Lymphomas of the nasal cavity and paranasal sinuses are uncommon. 1–4 They appear to be heterogeneous with respect to pathologic and clinical behavior. 2,4–12 Their rarity and the nonuniformity of treatment make understanding the natural history and treatment outcome difficult. Some publicized studies suggest that radiation alone for early-stage disease provides good local–regional control, and that chemotherapy did not improve on the relapse rate and only should be used for more extensive disease. 2,13 Other more recent studies have shown that combined-modality treatment (CMT) with chemotherapy and local–regional radiation provides a significant improvement in both disease-free survival (DFS) and overall survival (OS) specifically for lymphomas of the nasal cavity and paranasal sinuses. 3,4 We have reviewed the experience at the Massachusetts General Hospital to further understand the outcomes in patients with this rare disease.
PATIENTS AND METHODS
A retrospective review was made of patients diagnosed and treated for paranasal sinus or nasal cavity lymphoma at the Massachusetts General Hospital, between the years 1960 and 1998. Patients were considered to have primary sinonasal lymphoma if they presented with symptoms related to the sinonasal area, if diagnostic tissue was obtained from the nasal cavity or sinuses, if there was predominate or exclusive involvement of the nasal cavity or paranasal sinuses, and if there was no prior history of lymphoma. A review of the records of the Massachusetts General Hospital, Massachusetts Eye and Ear Infirmary, and consultations submitted to the authors revealed 86 patients with this diagnosis. In 58 of these cases, slides and/or tissue blocks were available for review. Thirty-two of these 58 patients did not receive complete clinical evaluation and/or treatment at the Massachusetts General Hospital and were excluded from treatment-outcome analysis. The remaining 26 patients received treatment with either radiation alone or radiation combined with either surgery or chemotherapy. Of these 26 patients, there was 1 case each of extranodal marginal zone lymphoma, adult T-cell lymphoma/leukemia, and peripheral T-cell lymphoma, so these cases were also excluded from the final analysis. The remaining 23 patients comprise the cohort reported here. Surgery usually consisted of a diagnostic biopsy, with a few patients early in the series undergoing a partial resection.
Patients were clinically staged at the time of diagnosis and treatment according to the Ann Arbor system. 14 TNM staging according to the 1998 TNM classification of the American Joint Committee on Cancer was used for all tumors thought to arise from a sinus. Because no standardized T stage for nasal cavity exists, for purposes of this study to be able to assess extent of local disease on outcome, the T stage was assigned in a similar fashion as that for paranasal sinuses with, for simplicity, limiting it to only two stages. These two stages were for either those tumors confined to the nasal cavity (T1) or those with clear extension in surrounding structures (T4). The T stage assigned was based on the description given from examination and/or the radiographic findings, depending on the era of treatment. Clinical staging procedures included history, physical examination including ear/nose/throat evaluation, complete blood counts, chemistry panel, radiographic assessment including chest radiograph, bone marrow biopsy, computerized axial tomography and/or magnetic resonance imaging of head and neck, as the technology became available. All patients had either a lymphangiogram and/or computed tomography (post 1980) of the chest, abdomen, and pelvis performed as part of the staging workup. Age, gender, presence of absence of B symptoms, presenting symptoms, and primary site were recorded for each case.
Cases were classified pathologically according to the Revised European-American Lymphoma Classification and the criteria of the International Lymphoma Study Group for Nasal NK/T-cell lymphoma. 15,16 All cases were evaluated by light microscopic examination of hematoxylin and eosin–stained slides from one or more diagnostic biopsies and by immunohistochemistry. In situ hybridization for Epstein-Barr virus–encoded RNA was performed whenever possible. Southern blot hybridization to detect rearrangement of the T-cell receptor and/or immunoglobulin genes was performed in selected cases. The pathologic and immunophenotype findings of all 58 cases have been previously described in detail. 17
Two clinical outcomes were recorded: OS from the date of treatment of the disease, and DFS from the same date. For DFS, if a patient died of a cause other than disease or treatment, then the patient’s data were censored at the date of death. Kaplan-Meier survival curves, 18 along with the log-rank test, 19 were used to assess the effects of the following covariates: stage (IEA, IIEA), T stage (1, 2, 3, 4), histologic type of lymphoma (diffuse large B-cell lymphoma, T/NK), location of lymphoma (sinus, nasal, sinonasal), gender, age (≤60 years, >60 years), treatment (radiation alone, radiation + chemotherapy), and dose of radiation (≤45 Gy, >45 Gy). Proportional hazards modeling 20 was used to examine the effects of sets of covariates on the two outcomes. Statistical procedures were run with the SAS statistical software package (SAS Institute, Cary, NC, U.S.A.) on Wintel-based microcomputers. Final survival plots were constructed with the SigmaPlot graphics package (SPSS, Chicago, IL, U.S.A.).
Patient clinical characteristics are listed in Table 1. The majority of patients had nasal obstruction and localized soft-tissue swelling. Many patients had more than one symptom at presentation. The majority of patients treated had Ann Arbor Stage I disease with approximately one third having T1 primary lesions. Approximately one half of the patients had locally advanced disease with T3 or T4 lesions. For primary tumors evaluable and thought to arise from the nasal cavity, these tumors were classified as: T1 (n = 1): tumor confined to the nasal cavity; T4 (n = 4): tumor with extension outside the nasal cavity in surrounding structures. The male to female ratio was 1.4:1. The majority of tumors confined to a paranasal sinus involved the maxillary sinus. Three of the 18 paranasal sinus tumors were T1 stage.
Eight of the 23 cases were classified as nasal NK/T-cell lymphoma. Five of the eight involved the nasal cavity either alone or with extension to surrounding structures including the paranasal sinuses. One case involved the paranasal sinuses alone; the other two cases involved both the paranasal sinus and extension to surrounding structures. Fifteen cases were classified as diffuse large B-cell lymphoma. Of these, 3 involved the nasal cavity alone or nasal cavity with extension to surrounding structures, and the other 12 involved the paranasal sinuses alone or paranasal sinuses plus extension to surrounding structures.
All 23 patients received radiation as part of their initial curative treatment. Treatment was given with high-energy radiation either from a Co-60 machine or low megavoltage linear accelerators. Field arrangement was either a two-field anteroposterior/lateral wedge pair or a three-field technique (i.e., opposed lateral fields and an anteroposterior field approach). Total doses ranged from 34 Gy to 60 Gy, using standard once-daily 1.8 Gy to 2 Gy fractionation. Patients receiving treatment early in the series tended toward the higher end of the dose range, whereas patients treated in the more modern era had doses on the lower end of the range. No patients received central nervous system (CNS) prophylaxis with radiation. Three patients received combined treatment with systemic chemotherapy and radiation. Chemotherapy consisted of cyclophosphamide, doxorubicin, vincristine, and prednisone (n = 3). Sequencing of treatment was individualized and varied among patients.
The major univariate results are summarized in Table 2. The overall estimated survival rate for the entire group was 78% at either 5 or 10 years, and the estimated DFS rate at 5 or 10 years was the same. In fact, by coincidence, the predicted OS and DFS for both 5 years and 10 years were the same, no matter which group of patients was examined. Figures 1 and 2 include the survival plots for groups divided by stage and T stage (DFS plots are identical). The most important covariate was stage, followed by T stage for both survival and DFS. Earlier stage (IEA versus IIEA) and smaller T stage (I or II versus III or IV) were strong predictors of increased survival and DFS.
The difference in DFS and OS for stage IEA versus stage IIEA disease was significant at p = 0.0001. T1 or T2 versus T3 or T4 demonstrated significance for DFS and OS at p = 0.0243. This remained highly suggestive when stratified by stage (p = 0.0896). No other covariates yielded p values less than 0.05 for the log-rank test. The multivariate proportional hazards analysis confirmed the univariate analyses: with all the potential covariates entered in the model, only stage and T stage were significant (p < 0.05) covariates.
With only stage in the model, the addition of T stage significantly improved the model, as shown by the likelihood ratio test (p < 0.05). Therefore, the combination of stage and T stage is the best predictor of both survival and DFS. Figure 3 shows Kaplan-Meier plots for survival and DFS for four groups of patients: IEA & (T1 or T2), IEA & (T3 or T4), IIEA & (T1 or T2), and IIEA & (T3 or T4). The difference in survival and DFS between the IEA & (T1 or T2) and IIEA & (T3 or T4) groups was very large: the 10-year predicted OS and DFS rates for the former group were 100%, and for the latter group were 0% for both outcomes. The difference among the four groups for both outcomes was significant (p = 0.0001). The estimated median time to death for IEA & (T3 or T4) was 166 months; that for IIEA & (T3 or T4) was 24 months; and that for IEA & (T1 or T2) and IIEA & (T1 or T2) could not be estimated because there were no deaths. Thus, overall, the (T1 or T2) patients have an excellent prognosis, irrespective of the Ann Arbor stage, and the (T3 or T4) patients have a much poorer prognosis as the Ann Arbor Stage increases from IEA to IIEA. The median follow-up time for the group of 23 patients (time to death or last contact) was 84 months, with a range from 3 to 288 months.
Sites of relapse include both local and distant for patients treated with either radiation alone or radiation and chemotherapy. Sites of distant failure included distant nodal sites, bone, skin, testes, gastrointestinal, and the CNS. Fifteen of the 20 patients treated with radiation alone had no evidence of disease at last follow-up. One patient treated with radiation alone failed to respond to treatment locally. However, 4 patients treated with radiation alone failed to respond at distant sites. Of the three patients treated with radiation and chemotherapy, none had either a local or distant failure to respond.
A fistula involving the bridge of the nose developed in three patients after completion of treatment. Cystitis thought to be related to chemotherapy treatments developed in one patient. A nasal myxoid fibrosarcoma developed in one patient 8 years after treatment for a primary nasal cavity lymphoma.
Treatment of patients with early-stage non-Hodgkin’s lymphoma of the nasal cavity and paranasal sinuses has evolved from local–regional irradiation to a combination approach using both chemotherapy and radiation. There is no question that patients with advanced-stage disease require systemic treatment and local treatment with radiation for bulky disease or disease not responsive to chemotherapy. However, there are some conflicting data on the best treatment for patients with early-stage disease. In an effort to answer the question of appropriate treatment for such patients, we have limited our study to those patients treated with early-stage disease, most of whose lymphomas were Ann Arbor Stage I and who received treatment with radiation alone.
Effective treatment of early-stage disease with radiation alone has been reported. Robbins et al. 5 indicated that for early-stage disease, T1 and T2, radiotherapy alone was effective treatment and that multiagent chemotherapy was needed for advanced stages T3 and T4, regardless of whether there was cervical lymphadenopathy. 21
A more recent report by Liang et al. 2 drew the same conclusion: that no benefit was derived with the addition of chemotherapy to radiation, albeit their patients had only nasal T-cell lymphomas. 22 Despite these reports, most sinonasal lymphomas are treated with a combined approach with systemic chemotherapy and locoregional radiation.
For non-Hodgkin’s lymphoma in general, a CMT approach for early-stage aggressive lymphomas has been shown to improve DFS and OS in two recent randomized studies reported. 23,24 Earlier reports have shown improvement in outcome with CMT for non-Hodgkin’s lymphoma of the head and neck region, although a minority of the patients studied had involvement of the nasal cavity and paranasal sinuses that were treated with radiation alone. 1,25 Moreover, recent reports have supported the use of chemotherapy and radiation specifically for lymphomas of the sinonasal region. 3,4
Logston et al. 3 have reported a significant improvement in freedom from progression (FFP) with CMT over radiation alone with patients with IEA and IIEA disease having an actuarial 5-year freedom from relapse (FFP) of 83% and OS of 67%. A recent report by Hausdorff et al. 4 on paranasal sinus lymphomas, which included only 13 patients with IEA or IIEA disease, found that even with CMT including CNS prophylaxis, improvement in the 5-year survival rate was only 29% with a median survival of only 18 months. This group also found a worse prognosis for the 6 patients with T- or T/NK lineage tumors compared with the 10 patients with B-lineage tumors.
In our study, patients with IEA disease did significantly better than patients with IIEA disease in both control of their local disease and survival with local–regional radiation alone. Furthermore, both IEA and IIEA patients with primary tumors of stage T1 or T2 do significantly better than T3 or T4 with respect to DFS and OS. Our results did not show any difference between patients treated with CMT versus radiotherapy alone, but no definite conclusions can be made with so few patients treated with CMT. In addition, despite the good local control seen in patients treated with radiation alone, the fact that 4 of 20 (20%) had distant failure suggests the need for systemic treatment. Furthermore, two patients in whom distant failure developed had CNS involvement. No prophylactic CNS radiation was done in this series of patients, but this has been advocated as a way to prevent relapse within the CNS. 4,26 Others, however, do not believe that prophylaxis is necessary based on the patterns of recurrence seen. 2,3,27 Our results confirm earlier reports that achieved good results with radiation alone. Additionally, our results support the conclusion by Logsdon et al. 3 for patients treated with radiation, that is, that having a greater extent of local disease (i.e., T4) is a clinical characteristic that is a significant prognostic indicator for a poor outcome. They reported an actuarial 5-year freedom from progression of 87% and 70% for patients with stage IEA and IIEA, respectively, treated with CMT, and only 34% for patients with stage IIIEA and IVEA disease. For those patients treated with radiotherapy alone, the 5-year freedom from progression was significantly worse, with only 38% for stage IEA & T4 patients and 0% for stage IIEA, IIIEA, and IVEA patients.
An earlier report by Robbins et al. 21 also demonstrated a significant difference in outcome by extent of disease for patients treated with radiation alone, with a 5-year DFS for T1 or T2 versus T3 or T4 of 78% and 19%, respectively. They also noted no difference with the addition of systemic chemotherapy in outcome for the T1 or T2 group but did note improvement for the T3 or T4 group. 5 This phenomenon of local extent of disease influencing outcome has been noted by others and has been used as an argument for implementing a TNM classification of lymphomas involving these regions to better predict prognosis and make therapeutic decisions. 28,29
Complete histopathologic details of all patients reported here as being treated at the Massachusetts General Hospital, have been previously reported along with other patients who were diagnosed with lymphoma of either the nasal cavity or paranasal sinuses. 15 Only 1 patient of 8 with an NK/T cell lymphoma had paranasal sinus alone involvement, whereas only 3 of 15 patients with diffuse large B-cell lymphoma had the primary site as the nasal cavity.
These results suggest that lymphomas of the nasal cavity may have a different cellular origin with a predilection for NK/T cell origin. However, as others have reported, our results show the largest percentage of sinonasal tumors as a group to be of B-cell origin. 27
In addition to a possible difference in cellular origin, others have reported a possible difference in clinical behavior with B-lineage lymphoma having a more favorable outcome. 4,30–32 However, in both univariate and multivariate analysis, histologic differences were not found to be prognostic for either DFS or OS. This result differs from others who have suggested a different clinical behavior for T- or T/NK lineage. Indeed, the Stanford group report that all six patients with a T- or T/NK lineage relapsed or died within 6 months. 4
Several other variables were analyzed for prognostic significance in our study. No apparent differences in DFS and OS were noted by location of disease, whether involvement was sinus or nasal cavity. This may reflect that a more significant factor for outcome is extent of disease and not location by itself, because effective radiation treatment can be technically delivered to either of the sites. Radiation doses were also evaluated and did not demonstrate any difference between doses more than or less than 45 Gy as long as the dose is more than that which is expected to provide control of localized non-Hodgkin’s lymphoma (i.e., >36 Gy). Although others have suggested a potential difference in both DFS and OS between males and females and between older (>60 years) and younger (≤60 years) patients, neither of these factors was found to be significant in our series. The complication rate in our series was low, and complications were predominately localized to the initial site of radiation treatment. With more modern radiation treatment systems using computed tomography/magnetic resonance imaging technology and three-dimensional treatment systems, few side effects should be expected, with the likelihood that less radiation will be delivered to nontarget tissues.
The combination of lower radiation doses in the more modern era and improved therapies to better prevent and treat side effects from both radiation and chemotherapy should contribute to a reduction in the incidence of both early and late side effects. This reduction in the more modern era of treatment has already been noted by others. 3
Sinonasal lymphomas appear to be well controlled locally with radiation alone. However, they are prone to fail distantly despite good local control and therefore require systemic treatment to reduce distant failure rates. Patients treated with CMT in our series experienced no survival benefit. Also histology did not appear to be a significant prognostic factor for outcome. Patients with both earlier T stage and Ann Arbor stage do significantly better than patients with more advanced stages in terms of both DFS and OS. Furthermore, the combination of early-stage disease as classified by the Ann Arbor system and the T stage adopted here for nasal cavity and that of the American Joint Committee on Cancer for sinus was the best prognostic factor for OS and DFS. With the small number of patients treated in our series, particularly with CMT, these results would need to be studied in a larger prospective randomized trial before any definitive conclusions could be drawn.
We thank Mr. John Parsons, William Daley, and Jeannine Park for database management.
1. Jacobs C, Hoppe RT. Non-Hodgkin’s lymphomas of head and neck extranodal sites. Int J Radiat Oncol Biol Phys 1984; 11: 357–364.
2. Liang R, Todd D, Chan TK, et al. Nasal lymphoma. Cancer 1990; 66: 2205–2209.
3. Logsdon MD, Ha CS, Kavadi VS, et al. Cancer 1997; 80: 477–488.
4. Hausdorff J, Davis E, Long G, et al. Non-Hodgkin’s lymphoma of the paranasal sinuses: clinical and pathological features, and response to combined-modality therapy. Cancer J Sci Am 1997; 3: 303–311.
5. Robbins KT, Fuller LM, Vlasak M, et al. Primary lymphomas of the nasal cavity and paranasal sinuses. Cancer 1985; 56: 814–819.
6. Makepeace AR, Fermont DC, Bennett MH. Non-Hodgkin’s lymphoma of the nasopharynx, paranasal sinus and palate. Clin Radiol 1989; 40: 144–146.
7. Ferry JA, Sklar J, Zukerberg LR, et al. Nasal lymphoma, a clinicopathologic study with immunophenotypic and genotypic analysis. Am J Surg Pathol 1991; 15: 268–279.
8. Tran LM, Mark R, Fu YS, et al. Primary non-Hodgkin’s lymphoma of the paranasal sinuses and nasal cavity, a report of 18 cases with stage IE disease. Am J Clin Oncol 1992; 15: 222–225.
9. Campo, Cardesa A, Alos L, et al. Non-Hodgkin’s lymphomas of nasal cavity and paranasal sinuses, an immunohistochemical study. Am J Clin Pathol 1991; 96: 184–190.
10. Johnson CD, Kent DM, Varjabedian GC, et al. Malignant lymphoma of the maxillary sinus. J Am Osteopath Assoc 1993; 93: 255–258.
11. Hoppe RT, Burke JS, Kaplan HS, et al. Non-Hodgkin’s lymphoma: involvement of Waldeyer’s ring. Cancer 1978; 42: 1096–1104.
12. Wang CC. Malignant lymphoma of Waldeyer’s ring. Radiology 1969; 92: 1335–1339.
13. Ruijs CDM, Dekker AW, van Kempen-Harteveld ML, et al. Treatment of localized non-Hodgkin’s lymphomas of the head and neck. Cancer 1994; 75: 1281–1291.
14. Carbone PP, Kaplan HS, Mushoff K, et al. Report of the Committee on Hodgkin’s Disease Staging. Cancer Res 1971; 31: 1860–1861.
15. Harris NL, Jaffe ES, Stein H, et al. A revised European-American classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group. Blood 1994; 84: 1361–1392.
16. Jaffe E, Chan J, Su I, et al. Report of the workshop on nasal and related extranodal angiocentric T/NK cell lymphomas: definitions, differential diagnosis, and epidemiology. Am J Surg Pathol 1996; 20: 103–111.
17. Caudra-Garcia I, Proulx GM, Wu CL, et al. Sinonasal lymphoma: a clinicopathologic analysis of 58 cases from the Massachusetts General Hospital. Am J Surg Pathol 1999; 23: 1356–1369.
18. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457–481.
19. Peto R, Peto J. Asymptotically efficient rank invariant test procedures. J R Stat Soc 1972; 135A: 185–206.
20. Cox DR. Regression models and life tables. J R Stat Soc 1972; 34B: 187–220.
21. Robbins KT, Fuller LM, Vlasak M, et al. Primary lymphomas of the nasal cavity and paranasal sinuses. Cancer 1985; 56: 814–819.
22. Liang R, Todd D, Chan T, et al. Treatment outcome and prognostic factors for primary nasal lymphoma. J Clin Oncol 1995; 13: 666–670.
23. Miller TP, Dahlberg S, Cassady JR, et al. Chemotherapy alone compared with chemotherapy plus radiotherapy for localized intermediate- and high-grade non-Hodgkin’s lymphoma. N Engl J Med 1998; 339: 21–26.
24. Glick J, Kim K, Earle J, et al. An ECOG randomized phase III trial of CHOP vs CHOP + radiotherapy (XRT) for intermediate grade early stage non-Hodgkin’s lymphoma (NHL). In: Perry MC, ed. Proceedings of the Thirty-First Annual Meeting of the American Society of Clinical Oncology
, Los Angeles, California, May 20–23, 1995.
25. Plantenga KF, Hart G, Van Heerde Peter, et al. Non-Hodgkin’s malignant lymphomas of upper digestive and respiratory tracts. Int J Radiat Oncol Biol Phys 1981; 7: 1419–1427.
26. Burton G, Atwater S, Borowitz M, et al. Extranodal head and neck lymphoma: prognosis and patterns of recurrence. Arch Otolaryngol Head Neck Surg 1990; 116: 69–73.
27. Abbondanzo S, Wenig B. Non Hodgkin’s lymphoma of the sinonasal tract. Cancer 1995; 75: 1281–1291.
28. Mill WB, Lee FA, Franssila KO. Radiation therapy treatment of stage I and II extranodal non-Hodgkin’s lymphoma of the head and neck. Cancer 1980; 45: 653–661.
29. Shirato H, Tsujii H, Arimoto T, et al. Early stage head and neck non-Hodgkin’s lymphoma. The effect of tumor burden on prognosis. Cancer 1986; 58: 2312–2319.
30. Weiss LM, Gaffey MJ, Chen Y-Y, et al. Frequency of Epstein-Barr viral DNA in “Western” sinonasal and Waldeyer’s ring non-Hodgkin’s lymphomas. Am J Surg Pathol 1992; 16: 156–162.
31. Chan JK, Yip TT, Tsang WY, et al. Detection of Epstein-Barr viral RNA in malignant lymphoma of the upper aerodigestive tract. Am J Surg Pathol 1994; 18: 938–946.
32. Shibuya H, Kamiyama R-I, Watanabe I, et al. Stage I and II Waldeyer’s ring and oral-sinonasal non-Hodgkin’s lymphoma. Cancer 1987; 59: 940–946.