Clinical science
Proton beam irradiation for non-AMD CNV: 2-year results of a randomised clinical trial Ling Chen,1,2,3 Ivana K Kim,1 Anne M Lane,1 Danny Gauthier,4 John E Munzenrider,5 Evangelos S Gragoudas,1 Joan W Miller1 1
Retina Service, Massachusetts Eye and Ear Infirmary, Department of Ophthalmology, Harvard Medical School, Boston, Massachusetts, USA 2 Department of Ophthalmology, The University of Hong Kong, Hong Kong, China 3 Department of Ophthalmology & Vision Science, Eye & ENT Hospital, Shanghai Medical School, Fudan University, Shanghai, China 4 Retina Service, Ophthalmology Department, University of Montreal, Hôpital Notre-Dame, Montreal, Quebec, Canada 5 Department of Radiation Oncology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA Correspondence to Dr Joan W Miller, Chair of Ophthalmology, Massachusetts Eye and Ear Infirmary, 243 Charles Street, Boston, MA 02114, USA; [email protected] Received 8 December 2013 Revised 28 February 2014 Accepted 20 March 2014 Published Online First 12 May 2014
ABSTRACT Aims To evaluate safety and visual outcomes after proton beam irradiation (PBI) therapy for subfoveal choroidal neovascularisation (CNV) secondary to causes other than age-related macular degeneration (AMD). Methods This study is a prospective, unmasked and randomised clinical trial using two dosage regimens, conducted in the Massachusetts Eye and Ear Infirmary. The study included 46 patients with CNV secondary to non-AMD and best-corrected visual acuity of 20/320 or better. Patients were randomly assigned to receive 16 or 24 cobalt gray equivalents (CGE) of PBI in two equal fractions. Complete ophthalmological examinations, fundus photography and fluorescein angiography were performed at baseline and 6, 12, 18 and 24 months after treatment. Results At 1 year after treatment, 82% and 72% lost fewer than 1.5 lines of vision in the 16 CGE and in 24 CGE groups, respectively. At 2 years after therapy, 77% in the lower dose group and 64% in the higher dose group lost fewer than 1.5 lines of vision. Mild radiation complications such as radiation vasculopathy developed in 17.6% of patients. Conclusions PBI is a safe and efficacious treatment for subfoveal CNV not due to AMD. The data with respect to visual outcomes and radiation complications trend in favour of the 16 CGE group, although differences do not reach statistical significance. PBI may be considered as an alternative to current therapies.
INTRODUCTION
To cite: Chen L, Kim IK, Lane AM, et al. Br J Ophthalmol 2014;98: 1212–1217. 1212
Choroidal neovascularisation (CNV) is a severe sight-threatening complication that develops most commonly in the setting of age-related macular degeneration (AMD). However, it may result from other conditions associated with abnormalities in Bruch’s membrane, including pathological myopia, presumed ocular histoplasmosis and angioid streaks. Natural history is poor in these patients,1–5 and response to treatment has been shown to be variable.2 5–10 Laser photocoagulation may be effective for extrafoveal lesions, but is not desirable for subfoveal membranes. Photodynamic therapy (PDT) using verteporfin has been proven effective for subfoveal CNV due to myopia. Patients with myopic subfoveal CNV treated with PDT experienced less visual loss than a sham-treated group at 1-year follow-up2; however, this beneficial effect was no longer significant at 2-year follow-up.11 Therapies targeting vascular endothelial growth factor (VEGF) have emerged as the standard of care for CNV secondary to AMD,12 13 but the long-term efficacy of anti-VEGF therapy has not been as well evaluated in clinical trials for patients
with subfoveal CNV due to aetiologies other than AMD. In addition, both PDT and anti-VEGF therapies require frequent re-treatments, which may impose additional risk and inconvenience for patients. Radiation represents a treatment modality with the benefit of a single treatment and the potential for more selectivity due to the relative resistance of neural tissue to radiation damage. It has been investigated mainly for the treatment of subfoveal CNV secondary to AMD14–17 based on the rationale that radiation can inhibit endothelial cell proliferation, angiogenic cytokine-producing inflammatory cells in CNV complexes and reduce fibroblast proliferation involved in scar formation.18 19 Radiation exposure to normal tissues is the main cause of complications, but unlike other sources of external beam radiation, proton beam irradiation (PBI) is known to deliver more than 90% of the radiation dose to the target tissue, and therefore collateral tissue damage is minimised.20 A number of studies using PBI for neovascular AMD have demonstrated possible benefit.14–16 We report here the results of a randomised clinical trial to evaluate safety and visual outcomes after 16 cobalt gray equivalents (CGE) versus 24 CGE of PBI therapy for subfoveal CNV secondary to aetiologies other than AMD.
MATERIALS AND METHODS Study design This was a prospective, unmasked, randomised clinical trial of two radiation doses for patients with subfoveal CNV, with a classic component, secondary to causes other than AMD. Patients were assigned to receive a total dose of 16 or 24 CGE PBI fractionated in two equal doses over a 2-day or 3-day period. This study was conducted in the Retina Service of Massachusetts Eye and Ear Infirmary (MEEI). Recruitment efforts were initiated in May 1996, and the last patient was enrolled in October 1999. There were no other effective, proven treatments available at that time. The Institutional Review Board approval was obtained before initiation of the study. Informed consent was obtained from the subjects.
Patient selection, entry evaluations and follow-up assessment All patients with subfoveal CNV secondary to causes other than AMD demonstrated by fluorescein angiography with best-corrected visual acuity of 20/320 or better using the Early Treatment Diabetic Retinopathy Study (ETDRS) charts were eligible for the study. Sixteen patients (34.8%) had received prior laser treatment (two treated OU)
Chen L, et al. Br J Ophthalmol 2014;98:1212–1217. doi:10.1136/bjophthalmol-2013-304761
Clinical science and two (4.3%) patients had undergone PDT prior to proton irradiation. Patients with any other macular diseases that might have contributed to visual loss were excluded from the study. Best-corrected visual acuity testing, colour fundus photography, fluorescein angiography and a complete ophthalmological examination at baseline were performed to determine eligibility. Eligibility was confirmed by completion of an eligibility review form. This review form was submitted to a study coordinator who randomly assigned (1:1) each patient to receive a dose of 16 or 24 CGE. Patients were asked to return to the MEEI Retina Service for safety and clinical assessments 6, 12, 18 and 24 months after treatment. Radiation retinopathy and radiation papillopathy were assessed at follow-up visits using colour fundus photographs and fluorescein angiograms in addition to slit lamp biomicroscopy.
Proton beam irradiation therapy The light-field technique for PBI was used as previously published.14 21 Treatment parameters were generated by the ophthalmologist (ESG or JWM) from the baseline fluorescein angiogram. The area of the CNV lesion was treated with margin of 2 mm to 90% of the dose. All treatments were performed at the Harvard Cyclotron Laboratory.
Vision testing Best-corrected visual acuity was recorded as a visual acuity score based on the total number of letters read at 2 m using a modified ETDRS protocol. If the patient read fewer than 20 letters at 2 m, visual acuity was measured at 1 m.
Radiation vasculopathy assessment Radiation retinopathy and radiation papillopathy were assessed retrospectively at each time point using colour fundus photographs and fluorescein angiograms. Features of radiation retinopathy that were identified included macular oedema, haemorrhages, microvascular changes, nerve fibre layer infarcts, neovascularisation, retinal pigment epithelial atrophy, subretinal lipid, capillary non-perfusion, arteriolar occlusion, cotton wool spots and vascular sheathing. The following radiation papillopathy features were indentified: optic disc oedema, haemorrhage, atrophy, pallor and exudate.
Statistical analysis The primary outcome measurement was the proportion of eyes losing or gaining 1.5 lines (eight letters) of visual acuity from baseline. Cumulative rates of vision loss were calculated using Kaplan–Meier estimates, and meaningful differences in rates between dose groups were tested using the log-rank test. Restricting analysis to patients with CNV due to pathological myopia (N=27), differences in visual outcomes between dose groups at annual time points were evaluated using Fisher’s exact test.
RESULTS In total, 51 eyes of 46 patients with CNV due to aetiologies other than AMD were included in this study. Five patients received treatment in both eyes by the completion of the study. Twenty-nine eyes received 16 CGE, and 22 eyes received 24 CGE. Patients were between 16 and 79 years of age, with mean ages of 48.1 years (16 CGE) and 45.7 years (24 CGE) at the time of study entry. Myopia constituted the most common diagnosis (52.9%), followed by angioid streaks (17.6%) and presumed ocular histoplasmosis (9.8%). Both groups were comparable in terms of visual acuity and lesion size at
Table 1 Baseline characteristics of 46 patients with subfoveal CNV secondary to causes other than AMD Characteristic Number of eyes/patients Median age, years (range) Caucasian Male, N (%) Myopia Visual acuity, N (%) 20/20–20/100 Classic CNV, no. (%) CNV size, no. (%) ≤2 MPS DAs Prior treatment
16 cobalt grey equivalent
24 cobalt grey equivalent
p Value
29/27 48 (16–77)
22/19 47 (28–79)
0.96
24 (88.9) 13 (48.1) 16 (59.3)
19 (100) 12 (63.2) 6 (31.6)
0.26 0.38 0.08
19 (67.9) 19 (65.5)
15 (68.2) 17 (77.3)
1.0 0.54
14 (48.3) 13/12 (44.8)
15 (68.2) 5/4 (22.7)
0.25 0.14
AMD, age-related macular degeneration; CNV, choroidal neovascularisation; MPS DAs, macular photocoagulation study disc areas.
presentation, and treatment procedures were equally distributed between right and left eyes. Eighty per cent of patients completed 12-month follow-up, while 61% completed the 24-month follow-up. Characteristics of the 46 patients (51 eyes) at baseline are summarised in table 1.
Visual acuity outcomes Mean change in visual acuity from the baseline screen to the 12-month follow-up visit was +4.5 letters and +1.7 letters for the 16 Gy (n=22) and 24 Gy (n=18) dose groups, respectively. A visual acuity score was available for 31 patients at the end of the study (24-month follow-up visit). Of these, mean change in visual acuity had diminished to +0.24 letters in patients treated with 16 Gy and −7.6 letters in the 24 Gy group.
Vision loss At 12 months after treatment, the proportion of eyes with vision maintenance or gain (loss of 2 disc areas (DA) versus those with lesions ≤2 DA size (log-rank test, p=0.39). Similar rates of vision loss were also seen in patients with baseline visual acuity of 20/100 or worse compared with those with baseline visual acuity better than 20/100 (log-rank test, p=0.49).
Vision gain The proportion of patients with a visual acuity gain (eight or more letters or 1.5 lines) was 36% in the 16 CGE group and 28% in the 24 CGE at 12 months after treatment. This trend was also seen at the 24-month visit, with 4 of 17 eyes (24%) in the 16 CGE group versus 1 of 14 eyes (7%) in the 24 CGE group gaining ≥1.5 lines of visual acuity. Meaningful visual acuity improvement (≥15 letters) was observed in 5 of 22 eyes (23%) in
Chen L, et al. Br J Ophthalmol 2014;98:1212–1217. doi:10.1136/bjophthalmol-2013-304761
1213
Clinical science Radiation complications were not statistically related to dose assignment. Complications occurred in 3 of 29 eyes (10%) in the 16 CGE group and in 6 of 22 eyes (27%) in the 24 CGE group ( p=0.28, Fisher’s exact test). Additionally, there was not a statistically significant association between signs of radiation vasculopathy and vision loss of 1.5 lines or more at 24-month follow-up visit. Three out of nine eyes (33%) with evidence of radiation vasculopathy experienced vision loss of 1.5 lines or more at the 24-month follow-up visit compared with 29% of eyes without such signs ( p=1.0, Fisher’s exact test).
CASE REPORTS Figure 1 Changes in visual acuity from baseline by radiation dose. the lower dose group and 5 of 18 eyes (28%) in the higher dose group at 12 months after therapy. This improvement was not sustained at 24 months after treatment, with 3 of 17 eyes (18%) and 1 of 14 eyes (7%) in the lower dose group and higher dose group, respectively, achieving a three-line gain (figure 1). A subgroup analysis was performed for patients with CNV due to myopia (table 2). Results in this subgroup were similar to those of the overall group. Greater proportions of patients lost 20/100 BLVA ≤20/100
12-month follow-up (N=23) N (%)
24-month follow-up (N=18) N (%)
Loss of
Proton beam irradiation for non-AMD CNV: 2-year results of a randomised clinical trial Ling Chen,1,2,3 Ivana K Kim,1 Anne M Lane,1 Danny Gauthier,4 John E Munzenrider,5 Evangelos S Gragoudas,1 Joan W Miller1 1
Retina Service, Massachusetts Eye and Ear Infirmary, Department of Ophthalmology, Harvard Medical School, Boston, Massachusetts, USA 2 Department of Ophthalmology, The University of Hong Kong, Hong Kong, China 3 Department of Ophthalmology & Vision Science, Eye & ENT Hospital, Shanghai Medical School, Fudan University, Shanghai, China 4 Retina Service, Ophthalmology Department, University of Montreal, Hôpital Notre-Dame, Montreal, Quebec, Canada 5 Department of Radiation Oncology, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts, USA Correspondence to Dr Joan W Miller, Chair of Ophthalmology, Massachusetts Eye and Ear Infirmary, 243 Charles Street, Boston, MA 02114, USA; [email protected] Received 8 December 2013 Revised 28 February 2014 Accepted 20 March 2014 Published Online First 12 May 2014
ABSTRACT Aims To evaluate safety and visual outcomes after proton beam irradiation (PBI) therapy for subfoveal choroidal neovascularisation (CNV) secondary to causes other than age-related macular degeneration (AMD). Methods This study is a prospective, unmasked and randomised clinical trial using two dosage regimens, conducted in the Massachusetts Eye and Ear Infirmary. The study included 46 patients with CNV secondary to non-AMD and best-corrected visual acuity of 20/320 or better. Patients were randomly assigned to receive 16 or 24 cobalt gray equivalents (CGE) of PBI in two equal fractions. Complete ophthalmological examinations, fundus photography and fluorescein angiography were performed at baseline and 6, 12, 18 and 24 months after treatment. Results At 1 year after treatment, 82% and 72% lost fewer than 1.5 lines of vision in the 16 CGE and in 24 CGE groups, respectively. At 2 years after therapy, 77% in the lower dose group and 64% in the higher dose group lost fewer than 1.5 lines of vision. Mild radiation complications such as radiation vasculopathy developed in 17.6% of patients. Conclusions PBI is a safe and efficacious treatment for subfoveal CNV not due to AMD. The data with respect to visual outcomes and radiation complications trend in favour of the 16 CGE group, although differences do not reach statistical significance. PBI may be considered as an alternative to current therapies.
INTRODUCTION
To cite: Chen L, Kim IK, Lane AM, et al. Br J Ophthalmol 2014;98: 1212–1217. 1212
Choroidal neovascularisation (CNV) is a severe sight-threatening complication that develops most commonly in the setting of age-related macular degeneration (AMD). However, it may result from other conditions associated with abnormalities in Bruch’s membrane, including pathological myopia, presumed ocular histoplasmosis and angioid streaks. Natural history is poor in these patients,1–5 and response to treatment has been shown to be variable.2 5–10 Laser photocoagulation may be effective for extrafoveal lesions, but is not desirable for subfoveal membranes. Photodynamic therapy (PDT) using verteporfin has been proven effective for subfoveal CNV due to myopia. Patients with myopic subfoveal CNV treated with PDT experienced less visual loss than a sham-treated group at 1-year follow-up2; however, this beneficial effect was no longer significant at 2-year follow-up.11 Therapies targeting vascular endothelial growth factor (VEGF) have emerged as the standard of care for CNV secondary to AMD,12 13 but the long-term efficacy of anti-VEGF therapy has not been as well evaluated in clinical trials for patients
with subfoveal CNV due to aetiologies other than AMD. In addition, both PDT and anti-VEGF therapies require frequent re-treatments, which may impose additional risk and inconvenience for patients. Radiation represents a treatment modality with the benefit of a single treatment and the potential for more selectivity due to the relative resistance of neural tissue to radiation damage. It has been investigated mainly for the treatment of subfoveal CNV secondary to AMD14–17 based on the rationale that radiation can inhibit endothelial cell proliferation, angiogenic cytokine-producing inflammatory cells in CNV complexes and reduce fibroblast proliferation involved in scar formation.18 19 Radiation exposure to normal tissues is the main cause of complications, but unlike other sources of external beam radiation, proton beam irradiation (PBI) is known to deliver more than 90% of the radiation dose to the target tissue, and therefore collateral tissue damage is minimised.20 A number of studies using PBI for neovascular AMD have demonstrated possible benefit.14–16 We report here the results of a randomised clinical trial to evaluate safety and visual outcomes after 16 cobalt gray equivalents (CGE) versus 24 CGE of PBI therapy for subfoveal CNV secondary to aetiologies other than AMD.
MATERIALS AND METHODS Study design This was a prospective, unmasked, randomised clinical trial of two radiation doses for patients with subfoveal CNV, with a classic component, secondary to causes other than AMD. Patients were assigned to receive a total dose of 16 or 24 CGE PBI fractionated in two equal doses over a 2-day or 3-day period. This study was conducted in the Retina Service of Massachusetts Eye and Ear Infirmary (MEEI). Recruitment efforts were initiated in May 1996, and the last patient was enrolled in October 1999. There were no other effective, proven treatments available at that time. The Institutional Review Board approval was obtained before initiation of the study. Informed consent was obtained from the subjects.
Patient selection, entry evaluations and follow-up assessment All patients with subfoveal CNV secondary to causes other than AMD demonstrated by fluorescein angiography with best-corrected visual acuity of 20/320 or better using the Early Treatment Diabetic Retinopathy Study (ETDRS) charts were eligible for the study. Sixteen patients (34.8%) had received prior laser treatment (two treated OU)
Chen L, et al. Br J Ophthalmol 2014;98:1212–1217. doi:10.1136/bjophthalmol-2013-304761
Clinical science and two (4.3%) patients had undergone PDT prior to proton irradiation. Patients with any other macular diseases that might have contributed to visual loss were excluded from the study. Best-corrected visual acuity testing, colour fundus photography, fluorescein angiography and a complete ophthalmological examination at baseline were performed to determine eligibility. Eligibility was confirmed by completion of an eligibility review form. This review form was submitted to a study coordinator who randomly assigned (1:1) each patient to receive a dose of 16 or 24 CGE. Patients were asked to return to the MEEI Retina Service for safety and clinical assessments 6, 12, 18 and 24 months after treatment. Radiation retinopathy and radiation papillopathy were assessed at follow-up visits using colour fundus photographs and fluorescein angiograms in addition to slit lamp biomicroscopy.
Proton beam irradiation therapy The light-field technique for PBI was used as previously published.14 21 Treatment parameters were generated by the ophthalmologist (ESG or JWM) from the baseline fluorescein angiogram. The area of the CNV lesion was treated with margin of 2 mm to 90% of the dose. All treatments were performed at the Harvard Cyclotron Laboratory.
Vision testing Best-corrected visual acuity was recorded as a visual acuity score based on the total number of letters read at 2 m using a modified ETDRS protocol. If the patient read fewer than 20 letters at 2 m, visual acuity was measured at 1 m.
Radiation vasculopathy assessment Radiation retinopathy and radiation papillopathy were assessed retrospectively at each time point using colour fundus photographs and fluorescein angiograms. Features of radiation retinopathy that were identified included macular oedema, haemorrhages, microvascular changes, nerve fibre layer infarcts, neovascularisation, retinal pigment epithelial atrophy, subretinal lipid, capillary non-perfusion, arteriolar occlusion, cotton wool spots and vascular sheathing. The following radiation papillopathy features were indentified: optic disc oedema, haemorrhage, atrophy, pallor and exudate.
Statistical analysis The primary outcome measurement was the proportion of eyes losing or gaining 1.5 lines (eight letters) of visual acuity from baseline. Cumulative rates of vision loss were calculated using Kaplan–Meier estimates, and meaningful differences in rates between dose groups were tested using the log-rank test. Restricting analysis to patients with CNV due to pathological myopia (N=27), differences in visual outcomes between dose groups at annual time points were evaluated using Fisher’s exact test.
RESULTS In total, 51 eyes of 46 patients with CNV due to aetiologies other than AMD were included in this study. Five patients received treatment in both eyes by the completion of the study. Twenty-nine eyes received 16 CGE, and 22 eyes received 24 CGE. Patients were between 16 and 79 years of age, with mean ages of 48.1 years (16 CGE) and 45.7 years (24 CGE) at the time of study entry. Myopia constituted the most common diagnosis (52.9%), followed by angioid streaks (17.6%) and presumed ocular histoplasmosis (9.8%). Both groups were comparable in terms of visual acuity and lesion size at
Table 1 Baseline characteristics of 46 patients with subfoveal CNV secondary to causes other than AMD Characteristic Number of eyes/patients Median age, years (range) Caucasian Male, N (%) Myopia Visual acuity, N (%) 20/20–20/100 Classic CNV, no. (%) CNV size, no. (%) ≤2 MPS DAs Prior treatment
16 cobalt grey equivalent
24 cobalt grey equivalent
p Value
29/27 48 (16–77)
22/19 47 (28–79)
0.96
24 (88.9) 13 (48.1) 16 (59.3)
19 (100) 12 (63.2) 6 (31.6)
0.26 0.38 0.08
19 (67.9) 19 (65.5)
15 (68.2) 17 (77.3)
1.0 0.54
14 (48.3) 13/12 (44.8)
15 (68.2) 5/4 (22.7)
0.25 0.14
AMD, age-related macular degeneration; CNV, choroidal neovascularisation; MPS DAs, macular photocoagulation study disc areas.
presentation, and treatment procedures were equally distributed between right and left eyes. Eighty per cent of patients completed 12-month follow-up, while 61% completed the 24-month follow-up. Characteristics of the 46 patients (51 eyes) at baseline are summarised in table 1.
Visual acuity outcomes Mean change in visual acuity from the baseline screen to the 12-month follow-up visit was +4.5 letters and +1.7 letters for the 16 Gy (n=22) and 24 Gy (n=18) dose groups, respectively. A visual acuity score was available for 31 patients at the end of the study (24-month follow-up visit). Of these, mean change in visual acuity had diminished to +0.24 letters in patients treated with 16 Gy and −7.6 letters in the 24 Gy group.
Vision loss At 12 months after treatment, the proportion of eyes with vision maintenance or gain (loss of 2 disc areas (DA) versus those with lesions ≤2 DA size (log-rank test, p=0.39). Similar rates of vision loss were also seen in patients with baseline visual acuity of 20/100 or worse compared with those with baseline visual acuity better than 20/100 (log-rank test, p=0.49).
Vision gain The proportion of patients with a visual acuity gain (eight or more letters or 1.5 lines) was 36% in the 16 CGE group and 28% in the 24 CGE at 12 months after treatment. This trend was also seen at the 24-month visit, with 4 of 17 eyes (24%) in the 16 CGE group versus 1 of 14 eyes (7%) in the 24 CGE group gaining ≥1.5 lines of visual acuity. Meaningful visual acuity improvement (≥15 letters) was observed in 5 of 22 eyes (23%) in
Chen L, et al. Br J Ophthalmol 2014;98:1212–1217. doi:10.1136/bjophthalmol-2013-304761
1213
Clinical science Radiation complications were not statistically related to dose assignment. Complications occurred in 3 of 29 eyes (10%) in the 16 CGE group and in 6 of 22 eyes (27%) in the 24 CGE group ( p=0.28, Fisher’s exact test). Additionally, there was not a statistically significant association between signs of radiation vasculopathy and vision loss of 1.5 lines or more at 24-month follow-up visit. Three out of nine eyes (33%) with evidence of radiation vasculopathy experienced vision loss of 1.5 lines or more at the 24-month follow-up visit compared with 29% of eyes without such signs ( p=1.0, Fisher’s exact test).
CASE REPORTS Figure 1 Changes in visual acuity from baseline by radiation dose. the lower dose group and 5 of 18 eyes (28%) in the higher dose group at 12 months after therapy. This improvement was not sustained at 24 months after treatment, with 3 of 17 eyes (18%) and 1 of 14 eyes (7%) in the lower dose group and higher dose group, respectively, achieving a three-line gain (figure 1). A subgroup analysis was performed for patients with CNV due to myopia (table 2). Results in this subgroup were similar to those of the overall group. Greater proportions of patients lost 20/100 BLVA ≤20/100
12-month follow-up (N=23) N (%)
24-month follow-up (N=18) N (%)
Loss of
