Radiotherapy for extranodal, marginal zone, B-cell lymphoma of mucosa-associated lymphoid tissue originating in the ocular adnexa

A multiinstitutional, retrospective review of 50 patients

Abstract

BACKGROUND

Due to the small number of patients and differences in the pathologic classification in most radiotherapy series, information regarding the adequacy of tumor control in patients with ocular-adnexal marginal zone B-cell lymphoma of mucosa-associated lymphoid tissue (MALT lymphoma) is limited.

METHODS

A multiinstitutional, retrospective study was performed on 50 patients with Stage IE ocular-adnexal MALT lymphoma who were treated with radiotherapy between 1989 and 1999. The impact of patient characteristics and other variables on tumor control was analyzed.

RESULTS

Responses to radiotherapy include a complete response (CR) in 26 patients, a partial response (PR) in 20 patients, and no change in 4 patients. Forty-nine of 50 patients obtained tumor control in the ocular adnexa at 24 months. Overall, 6 patients exhibited disease recurrence at 4–97 months. Three patients developed recurrence in the ocular adnexa. Two patients had isolated extranodal failure involving the oral floor and the submandibular gland, and one patient experienced failure in the neck lymph node. The initial tumor response had a marginal impact on the development of recurrence. None of the 26 patients who achieved a CR experienced ocular-adnexal recurrence. All three patients who experienced local treatment failure belonged to the initial PR group. In total, five of six patients who developed recurrent disease had obtained a PR after initial radiotherapy. Age, gender, tumor location, and dose of radiotherapy did not influence the development of recurrence. There was only one death due to lymphoma. The 5-year overall survival rate was 91% with a median follow-up of 46 months.

CONCLUSIONS

Radiotherapy offers excellent local control with a prolonged clinical course for patients with MALT lymphoma in the ocular adnexa. The initial response to radiotherapy marginally influenced the probability of recurrence. Cancer 2003;98:865–71. © 2003 American Cancer Society.

DOI 10.1002/cncr.11539

Lymphomas in the ocular adnexa predominantly result from B-cell proliferation and include a variety of histologic types. Although diffuse large cell lymphomas occasionally are encountered, the most common type of ocular-adnexal lymphoma is an extranodal marginal zone B-cell lymphoma of mucosa-associated lymphoid tissue (MALT lymphoma).1, 2 The concept of MALT lymphoma was described first in 1983 by Isaacson and Wright3 and is now well established in the Revised European–American Classification of Lymphoid Neoplasms1 and the more recently published World Health Organization classification system.4 It is recognized that this type of marginal zone B-cell lymphoma arises in lymphoid tissue that has been acquired as a result of some preexisting disorder, such as Helicobacter pylori colonization in the stomach,5 Sjogren disease in the salivary gland,6 and Hashimoto thyroiditis.7 Although patients with MALT lymphoma exhibit an indolent natural history and have a favorable prognosis, it was reported that nongastrointestinal MALT lymphomas showed slightly different behavior from that of MALT lymphomas in the gastrointestinal tract.8 Results from radiotherapy series have demonstrated relatively good local control with a prolonged clinical course in patients with localized orbital lymphoma.9, 10 However, the small patient samples and the differences in pathologic classification may obscure the effects and sequelae of radiotherapy. Thus, there are scant data currently available regarding the prognostic factors that influence the probability of successful tumor control in patients with ocular-adnexal MALT lymphoma.

In the current retrospective study, we assessed the results of radiotherapy in 50 patients from 4 institutions. The objective of the study was to analyze the impact of patient characteristics and other variables on the degree of tumor control in patients with ocular-adnexal MALT lymphoma.

MATERIALS AND METHODS

Between 1989 and 1999, 54 patients with MALT lymphoma of the ocular adnexa were treated at 4 Japanese academic institutions. Four patients were excluded from this study: one patient with Stage II disease, because of simultaneous parotid gland involvement, and three patients who were treated initially with chemotherapy. Thus, the remaining 50 patients with MALT lymphoma originating from and limited to the ocular adnexa were the focus of the current study. The diagnosis was made on the basis of hematoxylin and eosin-stained biopsy specimens (incisional biopsy specimens, 45 patients; excisional biopsy specimens, 5 patients) supported by immunohistochemical analysis. The routine immunohistochemical studies applied for most patients have been described elsewhere.11 Seventy percent of all diagnostic biopsy specimens were reviewed centrally by expert Japanese hematologic pathologists (A. M. and K. I.). Patients were required to fulfill all histologic criteria, as defined previously, to be included in this study.3, 4, 12 All patients underwent a complete clinical and laboratory evaluation and were staged with Stage IE disease according to the Ann Arbor system. These evaluations included computed tomography (CT) scans of the orbit to exclude local spread outside the ocular adnexa; CT scans of the neck, chest, abdomen, and pelvis; gallium scintigraphy; iliac bone marrow biopsy; endoscopic examination of the gastrointestinal tract; and a serum lactate dehydrogenase level. In general, the greatest tumor dimension was measured clinically for conjunctival lesions and was determined on an axial CT slice for tumors in the orbital soft tissue. In patients with bilateral involvement, only the largest lesion was taken into account for measurement. A complete ophthalmic examination, including double eversion of the upper eyelids to examine the upper fornix, was performed in all patients. Patients who had simultaneous bilateral lesions were classified with Stage IE disease. All patients were immunocompetent.

Radiotherapy

All patients were treated with definitive radiotherapy. Sixteen patients, including 14 of 29 patients with conjunctival disease and 2 of 17 patients with orbital soft tissue lesions, were treated with electrons (energy, 6–12 megaelectron volts [MeV]) from a single anterior field. Of these 16 patients, 11 patients were treated with a cylindrical lens shielding approximately 6–12 mm thick, depending on the electron beam energy. Lens shielding was placed 1 cm above the cornea. In 15 of 16 patients, the planning target volume included the entire conjunctiva. The electron field included the local tumor with a generous margin instead of the entire conjunctiva in one patient. Megavoltage radiotherapy with 4–10-megavolt photons or Cobalt-60 γ-rays were used in 34 patients. In these patients, beam alignment was determined by the CT-delineated tumor contour. The field arrangement was individualized. A single lateral field, a lateral oblique field, or multiple fields sparing the lens were used if tumor coverage was not compromised. The field size ranged from 5 cm × 5 cm to 6 cm × 7 cm to include the tumor contour with an adequate margin. Conventional fractionation with a daily dose of 2.0 grays (Gy) was used for all patients. The total radiotherapy dose ranged from 20 Gy to 46 Gy with a median of 36 Gy.

Statistical Analyses

Responses were evaluated 4–6 weeks after the end of radiotherapy. A complete response (CR) was defined as complete resolution of clinical evidence of lymphoma in the ocular adnexa by both physical examination and orbital CT scan. A reduction of at least 50% in tumor size was defined as a partial response (PR), whereas a reduction in tumor size < 50% was defined as no response (NC).

After the completion of radiotherapy, patients were followed at regular intervals. Careful clinical and ophthalmologic examinations were performed every 1–3 months for the first 2 years, every 4–6 months through year 5, and annually thereafter. Orbital CT scanning or magnetic resonance imaging was recommended at 1 year after radiotherapy but was not required. Other radiographic studies were performed as indicated clinically.

In the current study, we analyzed the impact of the following variables on the development of recurrence: age (younger than 60 years or 60 years and older), gender, location (conjunctiva or nonconjunctiva), laterality (unilateral or bilateral), initial response to radiotherapy (CR or non-CR), and the dose of radiotherapy (≤ 30 Gy or > 30 Gy). The recurrence rate in each patient group was compared using the chi-square test for values > 5 (any recurrence) and Fisher exact test for values < 5 (ocular-adnexal recurrence). A probability level of 0.05 was chosen for statistical significance, and reported P values are 2-sided. Overall survival was measured from the start of radiotherapy to death from any cause, with follow-up of surviving patients censored at the last contact date. Estimates of survival distribution were calculated actuarially according to the method of Kaplan and Meier.13 Statistical analysis was performed with the SPSS software package (Version 7.5; SPSS Inc., Chicago, IL).

RESULTS

Patients Characteristics

Table 1 lists patient characteristics. There were 33 men and 17 women, with a median age of 61 years (range, 31–83 years). Only 2 patients had an Eastern Cooperative Oncology Group performance status of 2 or worse. Tumor locations within the orbit included the conjunctiva (n = 29 patients), orbital soft tissue (n = 17 patients), eyelid (n = 2 patients), and lacrimal gland (n = 2 patients). Seven patients had bilateral lesions. No patient exhibited an elevated serum lactate dehydrogenase level or systemic symptoms. The greatest tumor dimension was determined in 36 patients and ranged from 15 mm to 44 mm with a median of 22 mm.

Table 1. Characteristics of 50 Patients with Ocular-Adnexal Mucosa-Associated Lymphoid Tissue–Type Lymphoma

Characteristic	No. of patients
ECOG: Eastern Cooperative Oncology Group. a Data were available for 34 patients.
Age (yrs)
Median (range)	61 (31–83)
≥ 60	28
< 60	22
Gender
Male	33
Female	17
Performance status (ECOG)
0–1	48
≥ 2	2
Location
Conjunctiva	29
Orbital soft tissue	17
Eyelid	2
Lacrimal gland	2
Involved site
Unilateral	43
Bilateral	7
Greatest tumor dimension (mm)a
Median (range)	22 (15–44)

Initial Response, Local Control, and Recurrence

A flow chart of the tumor response in patients with MALT lymphoma of the ocular adnexa after radiotherapy is shown in Figure 1. The initial response rate was 92%, with 26 patients achieving a CR and 20 patients achieving a PR. Four patients exhibited NC. All patients with CR achieved ocular-adnexal tumor control at 24 months. Residual tumors present in 22 of 24 patients with PR or NC exhibited gradual regression after the evaluation of initial response. One patient who developed recurrent tumor in the eyelid at 4 months was salvaged by further radiotherapy. Thus, 49 patients obtained ocular-adnexal tumor control at 24 months. None of the variables, such as age, location, laterality, or dose of radiotherapy, influenced the initial tumor response. The mean dose of radiotherapy administered to patients with CR and non-CR was 36 Gy and 37.5 Gy, respectively (P = 0.44; t test). Gender had a significant impact on the initial tumor response, with a CR rate of 39% in men and 76% in women (P = 0.018). However, the difference in the CR rate between men and women did not translate into any difference in local tumor control or patient survival.

Figure 1. Patterns of initial tumor response, 2-year ocular-adnexal control, and recurrence in all 50 patients with extranodal marginal zone B-cell lymphoma of mucosa-associated lymphoid tissue. CR: complete response; PR: partial response; NC: no change. *: includes 1 patient who experienced out-of-field eyelid failure, which was controlled by additional radiotherapy, at 4 months. **: see Table 2.

Overall, 6 patients exhibited recurrent disease 4–97 months after radiotherapy. Clinical data on those six patients are detailed in Table 2. Three patients developed in-field recurrences in the ocular adnexa, affecting the conjunctiva in two patients (Patients 1 and 2) and the eyelid in one patient (Patient 3). All 3 had an initial conjunctival lesion and were treated with electron beams of 6–15 MeV using a lens shield. One patient (Patient 1) developed a recurrence adjacent to the shielded area at 4 months. She also developed a new lesion in the nonirradiated outer (lateral) conjunctiva of the ipsilateral eye. She was the only patient whose electron field did not include the entire conjunctiva. She currently is followed closely with no active treatment due to her unstable psychopathy. Another patient (Patient 2) developed a recurrence at 97 months and achieved a second CR with radiotherapy using ⁹⁰Sr β-rays. One patient (Patient 3) who developed a recurrence in the eyelid at 4 months was treated successfully with further irradiation and currently is well with no evidence of disease. There were three failures outside the ocular adnexa. Two patients developed isolated extranodal failure (Patients 4 and 5; oral floor [81 months] and submandibular gland [11 months], respectively). Restaging procedures did not detect any evidence of systemic recurrence. Therefore, these patients were treated with localized radiotherapy and were disease-free at 49 months and 18 months after treatment, respectively. One patient (Patient 6) who developed a recurrence in a cervical lymph node was treated successfully with chemotherapy but subsequently died after a second failure in the lungs. The rates of recurrence according to patient characteristics and other variables are shown in Table 3. Patients with initial bilateral involvement exhibited a significantly greater incidence of disease recurrence. Two of seven patients with initial bilateral lesions developed local recurrences. However, those 2 patients were the ones who developed recurrences in the ocular adnexa at 4 months (Patients 1 and 3 in Table 2). The initial tumor response had a marginal impact on the development of subsequent recurrence. None of the 26 patients who achieved a CR experienced ocular-adnexal recurrence. All three patients who failed locally were in the PR group. In total, five of the six patients who developed any recurrence exhibited a PR after the initial radiotherapy. Age, gender, tumor location, and radiotherapy dose had no influence on the development of recurrence at any site.

Table 2. Clinical Data on Six Patients Who Experienced Treatment Failure after Radiotherapy

Patient	Age (yrs)	Gender	Initial location	Size (mm)	Dose (Gy)	Energy (MeV)	Lens shield	Response to RT	Failure site (mos)	Salvage therapy	Status
Gy: grays; MeV: megaelectron volts; RT: radiotherapy; F: female; M: male; blt: bilateral; E: electron; X: X-ray; Co: cobalt-60 γ-ray; PR: partial response; CR: complete response; CT: chemotherapy; AWD: alive with disease; NED: no evidence of disease; DOD: died of disease. a In-field recurrence. b ^90Sr β-ray was used.
1	46	F	Conjunctiva (blt)	15	30	6E	Yes	PR	Conjunctiva (4)a	None	AWD
2	44	M	Conjunctiva	20	44	12E	Yes	PR	Conjunctiva (97)a	RTb	NED
3	69	F	Conjunctiva (blt)	30	20	15E	Yes	PR	Blt eyelid (4)a	RT	NED
4	49	M	Conjunctiva	15	40	10X	Yes	PR	Oral floor (81)	RT	NED
5	47	F	Lacrimal gland (blt)	20	30	4X	No	CR	Submandibular gland (11)	RT	NED
6	75	M	Soft tissue	30	40	Co	No	PR	Neck lymph node (47)	CT	DOD

Table 3. Rates of Recurrence According to Patient and Treatment Factors

Characteristic	No. of patients (%)
	Ocular-adnexal recurrence	P valuea	Recurrence at any site	P valueb
RT: radiotherapy; CR: complete response; Gy: grays. a Fisher exact test. b Chi-square test
Laterality
Unilateral	1/43 (2)	0.048	3/43 (7)	0.029
Bilateral	2/7 (29)	—	3/7 (43)	—
Initial response to RT
CR	0/26 (0)	0.10	1/26 (4)	0.093
Non-CR	3/24 (13)	—	5/24 (21)	—
Age (yrs)
< 60	2/22 (9)	0.58	2/28 (7)	0.39
≥ 60	1/28 (4)	—	4/22 (18)	—
Gender
Male	1/33 (3)	0.26	3/33 (9)	0.40
Female	2/17 (12)	—	3/17 (18)	—
Location
Conjunctiva	3/29 (10)	0.25	4/29 (14)	1.00
Others	0/21 (0)	—	2/21 (10)	—
Dose of RT (Gy)
≤ 30	2/18 (11)	0.29	3/32 (9)	0.65
> 30	1/32 (3)	—	3/18 (17)	—

Survival

At the time of the data analysis, 44 patients were alive with no evidence of disease, and 1 patient was alive with disease. Only one death in this cohort was due to lymphoma. The other 4 patients died of intercurrent illnesses, including 1 death due to apoplectic stroke at 96 months. The 5-year overall survival rate for the entire group was 91% (Fig. 2) with a median follow-up of 46 months (range, 10–140 months).

Figure 2. Overall survival of the 50 patients with extranodal marginal zone B-cell lymphoma of mucosa-associated lymphoid tissue in the ocular adnexa.

Acute Toxicity and Late Complications

Radiotherapy was tolerated well. Thirty-nine patients developed minimal irritation that required no specific treatment, and 8 patients developed a mild conjunctivitis that resolved spontaneously. Late effects of radiotherapy were acceptable. Cataracts developed in 6 patients, with 2 patients undergoing cataract extraction 29 months and 36 months after radiotherapy. All six patients were treated with megavoltage photon beams, and the lens could not be spared because of the tumor location. The total dose for these 6 patients was 30 Gy in 1 patient and 40 Gy in 5 patients. Cataracts did not develop in any of the 11 patients who were treated with lens shielding. Retinal complications attributed to radiotherapy developed in two patients. One patient was diagnosed with radiation retinopathy by an ophthalmologist 72 months after receiving 40 Gy at 2 Gy per fraction. Another patient experienced mild retinal bleeding 120 months after 40 Gy of radiotherapy, and this was accompanied by decreased visual acuity. One patient who underwent an initial excisional biopsy for tumor in the orbital soft tissue developed a corneal ulcer after receiving 40 Gy. There were no severe, late, lacrimal complications attributable to radiotherapy. The mean total dose of radiotherapy in the 9 patients who developed cataracts, retinal complications, or corneal ulcer was slightly greater than the mean dose in the remaining 41 patients, although the difference was not statistically significant (38.9 Gy vs. 36.3 Gy; P = 0.29; t test).

DISCUSSION

As used in the current study, the term lymphoma in the ocular adnexa refers to lymphoma that arises in the extraocular orbital space and involves the conjunctiva, eyelids, lacrimal gland, or orbital soft tissue. Currently, most low-grade lymphoid tumors in the ocular adnexa diagnosed according to the traditional histopathologic classification can be considered MALT lymphoma.14 In contrast to the published literature, the current series exhibited a preponderance of males. However, the median age of patients in this study was similar to that reported previously.15

The therapeutic approach for patients with extranodal MALT lymphoma should be tailored according to the specific site. Local radiotherapy is the treatment of choice for patients with disease originating in the ocular adnexa.9, 10, 16, 17 However, there is no consensus regarding the optimal radiotherapy for patients with MALT lymphoma in the ocular adnexa. In particular, the optimal dose and target volume have yet to be determined. Although some authors recommend lower doses of 25–30 Gy for patients with low-grade orbital lymphomas, most radiotherapy series include lymphomas of various histologic subtypes,9, 18 and few reports have focused specifically on radiotherapy for patients with MALT lymphoma. Thus, there is insufficient current evidence to suggest the absence of a dose-response beyond 30 Gy for patients with Stage I MALT lymphoma in the ocular adnexa. Because of the relatively uncommon presentation of this disease, reports of radiotherapy tend to be small, retrospective series from a single institution with limited statistical analyses.

To our knowledge, the current study is the largest reported series to specifically address the treatment of patients with ocular-adnexal MALT lymphoma. Despite the fact that the initial tumor response marginally influenced local control, there was no significant difference in the mean dose of radiotherapy between patients with and without CR. A total dose of > 30 Gy had no impact on local control in the current study. Although retrospective studies are subject to several limitations, our results suggest that 30 Gy may be sufficient for the treatment of patients with MALT lymphoma in the ocular adnexa. Our data indicate that this tumor tended to show gradual regression after the evaluation of initial response, and the dose of radiotherapy may have been determined empirically at a level higher than required. Based on the results of the current study, we are planning a prospective study using a total dose of 30 Gy for patients with MALT lymphoma in the ocular adnexa. With regard to the adverse effects of radiotherapy, there was no significant difference in the total dose between patients with and without late complications. However, 8 of the 9 patients who developed cataracts, retinal complications, or corneal ulcer had received 40 Gy. It appears that radiotherapy doses ≥ 40 Gy have the potential for deleterious late effects on the ocular adnexa. It was decided that patients who already had received higher doses should be followed carefully and monitored for the late effects of radiotherapy.

Most radiotherapy series report an absence of in-field recurrence and 100% local control in patients with ocular-adnexal lymphoma.10, 18, 19, 20 In contrast to these reports, three patients in the current study developed in-field recurrences in the ocular adnexa. All three patients had an initial conjunctival lesion and underwent radiotherapy using a lens shield. Two of the 3 patients developed an early presentation of recurrence at 4 months, with 1 recurrent lesion located adjacent to the shielded area. It appears that the lens shield may have blocked a small part of the conjunctiva where the induced MALT developed neoplastic proliferation, forming an in-field recurrence. Another patient initially received only 20 Gy. Although 15 MeV electrons were used in that patient, the eyelid that situates in front of the conjunctiva may have received a somewhat lower dose from a single anterior field. We speculate that early local recurrence in these two patients may have been due to the radiotherapy technique used. These patients could have been treated successfully if a sufficient dose had been delivered to the entire conjunctiva and eyelid. Thus, we consider that only one patient in this study developed a true in-field recurrence. However, care must be taken in interpreting our results. Considering the fact that this was a multiinstitutional, retrospective study and that the patients were treated over a 10-year period, there were limitations in assessing the radiotherapy technique, especially in patients who were treated with electron-beam radiation using direct set-up.

There is no consensus regarding the optimal planning target volume of radiotherapy for the treatment of patients with ocular-adnexal lymphoma. The determination of target volume with regard to the necessity to treat the entire orbital component electively also is evolving21 but is not widely accepted. For patients with conjunctival lymphoma, Pelloski et al.22 used a planning target volume that included the entire conjunctiva. The conjunctiva is the only mucosal structure of the ocular adnexa that has intrinsic or acquired lymphoid tissue. Our results suggest that the low-dose area in the conjunctiva has the potential to develop early recurrence in patients with an initial conjunctival lesion. Based on the patterns of ocular-adnexal recurrence in the current study, we propose that the entire conjunctiva and eyelid should be included in the planning target volume of radiotherapy for MALT lymphoma originating and limited to the conjunctiva. The planning target volume for tumor in the orbital soft tissue should be balanced against the probability of late toxicity.

In the current study, we examined the effects of age, gender, tumor location, laterality, dose of radiotherapy, and initial response to radiotherapy on local control. Patients with MALT lymphoma generally exhibit a good performance status, as shown by Thieblemont et al.8 Only 2 patients in the current study had a performance status of 2 or worse. No patients had elevated serum lactate dehydrogenase levels or systemic symptoms. Thus, these well-known prognostic factors for patients with aggressive lymphoma were excluded from our statistical analyses. Patients with initial bilateral tumors exhibited a higher recurrence rate in the current study. However, as discussed above, two patients had early ocular-adnexal recurrences that may be attributable to the radiotherapy technique used. Thus, it is questionable that bilateral involvement itself is associated with tumor recurrence. Controversy remains concerning the prognostic importance of bilateral involvement.10, 21, 23 Whether bilateral involvement implies a worse prognosis should be evaluated in future prospective studies in which a quality-controlled, uniform radiotherapy technique is mandatory.

The initial response to radiotherapy marginally influenced subsequent tumor recurrence in this study. Because most of the tumors in patients who did not achieve a CR demonstrated continuous regression after the response evaluation, it appears preferable to refer to this type of tumor as a slow responder. It is not clear whether the gradual tumor response in these patients implies an intrinsic radioresistance generally observed in slow-growing tumors. Because there was only one death due to lymphoma, no survival disadvantage in the slow responders could be demonstrated. However, 3 recurrences in slow responders occurred after a long latent period after radiotherapy (47 months, 81 months, and 97 months). Thus, from an alternative viewpoint, we speculate that there are some populations among the slow responders with the potential to develop metachronous lesions. Therefore, we believe that longer follow-up is required for patients with ocular-adnexal MALT lymphoma who exhibit a slow response to radiotherapy.