-Erfurth26 (25%). Fackler and colleagues27 reported only 19% of their patients developed a rise in IOP after IVTA. Ruiz-Moreno et al.28 reported that up to 57% of eyes developed a rise in IOP in the newly treated group and 50% in the previously treated group. No cases of endophthalmitis were reported from combination PDT and IVTA in these studies. Combining IVTA and PDT appears promising by improving vision irrespective of lesion type, causing a cessation of leakage on fluorescein angiography and reducing the number of retreatments. Adverse outcomes were treatable and did not lead to any devastating complications. However, these studies were all short-term, nonrandomized pilot studies with small sample sizes.
There are contrasting studies that do not support the combined use of PDT and IVTA. Ergun and associates29 recently found that after 12 months, 57.6% of their 60 eyes treated with PDT and IVTA lost more than 3 lines of vision. In addition, 33% developed a rise in IOP and 21.6% developed severe cataracts that required surgery. Michel and colleagues,30 in another study, found that 2 of 45 patients treated with combination therapy developed RPE tears. These studies suggest tempering combination therapy with a hint of caution.
Figure 1. Schematic of angiogenic pathway initiated by VEGF. After the endothelial cell is activated, basement membrane degradation occurs via proteases, followed by endothelial cell proliferation and migration, tube formation, and finally elongation and remodeling.
Therefore, combining IVTA with PDT appears advantageous over PDT alone by reducing the number of retreatments and decreasing the degree of visual loss by tackling the various pathways that lead to CNV. However, the side effects of the 2 approaches appear to be additive, and the evidence to support the combination approach is limited to small pilot studies. A large, randomized, controlled trial is underway to further evaluate therapy with PDT and IVTA: the Visudyne with Intravitreal Triamcinolone Acetonide (VisTA) trial, which compares 0 mg, 1 mg, and 4 mg of IVTA with PDT in patients with occult, with no classic, and with minimally classic subfoveal CNV. The results of the VisTA study are eagerly awaited to guide retina specialists in determining under what clinical circumstances PDT with IVTA is most appropriate.
PDT + ANTI-VEGF THERAPY
The introduction of specific agents targeting VEGF has revolutionized the treatment of AMD (Figure 1). Ranibizumab (Lucentis, Genentech) is a genetically engineered 49-kDa antibody fragment that blocks all forms of VEGF-A. Ranibizumab has been shown to inhibit neovascularization in both animal models and humans (Figure 2).31 A phase 2/3 multicenter, controlled clinical trial reported an unprecedented gain of 15 letters in 26% of subjects with exudative AMD, with an excellent safety profile for up to 6 months after intravitreal injections.32 Approval by the FDA of ranibizumab for the treatment of AMD has made this line of treatment popular among retinal specialists. The Minimally Classic/Occult Trail of the Anti-VEGF Antibody Ranibizumab in the Treatment of Neovascular Age-Related Macular Degeneration (MARINA) trial is a phase 3, prospective, randomized, placebo-controlled clinical trial that compared ranibizumab for the treatment of minimally classic or occult CNV versus sham injection.33
MARINA enrolled 716 patients with over 90% follow-up of patients over the first 12 months and 80% to 90% of patients over 24 months. Of patients treated with ranibizumab and followed for 24 months, 92% of those receiving 0.3 mg and 90% of those receiving 0.5 mg lost fewer than 15 letters from baseline, as compared to 52.9% of patients in the sham group. More impressive was that 24% of patients receiving 0.3 mg and 33% of patients receiving 0.5 mg gained 15 letters or more in VA. This is in comparison to a dismal 5% in the sham group. Overall, the incidence of any side effect, ocular or systemic, was similar among the groups. Another phase 3, prospective, double-blinded, randomized study, the Anti-VEGF Antibody for the Treatment of Predominantly Classic Choroidal Neovascularization in Age-Related Macular Degeneration (ANCHOR) study, evaluated the efficacy of monthly ranibizumab injections as compared to PDT with verteporfin.34 Of patients treated with 0.3 mg and 0.5 mg of ranibizumab, 94.3% and 96.4% lost fewer than 15 letters as compared, respectively, compared to 64.3% in the verteporfin group. Furthermore, 35.7% in the 0.3 mg group and 40.3% in the 0.5 mg group gained more than 15 letters as compared to only 5.6% in the verteporfin group. Again, there was no statistical difference in any adverse outcome between the 2 groups.
Figure 2. Avastin and Lucentis are both anti-VEGF-A biologics developed from the same antibody.
In summary, results from the MARINA and ANCHOR studies have demonstrated that ranibizumab monotherapy is an effective means to preserve and improve vision in patients with CNV with limited adverse events over 24 months. However, subjecting patients to frequent intraocular injections may increase the cumulative risk of adverse events, such as retinal detachment or endophthalmitis. It is reasonable to suggest that small differences in adverse events may not have been identified since the studies were not powered to detect them (MARINA n=716; ANCHOR n=423). The Phase 3b, Multicenter, Randomized, Double-Masked, Sham Injection Controlled Study of the Efficacy and Safety of Ranibizumab in Patients with Subfoveal CNV With or Without Classic CNV Secondary to AMD (PIER) and the Prospective Optical coherence tomography imaging of patients with Neovascular age-related macular degeneration Treated with intra-Ocular Lucentis (PrONTO) trials are currently evaluating whether less frequent injections are as effective as monthly injections. Preliminary results have shown that quarterly dosing was inferior to monthly dosing in that fewer patients gained VA when given an injection every 3 months.35 Simply reducing the number of injections does not appear to be the solution. We must therefore continue to explore other ways to limit the frequency of injections through various delivery devices and/or a combination of different treatment modalities.
Husain et al.36 showed that intravitreal ranibizumab injections in combination with PDT cause a greater reduction in angiogenic leakage than PDT alone in primate models. Kim and associates37 demonstrated that this combination does not produce any additional detrimental effects on the choroid or retina, as compared to PDT alone. In a murine model of CNV, PDT combined with pegaptanib sodium (Macugen, [OSI] Eyetech/Pfizer) the same day increased the efficacy over either drug alone.38 The RhuFab V2 Ocular Treatment Combining the Use of Verteporfin to Evaluate Safety (FOCUS) trial (stage 1/2, prospective, randomized) combined PDT with ranibizumab and showed 90% stabilization of vision despite many of these patients having had previous PDT treatments, and 23.8% of patients gained >15 letters, with only 1% losing more than 30 letters.39
Conversely, Avery and colleagues40 performed a subgroup analysis on patients who had received PDT before bevacizumab (Avastin, Genentech) and found no differences in response to treatment, though the timing of PDT and the number of PDT treatments were not reported. They suggested that this particular combination may be no more beneficial than bevacizumab alone. A recent article by Dhalla et al.41 looked at the combination of PDT and intravitreal bevacizumab for CNV in treatment of naïve eyes and had 7 months of follow-up. Twenty-four eyes were included, and 7 months later, 20 eyes (83%) were stable and16 eyes (67%) had an improvement of vision from baseline, with the mean improvement of 2.04 lines and 15 eyes (63%) required only 1 treatment. The fact remains that the efficacy of combining PDT with anti-VEGF injections in improving visual outcomes beyond anti-VEGF therapy alone is unknown and still needs to be explored.
PDT + ANTI-VEGF + STEROID
Recently, an article by Augustin and colleagues42 showed that triple therapy may allow an increase in vision with a decrease in the number of treatments. In this study, PDT with reduced light dose was administered. Sixteen hours later a limited vitrectomy was performed and dexamethasone (800 mg) and bevacizumab (1.25 mg) were administered. One hundred four patients followed for 40 weeks showed a mean increase in VA of 1.8 lines, with a decrease in retinal thickness of 182 μm. A previous study by Liggett et al.43 showed similar results when PDT was combined with IVTA and pegaptanib.
As we are gaining a better understanding about the multitude of factors that lead to CNV in AMD, we are discovering new lines of treatment that are producing unprecedented results. But they are not without their added costs to society and risks to the patient. For these reasons, we must continue to find ways reduce the number of treatments needed and improve on the vision gained. Many prospective studies are being undertaken to explore in a prospective, randomized fashion whether combination therapy using full or reduced fluence (which may improve PDT selectivity)44 combined with ranibizumab (with and without steroids) will produce an improvement over ranibizumab monotherapy. Through the exploration of these various combinations of therapy, we may achieve such a goal in the near future. RP
1. Mitchell P, Smith W, Attebo K, Wang JJ. Prevalence of age-related maculopathy in Australia: the Blue Mountain Eye Study. Ophthalmology. 1995;102:1450-1460.
2. Schmidt-Erfurth U, Miller J, et al. Photodynamic therapy of subfoveal choroidal neovascularization: clinical and angiographic examples. Graefes Arch Clin Exp Ophthalmol. 1998;136:365-374.
3. Schmidt-Erfurth U, Michels A, Barbazetto I, Laqua H. The photodynamic effects of choroidal neovascularization and physiological choroids. Invest Ophthalmol Vis Sci. 2002;43:830-841.
4. Klein RJ, Zeiss C, Chew EY, et al. Complement factor H polymorphism in age-related macular degeneration. Science. 2005;308:385-389.
5. Haines JL, Hauser MA, Schmidt S, et al. Complement factor H variant increases the risk of age-related macular degeneration. Science. 2005;308:419-421.
6. Dastgheib K, Green WR. Granulomatous reaction to Bruch's membrane in age-related macular degeneration. Arch Ophthalmol. 1994;112:813-818.
7. Killingsworth MC, Sarks JP, Sarks JH. Macrophages related to Bruch's membrane in age-related macular degeneration. Eye. 1990;4:613-621.
8. Oh H, Takagi H, Takagi C, et al. The potential angiogenic role of macrophages in the formation of choroidal neovascular membranes. Invest Ophthalmol Vis Sci. 1990;40:1891-1898.
9. Macular Photocoagulation Study Group. Argon laser photocoagulation for neovascular maculopathy. Five-year results from randomized clinical trials. Arch Ophthalmol. 1991;109:1109-1114.
10. Treatment of Age-Related Macular Degeneration with Photodynamic Therapy (TAP) Study Group. Photodynamic therapy of subfoveal choroidal neovascularization in age-related macular degeneration with verteporfin: one year results of 2 randomized clinical trials-TAP report 1. Arch Ophthalmol. 1999;117:1329-1345.
11. Verteporfin in Photodynamic Therapy Study Group. Verteporfin therapy of subfoveal choroidal neovascularization in age-related macular degeneration: two year-results of a randomized clinical trial including lesions with occult with no classic choroidal neovascularization-verteporfin in photodynamic therapy report 2. Am J Ophthalmol. 2001;131:541-560.
12. Arnold JJ, Blinder KJ, Bressler NM, et al; Treatment of Age-Related Macular Degeneration with Photodynamic Therapy Study Group; Verteporfin in Photodynamic Therapy Study Group. Acute severe visual acuity decrease after photodynamic therapy with verteporfin: case reports from randomized clinical trials-TAP and VIP report no. 3. Am J Ophthalmol. 2004;137:683-696.
13. Treatment of Age-Related Macular Degeneration with Photodynamic Therapy (TAP) Study Group. Verteporfin therapy for subfoveal choroidal neovascularization in age-related macular degeneration. Three year results of an open-label extension of 2 randomized clinical Trials –Tap report no.5 Arch Opthalmol. 2002;120:307-314.
14. Recchia FM, Greenbaum S, Recchia CA, Ruby AJ, Alldredge CD, Hassan TS.. Self-reported acute decrease in visual acuity after photodynamic therapy for age-related macular degeneration. Retina. 2006;26:1042-1048.
15. Tatar O, Shinda K, Adam A, et al Effect of verteporfin photodynamic therapy on endostatin and angiogenesis in human choroidal neovascular membranes. Br J Ophthalmol. 2007;91:166-173.
16. Tatar O, Adam A, Shinoda K, et al. Expression of VEGF and PEDF in choroidal neovascular membranes following verteporfin photodynamic therapy. Am J Ophthalmol. 2006;142:95-104.
17. Jonas JB, Hayler JK, Panda-Jonas S. Intravitreal injection of crystalline cortisone as adjunct treatment of proliferative vitreoretinopathy. Br J Ophthalmol. 2000;84:1064-1067.
18. Wilson CA, Berkowitz BA, Sato Y, et al. Treatment with intravitreal steroids reduces blood-retinal barrier breakdown due to retinal photocoagulation. Arch Ophthalmol. 1992;110:1155-1159.
19. Ishibashi T, Miki K Sorgente N, et al. Effects of intravitreal administration of steroids on experimental subretinal neovascularization in the subhuman primate. Arch Ophthalmol. 1985;103:708-711.
20. Ciulla TA, Criswell MH, Danis RP, Hill TE. Intravitreal triamcinolone acetonide inhibits choroidal neovascularization in a laser-treated rat model. Arch Ophthalmol. 2001;119:399-404.
21. Antoszyk AN, Gottlieb JL, Machemer R, Hatchell DL. The effects of intravitreal triamcinolone acetonide on experimental pre-retinal neovascularization. Graefes Arch Clin Exp Ophthalmol. 1993;231:34-40.
22. Penfold PL, Wen L, Madigan MC, et al. Modulation of permeability and adhesion molecule expression by human choroidal endothelial cells. Invest Ophthalmol Vis Sci. 2002;43:3125-3130.
23. McCuen BW II, Bessler M, Tano Y, et al. The lack of toxicity of intravitreally administered triamcinolone acetonide. Am J Ophthalmol. 1981;91:785-788.
24. Spaide RF, Sorenson J, Maranan L. Photodynamic therapy with verteporfin combined with intravitreal injection of triamcinolone acetonide for choroidal neovascularization. Ophthalmology. 2005;112:301-304.
25. Rechtman W, Danis RP, Pratt LM, Harris A. Intravitreal triamcinolone with photodynamic therapy for subfoveal choroidal neovascularization in age-related macular degeneration. Br J Ophthamol. 2005;88:344-347.
26. Augustin AJ, Schmidt-Erfurth U. Verteporfin therapy combined with intravitreal triamcinolone in all types of choroidal neovascularization due to age-related macular degeneration. Ophthalmology. 2006;113:14-22.
27. Fackler TK, Reddy S, Bearelly S, Stinnett S, Fekrat S, Cooney MJ. Retrospective review of eyes with neovascular age-related macular degeneration treated with photodynamic therapy with verteporfin and intravitreal triamcinolone. Ann Acad Med Singapore. 2006;35:710-715.
28. Ruiz-Moreno JM, Montero JA, Barile A, Zarbin MA. Photodynamic therapy and high-dose intravitreal triamcinolone to treat exudative age-related macular degeneration. 1-year outcome. Retina. 2006;26: 602-612.
29. Ergun E, Maar N, Ansari-Shahrezaei S, et al. Photodynamic therapy with verteporfin and intravitreal triamcinolone acetonide in the treatment of neovascular age-related macular degeneration. Am J Ophthalmol. 2006;142:10-16.
30. Michels S, Aue A, Simader C, et al. Retinal pigment epithelial tears following verteporfin therapy combined with intravitreal triamcinolone. Am J Ophthalmol. 2006;141:396-398.
31. Krzystolik MG, Afshari MA, Adamis AP, et al. Prevention of experimental choroidal neovascularization with intravitreal anti-vascular endothelial growth factor antibody fragment. Arch Ophthalmol. 2002;120:338-346.
32. Heier JS, Antoszyk AN, Pavan PR, et al. Ranibizumab for treatment of neovascular age-related macular degeneration: a phase I/II multicenter, controlled, multidose study. Ophthalmology. 2006;113:642.e1-4.
33. Rosenfeld PJ, Brown DM, Heier JS, et al. Ranibizumab for neovascular age-related macular degeneration. N Engl J Med. 2006;355:1419-1431.
34. Brown DM, Kaiser PK, Michels M, et al. Ranibizumab versus verteporfin for neovascular age-related macular degeneration. N Engl J Med. 2006;355:1432-1443.
35. Schmidt-Erfurth U. How often is re-treatment with Lucentis needed? Paper presented at: Joint Meeting of the American Academy of Ophthalmology and Asia Pacific Academy of Ophthalmology; November 11-14, 2006; Las Vegas, Nev.
36. Husain D, Kim I, Gauthier D, et al. Safety and efficacy of intravitreal injection of ranibizumab in combination with verteporfin PDT on experimental choroidal neovascularization in the monkey. Arch Ophthalmol. 2005;123:509-516.
37. Kim IK, Husain D, Michaud N, et al. Effect of intravitreal injection of ranibizumab in combination with verteporfin PDT in normal primate retina and choroid. Invest Ophthamol Vis Sci. 2006;47:357-363.
38. Ju M, Mailhos C, Ganley M, et al. Pegaptanib in combination with verteporfin-based PDT increases regression of murine ocular neovascularization. Invest Ophthalmol Vis Sci. 2006;47:E-Abstract 4175.
39. Heier JS, Boyer DS, Ciulla TA, et al; FOCUS Study Group. Ranibizumab combined with verteporfin photodynamictherapy in neovascular age-related macular degeneration:1 year results of the FOCUS study. Arch Ophthalmol. 2006;124:1532-1542.
40. Avery RL, Pieramici DJ, Rabena MD, Castellarin AA, Nasir MA, Giust MJ. Intravitreal bevacizumab (Avastin) for neovascular age-related macular degeneration. Ophthalmology. 2006; 113:363-372.e.5.
41. Dhalla MS, Shah GK, Blinder KJ, et al. Combined photodynamic therapy with verteporfin and intravitreal bevacizumab for choroidal neovascularization in age-related macular degeneration. Retina. 2006;988-993.
42. Augustin AJ, Puls S, Offerman I. Triple therapy for choroidal neovascularization due to age-related macular degeneration: veteporfin PDT, bevacizumab, and dexamethasone. Retina. 2007;27:133-140.
43. Liggett PE, Colna J, Chaudry NA, et al. Triple therapy of intravitreal triamcinolone, photodynamic therapy, and pegatanib sodium for choroidal neovascularization. Am J Ophthalmol. 2006;142:1072-1074.
44. Michels S, Hansmann F, Geitzenauer W, Scmidt-Erfurth U. Influence of treatment parameters on selectivity of verteporfin therapy. Invest Ophthalmol Vis Sci. 2006;47:371-376.