The treatment of thyroid eye disease (TED) remains difficult and controversial. The use of orbital radiation therapy (ORT) varies widely among clinicians, and its role may change in the near future with the development of new immunomodulating pharmacologic agents. Two experts now discuss the case for and against ORT in TED.
PRO—The Argument for Orbital Radiation Therapy for Thyroid Eye Disease Michael Kazim, MD
Despite progress in understanding the pathophysiology of TED, therapeutic advances have lagged. Although the therapeutic goals in the stable phase of disease are clearly defined (restoration of premorbid ocular function and appearance) and can be achieved through surgical maneuvers, the acute phase goals (mitigating the disease manifestations and abbreviating the acute phase) have been more elusive. Initially showing great promise, the adverse effects of corticosteroid therapy have limited its routine use to severe or vision-threatening complications of TED. When corticosteroids fail or are contraindicated, surgical options include orbital decompression for thyroid optic neuropathy (TON) and lid lengthening procedures for corneal exposure.
Far more challenging have been efforts to alter the natural history of TED. Except in a relatively small subgroup of patients, neither corticosteroids nor surgery alters the active phase of the disease, which typically persists for 12 months in nonsmokers and as long as 36 months in smokers. The longer the active phase persists, the greater the accumulated functional and cosmetic deformities and the less likely they are to reverse spontaneously. While we await treatments that will restore normal immune function and thereby avoid or abbreviate the course of TED, the only successfully tested therapeutic alternative to corticosteroids is collimated fractionated external beam orbital radiotherapy (ORT).
What is the rational basis for orbital radiotherapy?
A number of in vitro studies provide insight into the mechanism of ORT. It has been proposed that the major effect of ORT is reduction in orbital inflammation through a reduction in activated lymphocytes. Although appealing in it is simplicity, the explanation fails to account for the anticipated repopulation of the orbit with newly recruited lymphocytes after completion of treatment. In addition, while lymphocytes in general are radiosensitive, activated lymphocytes are relatively radioresistant to low doses of radiotherapy. As an alternative explanation, ORT induces terminal differentiation of progenitor fibroblasts (1). This phenotypic change may blunt the downstream consequences of fibroblast activation by reduction of glycosaminoglycan deposition, fat cell differentiation, and the inflammatory response. This model is particularly appealing because pluripotent orbital fibroblasts are involved in both the local inflammatory cascade and the proliferation of fat cells through differentiation of preadipocytes. In addition, radiotherapy in low doses has been shown to decrease adhesion of blood-borne lymphocytes to activated endothelial cells and reduce the secretion of proinflammatory cytokines from activated lymphocytes.
Is radiotherapy safe?
Charged with the mission to “first do no harm,” there are 2 areas for consideration. First is the question of radiation-induced tumorigenesis. Studies that have examined this issue have failed to demonstrate an increase in the rate of tumors in the treated field over a period of 40 years (2). The question of radiation-induced retinopathy is more difficult to answer, but it is a rare occurrence. Because of the retinopathic effects of diabetes mellitus and hypertension, we generally avoid ORT in these clinical settings.
What evidence is there of the value of orbital radiotherapy?
Both prospective and retrospective studies have examined the therapeutic value of ORT and still much uncertainty remains. A review of the studies prompts consideration of an “apples to apple's” comparison. While we presumably are examining the effect of the same treatment on the same disease, in truth, all published studies vary in their study metrics and inclusion criteria. This degree of inhomogeneity makes it difficult to compare outcomes and perform a meta-analysis of the data.
First, let us examine the study metrics. The NOSPECS classification of TED was never intended as a measure of disease severity or response to therapy. By using a total NOSPECS score to assess the efficacy of treatment, some therapeutic effects may be camouflaged or inappropriately highlighted. The Clinical Activity Score (CAS) has been developed by EUGOGO (European Group on Graves Orbitopathy) and measures disease activity. While a significant improvement over the NOSPECS score, it is predicated on response to corticosteroid therapy as the gold standard for disease activity and is not designed to predict therapeutic response. Hopefully, groups such as ITEDS (International Thyroid Eye Disease Society) and EUGOGO will resolve these difficulties in the near future. However, these scoring systems do provide a measure of therapeutic effect upon a particular clinical feature (e.g., motility, optic neuropathy) of TED.
Second, case selection is of paramount importance. Most patients with TED have limited therapeutic window (days to months) during which the natural history of the disease can be altered. Beyond this, we can only treat the patient symptomatically and strive to avoid adverse effects of the disease. At present, we do not know the optimal time to begin treatment; there is a general agreement that as the patient approaches the end of the acute phase of disease, the less likely will be the impact of treatment. Although most studies report treatment interventions in early phase of disease, there are no standardized methods (e.g., clinical, serologic, imaging) to quantify response to therapy.
Another case selection issue is the form of disease to be treated. While the majority of patients given ORT have severe inflammatory disease or TON, the severity of these cases has precluded a double-masked, therapeutic clinical trial with a placebo arm. As a consequence, we rely upon a number of retrospective studies. Petersen et al (3) reviewed 311 unclassified patients and reported that treatment with ORT resulted in arrested progression of TED in 95% of cases and eliminated the need for corticosteroids in 76%. Kazim et al (4) reviewed 84 patients with moderate-to-severe TED. The study focused on patients with TON. In 94% of patients treated with ORT, orbital decompression was avoided. In patients treated with corticosteroids, only 63% avoided surgery. Of note, there was no significant improvement in proptosis or strabismus.
All previous prospective studies have examined only mild-to-moderate TED. Perhaps as a result, the therapeutic effect is less compelling. Prummel et al (5) performed a randomized trial comparing a 3-month course of prednisone and sham ORT with placebo and 20 Gy ORT and found comparable (50%) improvement in motility and soft tissue swelling, the highest in NOSPECS class. The group treated with prednisone improved more quickly. Mourits et al (6) performed a double-masked randomized trial on 30 patients comparing ORT with sham ORT treatment. Inclusion criteria were NOSPECS class 2–4, and no TON patients were included. Sixty-three percent of ORT-treated patients and 31% of sham-treated patients had improvement in NOSPECS score. A significant improvement in ocular motility was noted in the treated group such that 25% were spared strabismus surgery. This result was confirmed by a more recent trial by Prummel et al (7) in a follow-up study.
The experience of Mayo Clinic (8) published in 2001 stands in stark contrast to the previous reports. In this study, one orbit of each patient received 20 Gy ORT, whereas the fellow orbit was sham treated. Six months later, the orbits were retreated in reverse. At each treatment interval and 6 months after the second treatment, the disease state was assessed by measuring proptosis, ocular motility, lid retraction, and CAS. Patients with TON were excluded. Although the study found that orbital ORT failed to provide benefit compared with sham treatment, this conclusion has been challenged. Patients in the study had TED for an average of 1.3 years, which, as noted above, is beyond the usual range of active TED. The measures of disease activity failed to worsen in the sham-treated group during the first 6 months of observation, supporting the belief that these patients were already in a stable phase of TED. Therefore, it is no surprise that the treated group showed no improvement.
The Ophthalmic Technology Assessment of the American Academy of Ophthalmology (AAO) (9) has concluded that ORT is safe and results in improved ocular mortality in patients with acute phase TED but has demonstrated little benefit for proptosis, lid retraction, and soft tissue swelling. This publication notes that there are no studies that have investigated the efficacy of ORT on TON. Although we lack class I evidence regarding the efficacy of ORT for TED, we must draw our conclusions from the best data available. In my opinion, ORT should not be used in the stable phase, near-stable phase, or in cases of mild TED. It is warranted in cases of progressive limitation of eye movements, and I use ORT with corticosteroids in cases of acute phase TON, to try to avoid surgical decompression.
CON—An Argument Against Orbital Radiation Therapy for Thyroid Eye Disease James A. Garrity, MD
Outcomes and cost have become hot-topic items regarding medical care. As an equation: value = outcomes/cost. Outcomes also include quality, safety, and service. This equation can also be applied to ORT for the treatment of TED. By many counts, there is limited value for this form of therapy as will be detailed in this “Con” argument.
From an historical perspective, radiation therapy for TED initially was directed at the pituitary gland (10). It was speculated that spillover radiation from the poorly collimated beams actually “treated” the orbits. Early reports of directed ORT describe dramatic improvement especially in patients with “marked clinical edema or intensive cellular infiltration on biopsy.” The response was, in general, proportional to the severity of the orbitopathy. The goal of ORT was “to obtain regression before fibrosis supervenes” (11). It is probably fair to say that this still remains a goal of ORT.
Several questions remain; however, the most obvious of which is, “Is ORT effective in TED?” Earlier studies were retrospective, had imperfect measurements, inconsistent reporting of results, confounding treatment variables, and lacked controls. All of these have to be considered in light of a disease process that is usually characterized by spontaneous improvement.
In 2001, investigators from the Mayo Clinic (8) led a double-blinded, randomized clinical trial with internal controls involving 42 patients with moderately severe TED. The study was designed to answer 3 questions: 1) Is 20 Gy ORT effective? 2) Does it make a difference with early vs late treatment? and 3) Are there any complications from ORT? The trial was also designed so that only one orbit was treated initially, and then 6 months later the contralateral orbit was treated. Thus, the untreated orbit served as an internal control. Retinal fluorescein angiography was performed at study entrance and study exit to examine for radiation retinopathy. Patients had a mean CAS of 6.2 (range, 3–10) using the 10-point scale of Mourits et al (12). Exclusion criteria were diabetes mellitus, steroid use within the 2 weeks before evaluation, previous orbital decompression, or presence of an optic neuropathy. At 6 months, there were no clinically or statistically significant differences between treated and untreated orbits for any of the 6 parameters studied. These included extraocular muscle volume, orbital fat volume, combined orbital fat plus extraocular muscle volume, proptosis, eyelid fissures, or monocular range of motion. In an accompanying editorial (13), the study was criticized for patient selection on the basis of time from onset of eye symptoms to study entrance (mean, 1.3 years; range, 0.2–16.0 years); that the control orbit did not change over 6 months, implying stable, inactive disease; and that several of the patients had been previously treated with steroids. In response, a subset analysis of patients with median time from onset of eye symptoms to study entrance less than 1.3 years was compared with the group with median time greater than 1.3 years; once again, there was no significant difference in any parameter. It was mentioned that a positive steroid response often predicts a positive ORT response, and it was assumed that many of the study patents had been steroid failures, which would imply that they would be ORT failures. At the 1-year follow-up evaluation, extraocular muscle volume and proptosis improved slightly more in the orbits that were initially treated with ORT, but these changes were of minimal clinical significance. They may have resulted either from treatment or from the disorder's tendency to spontaneously remit over time (14).
In a 3-year follow-up report of the Mayo Clinic study (15), 8 of 42 patients subsequently had undergone orbital decompression, 11 had strabismus surgery, and 18 had eyelid surgery. Although formal quality-of-life measurements were not part of this study, patients unanimously and spontaneously expressed gratitude for the perceived benefits of ORT despite the absence of any objective change. One should not overlook the possibility of a powerful placebo effect for ORT. When study-exit fluorescein angiograms were examined, retinal microvascular changes consistent with radiation retinopathy were noted in 5 eyes of 3 patients, although these patients were asymptomatic.
Prummel et al (7) compared 20 Gy ORT with sham ORT in patients with mild TED using major criteria (i.e., change of 8° monocular range of motion, change of one or more grades in diplopia score, and change in visual acuity of 1 or more lines) and minor criteria (i.e., 2 mm change in lid aperture, 2 mm change in proptosis, and change of 1 or more grades in soft tissue involvement). Patients with mild TED were specifically chosen “to determine whether ORT is a better option than the ‘wait and see’ policy that is usually adopted in these patients” (16). None of the 88 patients had received any previous treatment for TED, and the mean duration of TED before treatment was 15 months (range, 5–156 months). Treatment was given to 44 patients and 44 controls, and at 6 months, there was no difference between the 2 groups. At 12 months, however, the treatment group had a 52% (23/44) positive response rate vs 27% (12/44) in the sham treatment group for ocular motility and diplopia score. This same relative difference persisted when the duration of TED was less than 18 months, because a successful outcome was recorded in 58% (15/26) of the ORT group and 20% (5/25) in the sham treatment group. The need for additional treatment was lower in the ORT group (15/44 needed no additional treatment) than in the sham treatment group (7/44 needed no additional treatment). Notably, the frequency of worsening TED was similar in both groups. At the 12-month examination, changes on all quality-of-life subscales were similar between the 2 groups. Unfortunately, quality-of-life data were available for only 65 patients, which reduced the statistical power to compare the 2 groups. The authors also considered the total cost of treatment between the 2 groups, and ORT did not reduce treatment costs. The final conclusions of this study were that while ORT was an effective treatment of mild ophthalmopathy, the lack of effect on quality of life or treatment costs makes careful observation a good alternative strategy.
In addition, there are 5 recent observational studies including 947 total patients that should be considered. Radiation dosage was 20 Gy in 3 reports (2,17,18), from 16.8 to 24 Gy in 1 report (19), and up to 30 Gy in 1 report (20). Concomitant oral steroids were given in one study and to some of the patients in 2 other studies, and steroid use was not specified in the remaining 2 studies. The severity of TED was not included in 2 studies, severe in one, moderate to severe in another, and was described using the NOSPECS classification of TED in the last report. None of the studies mentioned the duration of TED before treatment.
Outcomes centered on clinical response for 2 studies, whereas the remaining studies focused on complications from treatment. An outcome study by Bartalena et al (21) reported results on 150 TED patients treated with both oral prednisone at an initial dose of 80–120 mg tapered over 6 months and ORT (20 Gy). Although treatment response criteria were not explicitly detailed and the length of follow-up not specified, an excellent or good response was noted in 40%, a moderate response was noted in 39%, 17% were unchanged, and 4% had worsened. No side effects of treatment were mentioned.
Marquez et al (20) reported on ORT treatment results in 184 of 453 TED patients treated with ORT doses ranging from 20 to 30 Gy. The NOSPECS score was used as the primary outcome, and statistically significant improvement was noted in all 5 NOSPECS categories, with a median of 7.2 years of follow-up. Complete soft tissue response was noted in 59%, complete proptosis response in 44%, and complete motility response in 44% of patients. Cataracts developed in 12% of patients, and solid tumors (outside of the field of radiation) developed in 5% of patients. Radiation retinopathy was noted in 2 patients, both of whom had received 30 Gy ORT. A high number of patients (n = 269) had no follow-up, which may have influenced the study results.
There were 3 observational studies that concentrated on complications; Schaefer et al (19) followed up 250 patients treated with 16.8–24 Gy per course (some treated with up to 3 courses) for a median of 31 years. Survival was no different from untreated patients, and deaths from malignancy were unrelated to tumors in the field of radiation. Another study by Wakelkamp et al (2) compared 159 patients treated with 20 Gy ORT (also treated with oral steroids) and 86 patients treated with oral steroids alone. After a mean follow-up of 11 years, 17% of ORT-treated patients and 12% of non–ORT-treated patients had died; this difference was not statistically significant. Radiation-related complications were either cataract or radiation retinopathy. Cataracts were seen in 29% of ORT-treated patients and 35% of patients treated with oral steroids. Radiation retinopathy was seen in 2% of ORT-treated patients and none of the steroid-treated patients.
The AAO issued a technology assessment in 2007 entitled Orbital Radiation for Graves Ophthalmopathy (9). The committee reviewed 14 studies (5 observational studies and 9 randomized clinical trials). They also considered cost, which at some institutions can exceed $15,000 for a 20-Gy course for a third-party payer and $3,500 for Medicare coverage. The report noted that evidence-based assessment of the role of ORT in the management of non–sight-threatening TED was limited by the heterogeneity and variable quality of published reports; however, data from randomized clinical trials suggest that ORT has a limited role in treating non–sight-threatening TED.
I would conclude that in non–sight-threatening TED, ORT has a negligible effect on proptosis or lid position. Motility may improve in some patients, but whether these changes translate into improved health-related quality of life is not clear. The only study to address this had inconclusive results (7). When one considers the cost, outcomes, potential for complications, and subsequent need for additional therapy, I agree with the AAO technology assessment that “ORT should serve in only rare circumstances, and further studies should evaluate more appropriate anti-inflammatory and immunomodulatory therapy” (22).
Rebuttal: Michael Kazim, MD
Dr Garrity's position on the role of ORT to treat TED is well considered and provokes further consideration. I believe that we are in agreement that the majority of TED patients should not be treated with ORT as they have mild, moderate or stable disease.
It is interesting to consider how our respective patient populations influence our therapeutic approach to TED. Due to the combination of geography and referral patterns of patients with TED seeking care at the Mayo Clinic, it is more likely they are in a stable phase in search of surgical rehabilitation. In addition, even if ORT is recommended, these patients may be unable to stay for the requisite two weeks of daily delivery of fractionated ORT. In contrast, patients seen in my consultative practice are almost exclusively in the early acute phase of disease. My recommendations focus on available therapeutic modalities to limit the morbidity and length of acute TED.
Currently, is only retrospective data to support ORT for the treatment of TON. At our medical center this form of treatment spares imminent, acute phase orbital decompression in 90 % of patients with TON. While surgical, anesthesia, and hospital fees vary, I would conservatively suggest that the $3,500 – $15,000 cost for ORT in these patients is far exceeded by those associated with the surgery. Added to the higher costs of surgery would also be those of surgically induced strabismus and any surgical complications such as recurrent sinusitis that can occur in some patients.
In summary, I believe Dr Garrity and I agree more that we disagree. While I would caution against the general use of ORT in TED patients, there is a subset of patients who benefit from this form of treatment.
Rebuttal: James A. Garrity, MD
Until the fundamental cause of TED is found, treatment is primarily directed at clinical symptoms. Dr. Kazim opines that if we can shorten the “active phase” of the disease, presumably with ORT, then the functional and cosmetic deformities can be ameliorated. Additionally, for ORT to be effective, it is optimally given during a brief, narrow therapeutic window, which is undefined and difficult to determine. As a critique to the Mayo Clinic study, Dr. Kazim points out that the disease was probably stable because the control orbit did not change CAS and median time to treatment from onset of eye symptoms to onset of ORT was 1.3 years (range, 0.2–16.0 years). However, length of disease duration, often difficult to quantify precisely, is probably less relevant than the CAS. The mean score was 6.2 (range, 3–10) and is consistent with most would consider “active” disease. Randomized controlled studies will be required to definitively address and assess the efficacy of ORT as treatment for TED. The published studies to date suggest that ORT's value, defined as outcomes per cost, is yet to be clearly established.
Conclusion: Andrew G. Lee, MD, and Valérie Biousse, MD
The indications for treating patients with TED with any therapeutic modality (e.g., immunosuppression, immunomodulation, ORT, or surgery) are severity and activity of disease. The studies that have treated patients with ORT who have inactive, stable, or fibrotic TED are less likely to show therapeutic benefit, and ORT should not be used for mild TED and “white, quiet eyes” in the fibrotic stage of Rundle curve. It remains controversial if ORT should be used for symptomatic patients with active inflammatory disease, to stabilize patients who cannot or will not take corticosteroids, to obviate the need for surgical decompression in patients with TON, or in patients who are not surgical candidates. This Point-Counter Point discussion may have succeeded in raising more questions than answers!
1. Trott KR, Kamprad F. Radiobiological mechanisms of anti-inflammatory radiotherapy. Radiother Oncol. 1999;51:197–203.
2. Wakelkamp IMMJ, Tan H, Saeed P, Schlingenmann RO, Verbraak FD, Blank LECM, Prummel MF, Wiersinga WM. Orbital irradiation for Graves' ophthalmopathy: is it safe? A long-term follow-up study. Ophthalmology. 2004;111:1557–1562.
3. Petersen IA, Kriss JP, McDougall IR, Donaldson SS. Prognostic factors in the radiotherapy of Graves' ophthalmopathy. Int J Radiat Oncol Biol Phys. 1990;19:259–264.
4. Kazim M, Trokel S, Moore S. Treatment of acute Graves' orbitopathy. Ophthalmology. 1991;98:1443–1448.
5. Prummel MF, Mourits MP, Blank L, Berghout A, Koornneef L, Wiersinga WM. Randomized double-blind trail of prednisone versus radiotherapy in Graves' ophthalmopathy. Lancet. 1993;342:949–954.
6. Mourits MP, van Kempen-Harteveld ML, Garcia MB, Koppeschaar HP, Tick L, Terwee CB. Radiotherapy for Graves' orbitopathy: randomized placebo-controlled study. Lancet. 2000;355:1505–1509.
7. Prummel MF, Terwee CB, Gerding MN, Baldeschi L, Mourits MP, Blank L, Dekker FW, Wiersinga WM. A randomized controlled trial of orbital radiotherapy versus sham irradiation in patients with mild Graves' ophthalmopathy. J Clin Endocrinol Metab. 2004;89:15–20.
8. Gorman CA, Garrity JA, Fatourechi V, Bahn RS, Petersen IA, Stafford SL, Earle JD, Forbes GS, Kline RW, Bergstralh EJ, Offord KP, Rademacher DM, Stanley NM, Bartley GB. A prospective, randomized, double-blind placebo-controlled study of orbital radiotherapy for Graves' ophthalmopathy. Ophthalmology. 2001;108:1523–1534.
9. Bradley EA, Gower EW, Bradley DJ, Meyer DR, Cahill KV, Custer PL, Holck DE, Wong JJ. Orbital radiation for Graves ophthalmopathy: a report by the American Academy of Ophthalmology. Ophthalmology. 2008;115:398–409.
10. Beierwaltes WH. X-ray treatment of malignant exophthalmos: a report on 28 patients. J Clin Endocrinol. 1953;13:1090–1100.
11. Jones A. Orbital X-ray therapy of progressive exophthalmos. Brit J Radiol. 1951;24:637–646.
12. Mourits MP, Koornneef L, Wiersinga WM, Prummel MF, Berghout A, van der Gaag R. Clinical criteria for the assessment of disease activity in Graves' ophthalmopathy: a novel approach. Br J Ophthalmol. 1989;73:639–644.
13. Feldon SE. Radiation therapy for Graves' ophthalmopathy: trick or treat? Ophthalmology. 2001;108:1521–1522.
14. Gorman CA. Radiotherapy for Graves' ophthalmopathy: results at one year. Thyroid. 2002;12:251–255.
15. Gorman CA, Garrity JA, Fatourechi V, Bahn RS, Petersen IA, Stafford SL, Earle JD, Forbes GS, Kline RW, Buettner H, Robertson DM, Bergstralh EJ, Offord KP, Rademacher DM, Stanley NM, Bartley GB. The aftermath of orbital radiotherapy for Graves' ophthalmopathy. Ophthalmology. 2002;109:2100–2107.
16. Wiersinga WM, Prummel MF. An evidence-based approach to the treatment of Graves' ophthalmopathy. Endocrinol Metab Clin North Am. 2000;29:297–319.
17. Bartalena L, Marcocci C, Tanda ML, Manelti C, Dell'Unto E, Bartolomei MP, Mardi M, Martino E. Cigarette smoking and treatment outcome in Graves ophthalmopathy. Ann Intern Med. 1998;129:632–635.
18. Marcocci C, Bartalena L, Rocchi R, Marino M, Mencoci F, Morabito E, Mazzi B, Mazzeo S, Sartini MS, Nardi M, Cartei F, Cionini L, Pinchera A. Long term safety of orbital radiation therapy and Graves ophthalmopathy. J Clin Endocrin Metab. 2003;88:3561–3566.
19. Schaefer U, Hesselmann S, Micke O, Schueller P, Bruns F, Palma C, Willich N. A long-term follow-up study after retro-orbital irradiation for Graves' ophthalmopathy. Int J Radiat Oncol Biol Phys. 2002;52:192–197.
20. Marquez SD, Lum BL, McDougall IR, Katkuri S, Levin PS, MacManus M, Donaldson SS. Long-term results of irradiation for patients with progressive Graves' ophthalmopathy. Int J Radiat Oncol Biol Phys. 2001;51:766–774.
21. Bartalena L, Marcocci C, Gorman CA, Wiersinga WM, Pinchera A. Orbital radiotherapy for Graves' ophthalmopathy: useful or useless? Safe or dangerous? J Endocrinol Invest. 2003;26:5–16.
22. Perry JD, Feldon SE. Rationale for radiotherapy in thyroid eye disease. Am J Ophthalmol. 2009;148:818–819.