It is estimated that lymphoma of the ocular adnexa represents approximately 8% of extranodal non-Hodgkin's lymphomas. Although much has been written about the subject, the data are difficult to interpret because many of the published series grouped lymphomas of various histologic subtypes together in their analyses. The inclusion of benign lesions of lymphoid hyperplasia and the use of the now out-of-vogue term “pseudolymphoma” in certain studies further clouded the understanding of the malignant counterpart.

Although any of the many histologic subtypes of non-Hodgkin's lymphoma may be encountered in the ocular adnexa, the extranodal marginal zone lymphoma, mucosa-associated lymphoid tissue (MALT) type has been found to be an important subtype. Isaacson and Wright   first introduced the concept of MALT in the 1980s when they made the seminal observation that certain extranodal lymphomas were related to mucosa-associated lymphoid tissue rather than to nodal lymphoid tissue. Since then, MALT lymphoma has gained general acceptance as a distinct clinicopathologic entity and is included in the Revised European–American Lymphoma classification under the broader term of marginal zone B-cell lymphoma and in the World Health Organization (WHO) classification as marginal zone B-cell lymphoma, MALT type.

This paper presents the Massachusetts General Hospital experience on a large cohort of consecutive patients with lymphoma of the ocular adnexa classified under the WHO classification scheme. Local control issues, as well as long-term clinical outcome, are discussed, with emphasis on the two most common histologic subtypes, marginal zone lymphoma/MALT type and follicular lymphoma.

To evaluate the clinical behavior and treatment outcome of ocular adnexal lymphomas classified by the World Health Organization system, with emphasis on marginal zone lymphoma of mucosa-associated lymphoid tissue (MALT).

The clinicopathologic materials from 98 consecutive patients treated for ocular adnexal lymphoma were reviewed. Fourteen patients had prior lymphoma and 84 patients had primary disease (75% Stage I, 6% Stage III, and 19% Stage IV). Radiation (photons/electrons) was administered to 102 eyes to a median dose of 30.6 Gy. The mean follow-up was 82 months.

Results

The most common subtypes among the primary patients were MALT (57%) and follicular (18%) lymphoma. The 5-year actuarial local control rate in 102 irradiated eyes was 98%. Among the low-grade lymphomas, the 5-year local control rate correlated with the radiation dose in the MALT lymphoma subgroup (n = 53): 81% for <30 Gy and 100% for 30 Gy (p <0.01). For the non–MALT low-grade lymphomas such as follicular lymphoma (n = 30), the local control rate was 100% regardless of dose. For 39 Stage I MALT lymphoma patients treated with radiation alone, the distant relapse-free survival rate was 75% at 5 years and 45% at 10 years. Distant relapses were generally isolated and successfully salvaged by local therapy. The overall survival for this subgroup was 81% at 10 years, with no deaths from lymphoma.

Eye complications

Orbital RT was generally well tolerated, with mild acute effects such as transient keratoconjunctivitis, dryness or excessive tearing, loss of eyelashes/eyebrow, and periorbital erythema and edema.

Most patients had longitudinal follow-up with an ophthalmologist in addition to the oncology team. Cataract development was documented in 16 of the 102 irradiated eyes. The median time to cataract diagnosis was 40 months after RT completion. Among the 57 eyes irradiated with photons (without lens shielding), 14 cataracts were documented: 1 in 8 eyes treated to <30 Gy, 1 in 12 eyes treated to 30.6–32.4 Gy, and 12 in 32 eyes treated to >32.4 Gy. In 45 eyes irradiated with electrons, 2 cataracts were documented to have developed despite the use of a lens shield.

Mild retinopathy was documented in 4 patients after doses of 39.6–46 Gy. One patient lost his sight immediately after orbital biopsy from orbital hemorrhage related to postoperative hypertension. Severe keratopathy and recurrent corneal abrasions developed in 1 patient with an aggressive diffuse large cell lymphoma who had received 40 Gy in 23 fractions with concomitant cyclophosphamide, vincristine, and prednisone chemotherapy, followed by three more cycles of CHOP. The RT was given primarily in 1.8–2-Gy daily fractions, although a few treatments were at 1.5 Gy twice daily. This patient subsequently underwent keratoplasty for a corneal perforation at 23 months and had restoration of good visual function.

Discussion 

This is the first large-scale outcome analysis of ocular adnexal lymphomas classified by the WHO classification scheme presenting information regarding the relative frequency of the various histologic subtypes, pattern of disease at presentation, response to RT, and 10-year treatment outcome.

In this cohort of consecutive patients with ocular adnexal lymphoma treated in a single institution, MALT lymphoma was by far the most common histologic subtype in patients with primary disease, accounting for slightly more than one-half of cases. Follicular lymphoma was the second most common subtype, occurring in 20% of cases. In comparison, the histologic distribution was quite different among patients who developed secondary orbital involvement after prior lymphoma elsewhere. In the secondary group, no cases of MALT lymphoma occurred at all. Follicular lymphoma and mantle cell lymphoma each accounted for about one-third of the cases.

Although most of the primary patients had Stage I disease, 25% had Stage III-IV, underscoring the importance of thorough staging studies such as body CT scans and bone marrow biopsy in this population. One-quarter of the Stage IV patients were upstaged from Stage I solely because of a positive bone marrow biopsy. In addition, CT and/or MRI of the orbit are of utmost importance. One-half of the patients who had evidence of disease only on the conjunctiva by physical examination were found to have synchronous involvement of retrobulbar soft tissues on CT or MRI. Lymphoma isolated to the conjunctiva can be treated well with electrons, which have relatively superficial depth–dose characteristics; however, involvement of the retrobulbar tissues generally calls for photons, which have a deeper beam penetration. Therefore, the knowledge of the full extent of orbital involvement is critical for proper radiation planning. Of further interest is that regional nodal involvement is uncommon in primary orbital lymphoma. Not one of the 63 early-stage patients in this study had regional nodal disease, corroborating the observations of Pelloski, who noted only 5% cervical nodal involvement in 21 patients. A slightly greater rate of 13% Stage II disease has been reported by others

Local control with RT

Although excellent local control rates have been reported for low-grade orbital lymphoma treated with RT alone, the optimal radiation dose is controversial. Bolek  observed no in-field failure in 9 patients with low-grade histologic features treated with 20–25 Gy of radiation. Bessell reported on low-grade lymphoma of the orbit and conjunctiva treated to 20–30 Gy, with 100% local control. In the study of Minehan , 95% of 18 patients with low-grade lymphoma had successful disease control after RT ranging from 16.9 to 50.6 Gy. The one failure occurred after 24 Gy. More recently, Martinet  reported a 97% local control rate with a median dose of 34.2 Gy in 90 patients with mostly low-grade disease, although intermediate- and high-grade lesions were also included in the analysis. On the basis of these and other studies, many authors have concluded that low-grade orbital lymphoma can be successfully controlled by radiation doses in the range of 25–34 Gy, although some have advocated doses as low as 15–20 Gy

Many of the published series were limited by small patient numbers and the grouping of low- and intermediate-grade histologic subtypes in the same analyses. Furthermore, the use of the Working Formulation classification in most studies made it difficult to extrapolate the results to current cases classified by the WHO lymphoma schema. For example, it is not immediately apparent whether extranodal marginal zone lymphoma of the ocular adnexa would respond to RT in the same manner as what is called “low-grade lymphoma” in the literature.

Published data are limited on the optimal radiation dose for MALT lymphoma of the ocular adnexa. Le  reported a 100% local control rate in 31 cases of MALT lymphoma of the orbit treated to 30–40 Gy. Hitchcock  observed equally successful local control in 11 patients with MALT lymphoma of the orbit treated to a median dose of 32 Gy. Because these dose ranges were fairly narrow, these studies did not address whether a dose of <30 Gy might be sufficient for MALT lymphoma of the orbit.

In the hope of answering these questions, we analyzed the radiation dose–response relation in 83 consecutive cases of low-grade orbital lymphoma, all treated with RT alone. Within the range of 16.2–45 Gy, a statistically significant dose–response relationship emerged. The actuarial 5-year local control rate was 100% for 30 Gy and 86% for <30 Gy. All the local failures were in patients with MALT lymphoma. In the subgroup analysis, the local control rate for follicular and B-CLL lymphoma was 100%, regardless of dose, within a range of 16.2–40.6 Gy. In contrast, for MALT lymphoma, <30 Gy was associated with only an 86% 5-year actuarial local control rate vs. 100% for 30 Gy.

The single local failure in the 30.6–32.4-Gy group (at 8 years) was in a patient treated with low-energy electrons for disease presumed to be limited to the conjunctiva. Because the patient did not undergo CT or MRI of the orbit for staging and subsequently developed recurrence in the retrobulbar tissue, one can conjecture that the recurrence was possibly a result of inappropriate therapy rather than a true radiation failure.

These results suggest that orbital MALT lymphoma may have a different dose–response characteristic than that of follicular and B-CLL lymphomas and that MALT lymphoma may require a slightly greater radiation dose than lymphomas of the latter types. Although doses in the mid 20-Gy range may be adequate for follicular lymphoma, our observations support the use of 30.6–32.4 Gy in 1.8-Gy fractions for MALT lymphoma. These dose ranges were well tolerated, with no serious long-term complications observed, other than a low risk of cataract formation.

Systemic disease control

It is well known that MALT lymphoma is an indolent disease, and MALT lymphoma of the ocular adnexa is no exception. However, our long-term results indicate that Stage I patients treated with RT face a continuous risk of distant relapse during the first 10 years. Recurrences are frequently isolated, involving other MALT sites, as well as nodal sites, and can generally be well controlled by local therapy such as surgery or RT. Patients therefore enjoy good lengths of disease-free intervals. Although the projected rate of remaining free of distant relapse was only 45% at 10 years, no patient died of lymphoma, and the 10-year actuarial disease-specific survival rate was 100%.

The high distant relapse rate for Stage I ocular adnexal MALT lymphoma stands in strong contrast to the results reported for localized MALT lymphoma of the stomach. MALT lymphoma occurs most frequently in the stomach, and treated gastric MALT lymphoma is associated with a very low risk of distant recurrence. In the series of localized MALT lymphoma by Tsang , no relapses were observed in 15 stomach cases. Schechter  also reported no relapses in 17 cases of localized gastric MALT lymphoma treated with RT. Our own experience indicated a 10-year relapse-free survival rate of 93% for 16 patients with Stage I gastric MALT lymphoma

Longer follow-up and larger patient numbers are needed to determine whether MALT lymphoma of the ocular adnexa truly has a greater tendency for distant relapse compared with MALT lymphoma of the stomach. If this difference in the relapse rate is borne out by future studies, the reasons remain to be explored. Several other differences between orbital MALT lymphoma and gastric MALT lymphoma are noteworthy. Although the bacteria Helicobacter pylori has been shown to play an etiologic role in gastritis and the development of gastric MALT lymphoma, no such outside antigenic stimulus has been associated with MALT lymphoma of the orbit. Ultimately, the basis for variations in clinical behavior may lie at the molecular level. The most frequent structural chromosomal aberration in MALT lymphoma is the t(11;18)(q21;q21) translocation. This translocation pattern has been observed in 12–48% of MALT lymphoma of the stomach, yet rarely in the ocular adnexa (27–32). In contrast, a different translocation pattern, t(14;18)(q32;q21), was detected in 3 of 8 lymphomas of the ocular adnexa and 0 of 10 lymphomas of the stomach (33). The involved gene on chromosome 18 has been termed the MALT1 gene and is thought to be specific for extranodal MALT lymphoma (34–36). Differences in chromosomal translocation patterns have also been observed for MALT lymphoma of the lung, thyroid, skin, and salivary gland (33). The variation in translocation patterns among the different sites, together with clinical observations, suggest that the clinical heterogeneity of MALT lymphomas of different sites of origin may depend on events at the molecular level.

The small number of Stage I follicular lymphomas of the orbit reported here precluded any definitive conclusion. Our observed 55% rate of freedom-from-distant relapse at 10 years is in keeping with the long-term results for localized follicular lymphoma in general, treated with RT alone. Likewise, the number of patients with primary Stage I diffuse large B-cell lymphoma was small. The recommendation is to treat localized diffuse large cell lymphoma of the ocular adnexa with three cycles of CHOP plus involved-field RT, as supported by the findings from the Southwest Oncology Group randomized trial. With regard to the radiation dose, Wilder et al.(39) reported that for <3.5-cm lymphomas that reached a complete response after three cycles of CHOP-based chemotherapy, 29.1–39.1 Gy in 1.8-Gy fractions yielded a 96% control rate, with no advantage to using higher doses.

Our study was limited by all the inherent weaknesses of any retrospective study. Because pertinent staging studies were performed in most, but not all, the primary patients, the distant relapse rates might have been slightly overestimated.

Conclusion

Dose–response data suggest that the optimal radiation dose for MALT lymphoma of the ocular adnexa is 30.6–32.4 Gy in 1.8-Gy fractions and follicular lymphoma is adequately controlled with doses in the mid-20 Gy range. The substantial risk of distant relapse in Stage I ocular adnexal MALT lymphoma underscores the importance of long-term follow-up for this disease and the need for additional comparative studies of MALT lymphoma of different anatomic sites.