Influence of Treatment Parameters on Selectivity of Verteporfin Therapy

Abstract

purpose. To improve selectivity of verteporfin therapy (PDT) in neovascular age-related macular degeneration (AMD) using modified treatment parameters.

methods. Nineteen consecutive patients with predominantly classic choroidal neovascularization (CNV) in AMD were treated with 6 mg/m² verteporfin given as bolus infusion. Patients received PDT with a fluence of either 25 or 50 J/cm². Choroidal perfusion changes were evaluated by indocyanine green angiography (ICGA) at baseline, day 1, week 1, week 4, and month 3. Secondary outcomes were CNV closure rate and therapy-induced leakage documented by fluorescein angiography (FA). The safety of the treatment was assessed with ETDRS visual acuity.

results. Complete CNV closure was achieved in all patients at day 1. Choroidal hypoperfusion was minimal in eyes treated with a reduced fluence of 25 J/cm². Most patients treated with 50 J/cm² showed significant choriocapillary nonperfusion at week 1, lasting as long as 3 months. A transient PDT-induced increase in leakage area in FA at day 1 was found to be more extensive in the 50-J/cm² group.

conclusions. Bolus administration of verteporfin combined with a reduced light dose achieved improved selectivity of photodynamic effects, avoiding collateral alteration of the physiologic choroid while obtaining complete CNV closure. An increased selectivity with decreased effect on the surrounding choroid should be of advantage in verteporfin monotherapy as well as in combination strategies.

Verteporfin therapy (PDT) has been the standard of care for many patients with exudative age-related macular degeneration (AMD). Even though its effectiveness had been shown in several multicenter controlled clinical trials, ^{1 2} patients treated with verteporfin therapy still lose a mean of 2.2 lines of visual acuity over 12 months. ³ Seventy-five percent of vision loss is seen within the first 6 months and is usually associated with progressive growth of the choroidal neovascularization (CNV). An additional dilemma is the need for repeated retreatments during follow-up with an unknown pathogenic mechanism. 

Verteporfin therapy induces a nonthermal occlusion of the CNV lesion without affecting the overlying neurosensory retina, as shown experimentally. ⁴ Only minor effects on the retinal pigment epithelium (RPE) were documented after a single treatment in animal models. ⁴ However, definite dose-dependent choriocapillaris closure is seen histologically in human eyes 1 week after verteporfin treatment according to the standard treatment parameters. ^{5 6} These morphologic findings correlate well with characteristic indocyanine green angiography (ICGA) features documented after standard verteporfin therapy. ⁷ Choroidal hypofluorescence reaching its maximum at 1 week after treatment, comparable in size with the treatment spot used, is regularly documented by ICGA. At least partial reperfusion of the choriocapillary layer was seen at 3 months follow-up; however, repeated verteporfin therapy in patients treated in the Treatment of Age-Related Macular Degeneration with Photodynamic Therapy (TAP) trial led to persistent choriocapillary nonperfusion in most eyes. ⁸ Associated with choriocapillary occlusion, an angiogenic response with upregulation of vascular endothelial growth factor (VEGF) was also documented in human eyes. ⁹ The collateral impact of verteporfin therapy on the physiologic choroid strongly compromises the concept of selectivity and may be the reason for additional vision loss and the need for repeated intervention. 

Ideally a selective approach should include maximum efficacy (i.e., complete closure of the CNV) and minimal damage to the physiologic choroid (i.e., absence of nonperfusion or vascular–retinal pigment epithelium [RPE] barrier breakdown). Based on the principles of photodynamic mechanisms, there are distinct strategies to enhance selectivity, including modification of the route of administration, the timing of laser exposure, and a reduction of fluence and/or irradiance. Previous clinical trials have provided solid information about the range of safe and effective dosages, such as the phase I/II trials and the verteporfin treatment of subfoveal minimally classic CNV in age-related macular degeneration (VIM) study. ^{10 11 12} However, selectivity in terms of avoiding damage to the physiologic choroid was never evaluated. In a prospective case series, we chose a bolus infusion of verteporfin, which has been shown safe and effective in the treatment of choroidal hemangioma. ¹³ A bolus administration is generally used in PDT tumor therapy, to achieve an optimal and selective biodistribution of the sensitizer. ^{14 15} Bolus injection was also used in the pilot work in experimental photothrombosis of ocular vasculature and was not associated with significant choroidal damage. ⁴ A reduced light dose was selected based on the results of the phase I/II studies and the VIM study. ¹²  

The purpose of this study was to demonstrate proof of the principle that verteporfin therapy can achieve selectivity, defined as complete CNV closure, with reduced choroidal occlusion or barrier breakdown. 

Methods

The presented study is the result of a prospective interventional case series. The treatment protocol was approved by the local ethics committee. The study was conducted according to the tenets of the Declaration of Helsinki and patients gave written informed consent before enrolling, after a detailed discussion of the study procedures, potential risks, and goals. 

Patient Selection

Nineteen consecutive patients were enrolled in the study. All patients presented with a subfoveal, predominantly classic CNV due to AMD, without any prior treatment. Visual acuity ranged between 20/200 and 20/40. The total lesion size was measured to have a greatest linear diameter (GLD) of less than 5400 μm, and the CNV component had to cover at least 50% of the total lesion size, according to the TAP study recommendations. ¹  

Treatment Parameters

All patients received a bolus infusion of 6 mg/m² body surface area over 1 minute. Patients were assigned to two treatment protocols, using a fluence of either 25 or 50 J/cm². In each group, one additional parameter was modified. In the 25-J/cm² group, patients received an irradiance of either 600 or 300 mW. Depending on the irradiance, the time of photosensitization was 42 or 83 seconds. In the 50-J/cm² group timing of photosensitization was 5 or 15 minutes after verteporfin bolus infusion. 

Detailed treatment parameters and the treatment groups are shown in Table 1 . 

Documentation

Patients were seen for regular follow-up visits within 1 week before and at day 1, week 1, week 4, and month 3 after treatment. A standardized evaluation was performed at each visit including best corrected visual acuity according to the guidelines of the Early Treatment Diabetic Retinopathy Study (ETDRS), confocal scanning laser fluorescein angiography (FA), ICGA (Heidelberg Engineering, Dossenheim, Germany), fundus photography, and a complete eye examination. Selected patients were imaged with optical coherence tomography (OCT). 

The main outcome measures were choroidal perfusion changes, as documented by early and late ICGA. A PDT-induced increase in collateral leakage area seen by late FA 1 day after PDT was defined as a secondary outcome, as was primary CNV closure documented by early FA. Best-corrected visual acuity was documented for safety evaluation. 

Data were statistically analyzed with the Wilcoxon signed-rank and Wilcoxon rank sum tests. Statistical significance was defined as P < 0.05. 

Procedures for Evaluation

The Heidelberg Eye Explorer (software 1.0; Heidelberg Engineering), an imaging software developed for analysis and visualization of images obtained with the Heidelberg Retina Angiograph (HRA), was used for planimetric evaluation of the area of hypofluorescence detected by ICGA and the area of PDT-induced leakage seen on FA. Choriocapillary hypoperfusion and nonperfusion were graded according to a scale (Table 2) . Primary CNV closure was evaluated with FA 1 day after PDT. Angiographies were evaluated by two masked readers, and planimetric and grading results of both readers were averaged. There was 100% agreement on CNV closure at day 1. 

Results

PDT-Induced Choroidal Perfusion Changes

The effect of treatment parameters on choroidal perfusion was evaluated based on angiographic loss of perfused choriocapillary patterns during early ICGA (Table 2)and hypofluorescence in late ICGA. 

At day 1, a significant loss of early choriocapillary perfusion patterns was seen in early ICGA in the 50-J/cm² light dose regimen (Table 3 , Fig. 1D ). Two of three patients with photosensitization 5 minutes after the end of the verteporfin infusion showed nonperfusion of some larger choroidal vessels (grade IV). None of the protocols showed any effect at the level of the retinal circulation. An occlusive effect on the choriocapillary layer was much less evident in the 25-J/cm² treatment group (Table 3 , Fig. 2D ). On average, patients treated in this protocol showed only minor grades of malperfusion of the choriocapillaris (≤grade II). At day 1 high grades of malperfusion (≥grade III) were significantly more common in the 50 J/cm² group compared with the 25-J/cm² group (P = 0.0048). At the 1-week follow-up no patient showed more than grade II choroidal perfusion changes in the 25-J/cm² treatment group (Table 3) . However, a significant nonperfusion effect (≥grade III) was evident in 66.6% of patients treated with 50 J/cm² at 1 week (Table 3 , Fig. 1E ). Perfusion changes of at least grade III were significantly more frequent in the 50-J/cm² group than in the 25-J/cm² group (P = 0.0027). During further follow-up, all groups showed progressive recovery of the choriocapillaris. The 3-month follow-up examination did not document any moderate perfusion changes (grade II) in the 25-J/cm² protocol. However, especially in patients with early photosensitization in the 50-J/cm² group, at least moderate perfusion changes were present at 3 months (60% of patients; Table 3 ). 

Late hypofluorescence on ICGA provides information about the completeness of choroidal and choriocapillary perfusion. Figure 3summarizes the follow-up data on the area of late hypofluorescence. At baseline, focal perfusion changes were visible due to the underlying disease and were restricted to the lesion site. Verteporfin therapy induced an increase in the hypofluorescent area in late ICGA in all regimens at 1 day and 1 week; however, most pronounced changes were documented in the 50-J/cm² protocol (Fig. 1F) . The 25-J/cm² protocol showed recovery to pretreatment levels at 4 weeks and 3 months, whereas in the 50-J/cm² protocol, large areas with persistent hypofluorescence were detectable during follow-up (Fig. 3) . At month 3, the treatment-induced hypofluorescent area was significantly smaller in the 25-J/cm² group than in the 50-J/cm² group (P = 0.0033). 

CNV Closure and Collateral Vascular–RPE Barrier Breakdown

The change in size of the total leakage area, including leakage from the CNV and new leakage arising from the entire PDT-exposed area, was used as a measure of increased permeability and breakdown of vascular barriers and was evaluated by FA. 

At baseline, the leakage area during late FA correlated well with CNV size, ranging from 5.05 to 8.25 mm². At day 1 after treatment, the higher light dose protocol (50 J/cm²) showed a more intensive increase in leakage area of 2.9 mm². The reduced-light protocol induced only a moderate increase in leakage area of 1.16 mm². This difference between both treatment protocols was not statistically significant (P = 0.2570). OCT, performed in selected cases, showed characteristically in the 50-J/cm² group an increase in mostly subretinal fluid at day 1, which resolved by week 1 (Fig. 4) . 

CNV size at baseline was well balanced between the 25- and 50-J/cm² group (2.72 mm² vs. 3.12 mm²). Complete closure of the CNV, a parameter for early-treatment efficacy, was achieved in all eyes of both treatment groups. Figures 1 and 2demonstrate complete CNV closure in FA in eyes treated with 50 and 25 J/cm². 

Further follow-up showed recurrent CNV in all patients at month 3, indicating requirement for retreatment. There was an increase in CNV lesion size in the 25- and 50-J/cm² groups to 6.37 mm² and 4.63 mm², respectively. 

Visual Acuity

Visual acuity monitoring served as a safety precaution for the different protocols. Best-corrected visual acuity using the ETDRS protocol was balanced between the groups at baseline. None of the patients experienced an early severe vision decrease or lost more than 2 lines in 3 months of follow-up. 

Discussion

This study was designed to provide proof of principle that CNV closure may be achieved while avoiding damage to the surrounding physiologic choroid. The data show that an occlusion of the surrounding choriocapillaris is not necessary to achieve closure of the CNV lesion, but that these effects can be separated by choosing an appropriate dosimetry. Achievement of complete CNV closure together with an intact choroidal perfusion is a qualitative observation, a quantitative analysis of different dosages was not intended and further patient enrollment in the 50-J/cm² group was stopped after evidence of a significant photodynamic effect on the choroidal circulation. 

So far, the only parameters used for the treatment of CNV in AMD are those recommended by the TAP guidelines. The fact that the TAP regimen regularly leads to damage to the physiologic choroid has been demonstrated by ICGA showing early and often persistent nonperfusion of the surrounding choroid, ^{7 8} by histology showing a dose-dependent thrombosis of the choriocapillaris, ⁶ and by immunostaining demonstrating a reactive upregulation of VEGF. ⁹ However, the present study clearly demonstrates that such consequences are the result of the parameter selection and that such overtreatment effects may be avoided. Selecting optimal parameters allows differentiation between intended effects on the pathologic neovasculature and unwanted effects on the physiologic choroid. Patients treated with a light dose of 25 J/cm² demonstrated, similar to patients treated with a light dose of 50 J/cm², a complete closure of CNV at the 1-day follow-up. However, the choriocapillaris was substantially less affected by treatment with the lower light dose, as seen in early and late ICGA at day 1 and during further follow-up (Figs. 1 2) . No overall difference was seen in the 300- and 600-mW subgroups in the 25-J/cm² regimen. Patients in the 50-J/cm² group, receiving photosensitization 5 minutes after verteporfin infusion tended to have earlier, more, and longer-lasting choroidal changes. A detailed study was performed earlier quantifying the early effects of standard verteporfin therapy by using the same FA and ICGA measurements. ⁷ Comparing results of the 50-J/cm² protocol with the previous standard regimen study demonstrates that the qualitative sequence of angiographic events is identical with regard to standard infusion therapy or bolus administration. ⁷  

However, choroidal damage appeared to be inseparable from CNV effects with the standard therapy. Modification of treatment parameters with bolus administration and reduced fluence allowed selective closure of the CNV and less effect on the physiologic choroidal vasculature. Changing treatment parameters appeared not to have a relevant effect on short-term safety. None of the patients in any group lost more than 2 lines of vision over a 3-month follow-up and no closure of retinal vessels was observed. 

A selective approach may be particularly useful if earlier retreatment is considered or when other lesion types that are primarily less sensitive to PDT are treated as shown in the VIM study. ¹² Combining verteporfin therapy with antiangiogenic and anti-inflammatory therapies is currently in debate. ^{16 17} The postulated benefits are a lower CNV recurrence rate after verteporfin PDT, improved durability, inhibition of the verteporfin PDT–induced angiogenic response, and eventually better functional outcomes. In experimental studies, PDT was shown to induce a rapid inflammatory response including infiltration of leukocytes, increased expression of cytokines (e.g., intracellular adhesion molecules [ICAM]-1 and interleukin [IL]-6). ¹⁸ This inflammatory response correlates with increased retinal edema detected especially by OCT. ¹⁹ Selective verteporfin therapy primarily induces less angiogenic and inflammatory side effects on the level of the choroid and should respond even better to combination therapy. The additional benefit of adjunct therapy provides further evidence to support the hypothesis that outcomes are improved when the collateral side effects are reduced or, ideally, primarily avoided. 

Anti-VEGF therapy by intravitreal injection or systemic therapy has shown promising initial results. ^{20 21 22 23 24} Adding selective verteporfin therapy may improve angiographic and functional outcomes. However, VEGF is not only a potent angiogenic and permeability inducing factor, ^{25 26} but is also essential in maintaining normal vascular structures. ²⁷ Inhibiting a physiologic angiogenic response secondary to choriocapillary hypoxia after standard PDT will most likely prevent choriocapillary recanalization and lead to more extensive and persistent choriocapillary closure. 

Bolus infusion and reduced light dose in verteporfin therapy have demonstrated an improved selectivity with complete angiographic closure of the CNV and absence of a significant effect on the choroid. Modified parameters in verteporfin therapy should be of particular interest for evaluating the potential of combination therapy of PDT and antiangiogenic drugs. 

 Supported by a Grant (schm836/2) from the German Research Foundation (DFG).

 Submitted for publication March 19, 2005; revised August 25 and September 14, 2005; accepted November 28, 2005.

 Disclosure: S. Michels, None; F. Hansmann, None; W. Geitzenauer, None; U . Schmidt-Erfurth (P)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Treatment Parameters and Distribution of Patients

	50 J/cm2			25 J/cm2
Timing (min)	5	15	15	15
Patients (n)	3	3	6	7
Fluence (mW)	600	600	600	300
Duration (sec)	83	83	42	83
Drug dose (mg/m2)	6	6	6	6

Table 2.

 

Grading of Verteporfin Therapy’s Effect on the Choriocapillaris as Documented by ICGA

 

Table 3.

 

Patients with Choriocapillary Perfusion Changes of at Least Grade II or III in ICGA According to Table 2

	Day 1	Week 1	Week 4	Month 3
50 J ≥grade II	100	100	66.6	60
50 J ≥grade III	83.3	66.6	25	20
25 J ≥grade II	61.5	45.5	16.6	0
25 J ≥grade III	15.4	0	0	0

 Data are expressed as percentages of the total sample.

Figure 1.

 

FA before verteporfin therapy (PDT) shows a subfoveal, predominantly classic CNV delineated during early FA (A) with significant leakage in late FA (B). One day after treatment with a light dose of 50 J/cm², early FA (C) showed a hypofluorescent spot consistent with significant choriocapillary nonperfusion and additional nonperfusion of some larger choroidal vessels in early ICGA (D). One week after PDT early (E) and late (F) ICGA documented a persistent choroidal perfusion deficit that persisted throughout weeks 4 and 12.

Figure 2.

 

Early FA before verteporfin therapy (PDT) identified predominantly classic CNV with subretinal hemorrhage along the superior edge (A) and leakage in the late phase (B). One day after PDT treatment with a light dose of 25 J/cm², FA demonstrated complete closure of the CNV (C). There was some subretinal blood, but only a moderate effect on the choriocapillaris in ICGA (D). Early FA at 1 week exhibited persistent closure of the CNV (E) and late ICGA demonstrates only mild perfusion changes (F). No perfusion changes were seen after weeks 4 and 12.

Figure 3.

 

Hypofluorescent area in late ICGA.

Figure 4.

 

Optical coherence tomography (OCT) showed exudative changes in three vertical scans after verteporfin therapy (PDT) with a light dose of 50 J/cm². At baseline (A) subretinal fluid and a thickened retinal pigment epithelium (RPE)–photoreceptor band was seen. One day after therapy (B) subretinal fluid increased, and resolved completely by week 1 (C).