Extranodal Marginal Zone Lymphoma of Ocular Adnexa: Outcomes following Radiation Therapy

Marginal zone B-cell lymphomas (MZL) are a group of lymphomas that originate from B-cell lymphocytes in the marginal zone of secondary lymphoid follicles [1]. MZL accounts for between 5 and 17% of all non-Hodgkin lymphomas (NHL) [2,3,4,5,6]. The International Lymphoma Study Group classified 3 subtypes of MZL. This includes extranodal marginal zone B-cell lymphoma (EMZL), splenic marginal zone B-cell lymphoma, and nodal marginal zone B-cell lymphoma [1,2,7,8]. EMZL is the most frequent, low-grade, indolent, small B-cell lymphoma with a prolonged course of the ocular adnexa [9,10]. Ocular adnexal lymphomas (OAL) of the EMZL type have a high response rate to various treatment modalities, such as radiation therapy, chemotherapy, and immunotherapy [2,9,10,11,12,13].

As EMZL is an indolent form of lymphoma, radiotherapy has long been the treatment of choice for localized disease [14,15,16]. Long-term survival and excellent local control have been documented with radiotherapy alone [4,9,10,14,15]. It is estimated that using low to moderate radiotherapy doses (25-36 Gy) can obtain 95-100% of local control [4,11,17]. Few authors have advocated radiotherapy doses of 25 Gy or less [16,18,19]. A recent British National Lymphoma Investigation randomized trial confirmed excellent local control rates at 24 Gy for indolent NHL and 30 Gy for aggressive NHL [20].

However, the optimal radiation dose that will achieve a high local control rate with a minimal risk of visually significant complications for the treatment of EMZL type OAL is not well known. We investigated the dose-response relationship vis-à-vis radiation complications in published studies and evaluated outcomes of a single institution cohort followed using a standard protocol.

After approval by our institutional review board, the study patient population was identified from our patient database at the Cleveland Clinic from January 1st, 1997, to December 31st, 2015 (19 years). The diagnosis was established in all patients by tissue biopsy performed at the Cleveland Clinic or reviewed by our pathology department if biopsy was performed elsewhere. In the setting of concurrent systemic involvement, biopsy from another involved site was considered adequate for diagnosis of ocular involvement. A complete history and ophthalmic exam was performed on all patients. Our assessment included best-corrected visual acuity, intraocular pressure, external examination (exophthalmometry and motility testing), slit-lamp examination, and dilated fundus exam. Slit-lamp, external, and fundus photography were routinely performed. B-scan ultrasonography and optical coherence tomography were obtained in patients with uveal involvement. All patient charts were reviewed for clinical features (age, sex, age at diagnosis, and details of lymphoma involvement), treatment type, radiation dose, fractionation, recurrence (local vs. distant), and complications from the radiation treatment. All information was placed on a secured, password-protected, encrypted database.

Since 2005, the Cleveland Clinic protocol for the EMZL type of OAL has been to treat unilateral ocular-only involved eyes with radiotherapy and bilateral ocular-only disease or systemic disease with single-agent rituximab. The average and median doses of radiation were 26.5 and 25.2 Gy, respectively. The range was 20-36 Gy delivered in fractions (150-200 cGy per fraction). Over the years, concomitant with innovations in the field of radiation oncology, radiation therapy was administered using 3 different techniques: external-beam radiation therapy, involved-field radiation therapy, and image-guided intensity-modulated radiation therapy (IGIMRT). Indolent lymphomas anterior to the orbital septum underwent partial orbital irradiation using a direct electron field and daily bolus; retrobulbar and lacrimal-gland OAL underwent whole orbital irradiation utilizing 3-dimensional conformal radiation therapy techniques (Fig. 1). Since 2014, we have employed IGIMRT based on the International Lymphoma Radiation Oncology Group consensus report [21].

Local control was defined as resolution of symptoms/signs at presentation supported by imaging studies. Recurrence was defined as lymphoma relapse after local control had been achieved and was divided into 2 categories: local and distant. Any recurrence that was in the field of radiation was considered a local recurrence. Any recurrence out of the field of radiation was considered a distant recurrence, which can further be categorized as the contralateral untreated eye or systemic. Patients who received radiation therapy only were included to control for the effects of chemotherapy or immunotherapy in combination with radiation therapy on the tumor recurrence data. A review of the literature for patient outcomes related to radiation treatment was performed (Table 1). Local control rate versus median dose and local control rate versus average follow-up were plotted.

All patients who were treated with radiation for EMZL type of OAL and who had at least 6 months of follow-up were included in this analysis.

Acute reactions to the radiation were not assessed in this study, which included common findings of radiation dermatitis, keratitis, and dry eyes, as these conditions are self-limiting and symptoms are easily treated by noninvasive measures. Visually significant complications analyzed were cataract formation/progression, radiation retinopathy, and optic neuropathy.

Radiation Cataracts

All eyes treated with radiation therapy were included in this analysis. Radiation cataract was defined as a latent response to any form of radiation, which may cause an early progression of the lens in any layer. Included patients were documented to have asymmetric, visually significant lenticular changes in the treated eye when compared to the contralateral eye. The presence of posterior subcapsular opacity in the absence of other causes was considered adequate to diagnose radiation-induced lenticular changes. By including patients who only had radiation therapy in 1 eye and excluding patients who had prior cataract surgery we could establish a control group. Here, we analyzed the incidence of cataract formation in the treated phakic eye versus the untreated phakic eye of the same individual.

Radiation Retinopathy/Optic Neuropathy

All eyes treated with radiation were included in this analysis. Findings of radiation retinopathy included retinal microaneurysms, retinal hemorrhages, retinal exudates, cystoid macular edema, cotton-wool spots, retinal neovascularization, or vitreous hemorrhage.

Retinal Tear/Detachment

Patients who sustained a rhegmatogenous retinal detachment after radiation treatment were identified.

A review of the literature for patient outcomes related to radiation treatment was performed. Twenty peer-reviewed articles were identified with similar patient populations (randomized trial for EMZL of the ocular adnexa). Local control rate versus median dose and local control rate versus average follow-up were plotted.

Outcomes were recorded from the date of tissue diagnosis to the date of the first event or, if no events occurred, the last follow-up time. For treatment-related analysis, outcomes were recorded from the date of treatment start to the date of the first event. The entire statistical analysis was conducted in R, version 3.2.3. Control rate was plotted against dose, as well as follow-up time. Each data point, i.e., study, was weighted by the number of patients included. Based on these weights, the line of best fit was estimated and plotted.

A total of 81 patients with 103 involved eyes seen at the Cleveland Clinic from January 1997 to December 2015 were identified with biopsy-proven EMZL type of OAL. Twenty-one patients were excluded because the patient was lost to follow-up after treatment or the patient continued treatment elsewhere after diagnosis. In total, 77 eyes of 60 patients were treated with radiation, which included 6 eyes with simultaneous uveal involvement (8%). Of the 60 total patients treated with radiation, there was a predominance of male patients; 39 patients were male (65%) and 21 patients were female (35%) with a mean age of 65 years. The average length of follow-up was 38.6 months with a range of 1-94 months and a median of 27 months (Table 2).

Forty-seven of the 60 patients (57 eyes) who were treated with radiation therapy only were analyzed for local control. All these were staged as IE (Ann Arbor staging system) at initial evaluation. None of these patients were identified with local recurrence during their follow-up period. However, 6 patients (13%) subsequently developed recurrence at a distant site; either systemic only (n = 5) or contralateral eye and systemic recurrence (n = 1).

The charts of 60 patients with OAL only who were treated with radiation therapy and who had at least 6 months of follow-up were reviewed to compare the visual outcomes. Phakic and pseudophakic patients were included. Three patients had 20/100 or worse vision at presentation. This includes 1 patient each with amblyopia, central retinal artery occlusion, and compressive optic neuropathy. Only 5 patients had 2 or more lines of vision loss after radiation therapy. In 3 of these patients, the visual decline was associated with cataract formation/progression and in 1 was due to dry eye.

Radiation Cataract

Radiation-induced cataract was the most common complication after radiation therapy to the globe and orbit. Nineteen of the 77 eyes (25%) that underwent radiation therapy developed a radiation-induced cataract. Mean radiation dose and fractions were 29.27 Gy and 16, respectively. The average number of months from treatment start to diagnosis of a radiation cataract was 40 months with a range of 2.5-147 months. The average change in logMAR from treatment start to diagnosis of radiation cataract was 0.26. To adequately assess the degree of radiation effect on the lens, we grouped our patients into those who were phakic in both eyes and only received radiation therapy to only 1 eye. There were 25 patients that fitted this description, and 10 patients developed a visually significant cataract in the treated eye compared to 1 cataract in the untreated eye. The average change in logMAR in the treated eye was 0.31 compared to 0.05 in the untreated eye group.

Radiation Retinopathy/Optic Neuropathy

After a careful review of the charts, only 1 patient out of 77 included patients that received radiation therapy was observed to have radiation retinopathy. This patient developed cystoid macular edema, retinal hemorrhages, and neovascularization of the iris and was subsequently treated with panretinal photocoagulation and bevacizumab injections. Although this patient was a poorly controlled diabetic, similar findings were not seen in the untreated eye, confirming the presence of radiation retinopathy. Six years after radiation therapy, visual acuity was counting fingers with chronic cystoid macular edema, hard exudates, and atrophic scar. Anti-VEGF injections were discontinued 3.5 years prior because of poor response to therapy.

Retinal Tear/Detachment

Five patients were identified who sustained a rhegmatogenous retinal detachment after radiation treatment. Three patients were excluded, 1 patient developed a retinal detachment after cataract surgery, 1 in the untreated eye, and the other patient developed a retinal detachment attributable to ciliary body biopsy. Two patients developed a retinal tear/detachment during the radiation treatment and in the absence of an intraocular biopsy.

One patient had a new posterior vitreous detachment, and the other patient had a preexisting posterior vitreous detachment. A complete dilated exam was performed in both patients, and no retinal tears or other peripheral retinal abnormalities were observed prior to radiotherapy treatment. Both patients had hyperopia. One patient had OAL, and the other patient had concomitant OAL and uveal lymphoma. There was no significant difference in radiation dosages (24-25.2 Gy) between these cases and the rest of the cohort.

Twenty peer-reviewed articles were reviewed from 1997 to 2013 [4,8,12,14,15,16,22,23,24,25,26,27,28,29,30,31,32,33,34,35]. All patients treated with radiation therapy only were recorded. When control rate was plotted against radiation dosage (Gy), the estimated line of best fit (when weighted by number of subjects in the study) was 30 Gy (Fig. 2a). When the control rate was plotted against follow-up time, an inversely proportional relationship was observed (Fig. 2b). Fractionation range for ophthalmic radiation varies from 150 to 200 cGy per fraction in standard clinical practice. Published reports did not all provide fractionation information.

In our study, we analyzed our 19-year radiation therapy treatment outcomes for biopsy-proven EMZL type OAL. We identified 77 eyes that were treated with radiation therapy, and 57 were treated with radiation therapy only. Local control rate was 100% in eyes treated with radiation therapy only with the mean dose of 26.5 Gy (median 25.2 [range 20-36] Gy). Results for local control are given only for those that received radiation therapy only, removing confounding effects of systemic therapy. When we plotted the control rate against the average dose data from 20 peer-reviewed articles with a similar patient population, our line of best fit, weighted by the number of subjects in the study, was at 30 Gy (Fig. 2a). When we plotted the local control against follow-up time, the line of best fit demonstrated an inverse trend (Fig. 2b). In our study sample, we achieved 100% of local control with an average dose of 26.5 Gy and a median of 25.2 (range 23.4-36) Gy. Fractionation range for ophthalmic radiation varies from 150 to 200 cGy per fraction in standard clinical practice. Published reports did not all provide fractionation information. Lower radiation doses (such as 4 Gy in 2 fractions) have also been evaluated for patients with indolent lymphomas, such as marginal zone lymphoma and follicular lymphoma of nonocular sites [36]. In a randomized phase 3 noninferiority trial, local progression-free survival was inferior with 4 Gy (2 fractions) compared to 24 Gy (12 fractions), establishing the higher dose as the standard of care [37].

Overall, there was a minimal risk of visually significant complications. Visual outcomes after radiation therapy for EMZL type OAL demonstrate the minimal effect of the radiation on visual potential. Only 4 patients lost 2 lines or more of vision after radiation. Few studies have reported severe radiation retinopathy with a total radiation dose of <40 Gy as observed in 1 of our patients (34 Gy in 20 fractions) [6,38]. Radiation-induced toxicity to the retina has been documented in 5 and 50% of patients at 5 years with fractionated doses of 45-50 and 55 Gy, respectively [39]. Our patient, who was diabetic, lost vison (final acuity of counting fingers) due to cystoid macular edema, cotton-wool spots, retinal hemorrhages, and neovascularization of the iris requiring full panretinal photocoagulation treatment and despite multiple anti-VEGF intravitreal injections. Diabetes has been reported to increases the risk of radiation retinopathy [38]. In our patient, development of unilateral radiation retinopathy may have been exacerbated by diabetic status.

Radiation-induced retinal tear or detachment is not a recognized risk from radiation therapy. Two (2%) of our patients were diagnosed with a new retinal tear or detachment during radiation therapy. None of the patients were myopic or had evidence of peripheral retinal abnormalities considered as risk factors for retinal tears at baseline examination. Even though the effects of radiation therapy on the vitreous are not known, our data suggest that the risk of a retinal tear or detachment from radiation treatment should be discussed with the patient given its potential for vision loss.

A limitation to our study is that this is a retrospective review rather than a prospective study. Our oncology service at the Cleveland Clinic was created in 2002. Prior to this time, many patients were lost to follow-up or they obtained treatment elsewhere (16 out of 97 patients), which is a significant proportion of our initial population. As shown in Figure 2b, there is an inverse relationship between local control rates and average length of follow-up time. Our average follow-up was 38 months and 100% local control. Longer follow-up times will be needed to see if our local control data follow a similar trend.

Over the years, concomitant with innovations in the field of radiation oncology, radiation therapy was administered using 3 different techniques: external-beam radiation therapy, involved-field radiation therapy, and IGIMRT. Strictly on theoretical grounds, IGIMRT, the current standard in the US and Europe, is not expected to influence disease control but is expected to reduce associated short- and long-term morbidity [40].

In conclusion, our data suggest that 25 Gy (median dose) is an effective (fractionated) radiation dose providing high local control and low risk of visually significant complications. Our findings are consistent with the recently published results of a phase III randomized trial [20] and guidelines from the International Lymphoma Radiation Oncology Group recommendations for extranodal lymphoma of all sites [21].

The study complied with the guidelines for human studies and animal welfare regulations. The subject gave informed consent, and the study protocol was approved by the institute's committee on human research.

The authors have no conflicts of interest to declare.