Radiation Therapy in the Treatment of Exudative AMD

Epimacular brachytherapy is still alive, and X-ray irradiation may provide yet another radiation modality for AMD.

Radiation Therapy in the Treatment of Exudative AMD

Epimacular brachytherapy is still alive, and X-ray irradiation may provide yet another radiation modality for AMD.

Darius M. Moshfeghi, MD

It seems that every five years or so, radiation therapy is revisited as a possible therapy for ocular disease, particularly for the treatment of choroidal neovascularization. We are currently in the middle of another one of these periods, and it is appropriate to review the rationale and technology and to ask when we may expect to get some answers.

The rationale for radiation therapy in CNV is manifold; it is: (1) antiangiogenic,1 (2) anti-inflammatory,2 and (3) antifibrotic.3 As most retina specialists are aware, the current interest in radiation as adjunct for treatment of exudative AMD has been largely driven by two competing technologies: (1) epimacular brachytherapy (EMBT) and (2) externally delivered low-voltage X-ray irradiation.


The epimacular brachytherapy approach involves pars plana vitrectomy, positioning of a shielded wand over the CNV in the macula, exposure to strontium (Sr) 90 for a period of approximately three minutes to deliver 24 Gy, and the use of intravitreal bevacizumab.4-6 The commercial name for the epimacular brachytherapy device is Vidion (Neovista, Inc., Newark, CA). The most obvious theoretical advantage of the epimacular brachytherapy approach is targeting certainty — the device is designed to be positioned directly over the CNV under direct visualization at the time of surgery.4-6

Darius Moshfeghi, MD, is director of pediatric vitreoretinal surgery and the director of ocular oncology in the Department of Ophthalmology, Stanford University School of Medicine in Stanford, CA.Dr. Moshfeghi reports minimal financial interest in Clarity Medical Systems, Compare Networks, Medavec, Genentech and InSitu Therapeutic, and moderate financial interest in Oraya. He can be reached via e-mail at

The externally delivered low-voltage X-ray treatment system delivers radiation through the inferior pars plana in three separate beams, which overlap on the macula.7 The patient is situated at the device with his or her chin in a modified slit-lamp head rest with a lens affixed to the treatment eye through suction, to allow for eye tracking and robotically controlled delivery of the beams.7 Prior to therapy, the patient receives an intravitreal injection of ranibizumab.

The commercial name for the low-voltage X-ray system is IRay (Oraya Therapeutics, Inc., Newark, CA). The obvious advantage of the low-voltage X-ray system is that it is nonsurgical (Figure 1).

Figure 1. The office-based IRay system designed by Oraya Therapeutics, Inc.

In the most simplistic terms, the two devices have been described as surgical (epimacular brachytherapy) and nonsurgical (low-voltage X-ray system). However, as we shall see, the differences in how the radiation is delivered may ultimately affect the success or failure of the technology. Finally, I would be remiss not to mention an external brachytherapy approach, championed by Salutaris Medical Devices, Inc. (Tucson, AZ), which remains in the early clinical trial design phase, and proton beam radiotherapy in conjunction with ranibizumab, both of which have enrolled six patients to date.8

For the purpose of this article, I will focus the discussion upon EMBT and low-voltage X-ray irradiation.


Epimacular brachytherapy uses a shielded beta radiation source that is unsheathed when it is positioned over the CNV in the eye. The Sr 90 delivers a radiation dose that drops off rapidly, moving away from the center of the source: a 10% decrease in dosage for every 100 µm.6 For comparison purposes, 100 µm is approximately 3 sheets of printing paper stacked together.

There are perils and advantages to this tight dosing profile. On the positive side, when properly positioned over the CNV, the radiation dose is directed toward the CNV with minimal dosage toward the normal retina and optic nerve. On the negative side, it is near impossible to maintain a z-distance above the macula within 100 µm tolerance for three minutes using a human-based delivery system.

This fact is gleaned from brachytherapy treatment for prostate cancer, in which human-guided needle placement on the x- and y- axes exceeds the 2 mm proscribed tolerance for repeatability and reproducibility greater than 10% of the time using 18-gauge needles, similar in size to the 19-gauge EMBT device.9

The gauge repeatability and reproducibility (GRR) index is felt to be acceptable if these numbers are less than 10%, marginally acceptable between 10% and 30%, and unacceptable if greater than 30%.9 McGill et al. noted that by physically constraining the side-to-side motion with a small aperture cuff, as well as slowing the insertion speed, they were able to optimize the GRR index.9 Obviously, 2 mm is 20-fold greater than the 100 µm tolerance required to deliver 90% to 110% of the desired dose of the EMBT device.6

Two side-by-side phase 1 trials were performed before the phase 3 CABERNET and MERLOT trials.5,6 Unlike in the in vitro prostate brachytherapy trial, there was no constraint on the movement on the x-, y-, and z-axes. These trials evaluated 24-Gy EMBT in treatment-naïve patients using radiation alone6 or in conjunction with two baseline injections of bevacizumab.5 These trials demonstrated no vision-limiting retinopathy or neuropathy after 12 months, with visual acuity improvement.5,6

Longer follow-up of 36 months for the EMBT plus bevacizumab group, on 19 of the original 34 patients, demonstrated several interesting trends.4 Mean visual acuity peaked at +15 ETDRS letters at 12 months, reached a nadir of -4.9 letters at 24 months (including all 34 patients), and rebounded to a mean of +3.9 letters at 36 months (19 patients, 17 losing <15 ETDRS letters).4

The drop-off noted at 24 months was attributable to cataract, which occurred in 50% of the phakic patients (12/24 patients; 10 patients pseudophakic at enrollment).4 One patient developed nonproliferative retinopathy at 36 months not affecting vision that had not progressed by 43 months.4

These promising results augured well for the treatment-naïve (CABERNET) and previously treated (MERLOT) phase 3 trials. However, results from the phase 3 CABERNET trial did not make the prespecified treatment outcomes. Specifically, CABERNET enrolled 457 treatment-naïve wet AMD patients in a 2:1 randomization (302 EMBT:155 modified PIER) to receive either 24 Gy of EMBT with two injections of ranibizumab followed by PRN ranibizumab vs a modified PIER protocol10 ranibizumab dosing regimen.11

This was a prospective trial with a noninferiority outcome aimed at a percentage of patients losing fewer than 15 ETDRS letters.11 At the two-year follow-up, patients randomized to EMBT received six ranibizumab injections and lost 2.5 letters vs 11 injections and a gain of 4.4 letters in the modified PIER protocol control.11 Vision favored the control group by >1 line on the ETDRS chart, with increased durability in the EMBT group.11 As a comparison, in the original PIER trial at 24 months patients lost 2.3 letters and received 10 ranibizumab injections.11,12

Other findings included 10 patients in the EMBT group who developed nonproliferative retinopathy that did not develop neovascular changes or result in an average decrease in vision at two years.12 One of the study leaders, Pravin U. Dugel, MD, speculated that there were two problems with the trial design: (1) inclusion of treatment-naïve patients; and (2) inaccurate probe placement.11

With respect to patient inclusion, Dr. Dugel felt that be cause the majority of patients who would likely be treated with the therapy would be previously treated, this patient population would likely have been the ideal population to have studied.11 To buttress this argument, a nonrandomized, prospective trial of previously treated wet AMD patients (MERITAGE) demonstrated that EMBT resulted in an improvement in vision in 63% of patients at six months, with a concomitant decreased need for anti-VEGF therapy.11,13 The MERITAGE trial has served as the basis for a government-sponsored phase 3 trial of EMBT in previously treated patients in the United Kingdom.11

With respect to probe placement, Dr. Dugel noted that the CABERNET trial did not monitor probe placement, yet he hypothesized, “If the surgeon was dedicated to placing the tip of the probe on the retina and delivering the radiation the way it was designed, there was a much better chance patients did well.”11 Similarly, with respect to the nonproliferative retinopathy, that adverse event was indicative of improper probe placement.11

However, one might reasonably argue the opposite, ie, that by decreasing the distance to the CNV by 100 µm would have resulted in overtreatment of 10%, increasing the effective dose from 24 to 26.4 Gy and increasing the likelihood of retinopathy.6 In either event, lack of knowledge of probe placement accuracy makes it difficult to evaluate the true efficacy of EMBT.


The externally delivered low-voltage X-ray irradiation system offers theoretical advantages to any surgically based system. First, the system is nonsurgical.7 Second, the system delivers background radiation or less to the treating physician, as opposed to direct handling, as occurs with any brachytherapy system.7 Third, the radiation delivery system is robotically controlled, eliminating the human element in probe placement abnormalities.7

This robotically controlled delivery system works in tandem with an eye-tracking system that monitors eye movements on x-, y- and z-axes and on rotational planes using an ocular fixation system based upon a suction-adhered lens.7 Any deviations beyond preset thresholds automatically and independently initiate a gating event, interrupting the radiation delivery until such time that the baseline conditions are restored.7

A key difference between the low-voltage X-ray irradiation system and EMBT is that in EMBT, the probe is placed directly over the CNV, with the inherent difficulties in maintaining placement, as noted above. While the dosing accuracy is unknown with EMBT, direct visualization of the CNV leads to targeting accuracy. In the low-voltage Xray irradiation system, targeting is based upon knowledge of the axial length of the eye (measured prior to treatment), as well as a shift from the visual axis to deliver radiation to the center of the macula (Figure 2).7

Figure 2. Illustration of the trajectory of the external beam radiation through the pars plana region into the macula, avoiding the lens and optic nerve.

This delivery system was evaluated using two different Monte Carlo simulations14,15 and an analysis of computed tomography data, as well as male and female head phantoms16 and a cadaver eye study,17 all of which indicated do sing accuracy such that a 90% isodose was delivered to the center of the macula, with subthreshold doses delivered to surrounding tissues. While targeting is not visualized directly, dosing accuracy is accurate and reproducible.14-17

Phase 1 clinical trials in Mexico evaluated three dosing regimens over a period of one year: (1) 16 Gy plus two doses of ranibizumab, followed by PRN ranibizumab; (2) 24 Gy plus two doses of ranibizumab, followed by PRN ranibizumab; and (3)16 Gy followed by PRN ranibizumab. To date, six-month follow-up has demonstrated that the 16- and 24-Gy doses, in conjunction with baseline loading of ranibizumab, resulted in improvement in visual acuity with a decreased need for injections.18,19

The 16 Gy radiation first, followed by as needed ranibizumab, demonstrated stability of vision with a decreased need for injections at six months, hinting at the independent effect of radiation upon CNV.20 There was no incidence of cataract progression, radiation retinopathy, optic neuropathy, or any other vision- or life-limiting event noted in any group.18-20

A randomized, prospective, double-blind, multicenter, controlled clinical trial (INTREPID) is being performed in Europe at 251 sites for previously treated patients. Eligibility for INTREPID included previously treated patients with CNV due to AMD with a diagnosis within three years and at least three ranibizumab or bevacizumab injections in the previous 12 months.

The INTREPID trial has enrolled 226 patients in a 2:1:2:1 randomization to receive either 16 Gy or 24 Gy of radiation (or matching sham radiation) at baseline in conjunction with a baseline injection of ranibizumab. The control groups receive sham radiation plus a baseline ranibizumab injection.

In both the experimental and control arms, retreatment with ranibizumab is guided by OCT findings (an increase of 100 µm in the central foveal subfield from the best previous exam), new or increased macular hemorrhage, or a >5 ETDRS letter decrease from baseline vision in conjunction with AMD activity.

The primary outcome of INTREPID was the number of injections given over a 52-week period, with secondary outcomes including changes in mean visual acuity, loss of <15 ETDRS letters, gain of ≥15 ETDRS letters, gain of ≥0 ETDRS letters, and change in CNV size.

One criticism of INTREPID is that the retreatment criteria are overly broad and do not represent the clinical practice here in the United States. However, the trial is European-based, and treatment indications are less stringent. Patient management and retreatment criteria have evolved since the trial was initiated, with no overarching standard. Additionally, because the primary outcomes are partially based upon on OCT findings, the 100-µm increase in central foveal subfield criteria (as opposed to treating for any fluid or cyst) works to increase the threshold for retreatment in both groups, potentially limiting the ability to detect a difference between the two groups. Results from INTREPID should be available in the late summer of 2012.


The year 2012 should be notable for providing further clarity to the role of radiation in the treatment of CNV from AMD. While the results from the CABERNET trial are underwhelming thus far, the data from the EMBT trial of previously treated patients (MERLOT) are pending. Similarly, data from the previously treated wet AMD population with the low-voltage X-ray irradiation system (INTREPID) are due soon and should give us greater insight into radiation as a therapy for wet AMD. RP


1. Hatjikondi O, Ravazoula P, Kardamakis D, Dimopoulos J, Papaioannou S. In vivo experimental evidence that the nitric oxide pathway is involved in the X-ray-induced antiangiogenicity. Br J Cancer. 1996;74:1916-1923.
2. Rodel F, Keilholz L, Herrmann M, Sauer R, Hildebrandt G. Radiobiological mechanisms in inflammatory diseases of low-dose radiation therapy. Int J Radiat Biol 2007;83:357-366.
3. Hart PM, Chakravarthy U, MacKenzie G, Archer DB, Houston RF. Teletherapy for subfoveal choroidal neovascularisation of age-related macular degeneration: results of follow up in a non-randomised study. Br J Ophthalmol. 1996;80:1046-1050.
4. Avila MP, Farah ME, Santos A, et al. Three-year safety and visual acuity results of epimacular 90 strontium/90 yttrium brachytherapy with bevacizumab for the treatment of subfoveal choroidal neovascularization secondary to age-related macular degeneration. Retina. 2012;32:10-18.
5. Avila MP, Farah ME, Santos A, Duprat JP, Woodward BW, Nau J. Twelve-month short-term safety and visual-acuity results from a multicentre prospective study of epiretinal strontium-90 brachytherapy with bevacizumab for the treatment of subfoveal choroidal neovascularisation secondary to age-related macular degeneration. Br J Ophthalmol. 2009;93:305-309.
6. Avila MP, Farah ME, Santos A, et al. Twelve-month safety and visual acuity results from a feasibility study of intraocular, epiretinal radiation therapy for the treatment of subfoveal CNV secondary to AMD. Retina. 2009;29:157-169.
7. Moshfeghi DM, Kaiser PK, Gertner M. Stereotactic low-voltage x-ray irradiation for age-related macular degeneration. Br J Ophthalmol. 2011;95:185-188.
8. Park SS, Daftari I, Phillips T, Morse LS. Three-year follow-up of a pilot study of ranibizumab combined with proton beam irradiation as treatment for exudative age-related macular degeneration. Retina. 2012 May;32:956-966.
9. McGill CS, Schwartz JA, Moore JZ, McLaughlin PW, Shih AJ. Precision grid and hand motion for accurate needle insertion in brachytherapy. Med Phys. 2011;38:4749-4759.
10. Regillo CD, Brown DM, Abraham P, et al. Randomized, double-masked, sham-controlled trial of ranibizumab for neovascular age-related macular degeneration: PIER study year 1. Am J Ophthalmol. 2008;145:239-248.e5.
11. Today R. Radiation treatments for neovascular AMD. Retin Today. 2012;7: 56-58.
12. Abraham P, Yue H, Wilson L. Randomized, double-masked, sham-controlled trial of ranibizumab for neovascular age-related macular degeneration: PIER study year 2. Am J Ophthalmol. 2010;150:315-324.e1.
13. Dugel P. Internal radiation for AMD: The CABERNET study. Paper presented at: Annual meeting of the American Academy of Ophthalmology; Orlando, FL; October 22-25, 2011.
14. Hanlon J, Firpo M, Chell E, Moshfeghi DM, Bolch WE. Stereotactic radiosur - gery for AMD: a Monte Carlo-based assessment of patient-specific tissue doses. Invest Ophthalmol Vis Sci. 2011;52:2334-2342.
15. Taddei PJ, Chell E, Hansen S, Gertner M, Newhauser WD. Assessment of targeting accuracy of a low-energy stereotactic radiosurgery treatment for age-related macular degeneration. Phys Med Biol. 2010;55:7037-7054.
16. Hanlon J, Lee C, Chell E, et al. Kilovoltage stereotactic radiosurgery for age-related macular degeneration: assessment of optic nerve dose and patient effective dose. Med Phys. 2009;36:3671-3681.
17. Gertner M, Chell E, Pan KH, Hansen S, Kaiser PK, Moshfeghi DM. Stereotactic targeting and dose verification for age-related macular degeneration. Med Phys 2010;37:600-606.
18. VM, Quiroz-Mercado H, Velez-Montoya R, et al. 16-Gy low-voltage X-ray irradiation with ranibizumab therapy for AMD: 6-month safety and functional outcomes. Ophthalmic Surg Lasers Imaging. 2011;42:1-6.
19. Canton VM, Quiroz-Mercado H, Velez-Montoya R, et al. 24-Gy low-voltage x-ray irradiation with ranibizumab therapy for neovascular AMD: 6-month safety and functional outcomes. Ophthalmic Surg Lasers Imaging. 2012;43:20-24.
20. Moshfeghi AA, Canton VM, Quiroz-Mercado H, et al. 16-Gy low-voltage X-ray irradiation followed by as-needed ranibizumab therapy for AMD: 6-month outcomes of a "radiation-first" strategy. Ophthalmic Surg Lasers Imaging. 2011; 42:1-8.