The frequency of marrow involvement in non-Hodgkin's lymphomas at diagnosis is highly variable, ranging from as low as 3% in primary mediastinal large B-cell lymphoma to as high as 73% in B-cell small lymphocytic lymphoma/chronic lymphocytic leukemia and lymphoplasmacytoid lymphoma (12
). It is generally higher in small cell lymphomas and follicular lymphomas. The reported frequency of marrow involvement by NK/T cell lymphoma is known to be low, and is based on morphologic examination in most studies (4
). However, morphologically normal marrow may harbor occult lymphoma cells as revealed only by more sensitive techniques for immunoglobulin or T-cell receptor gene rearrangements (13
). For example, in the study by Fraga et al. (14
), only 17% of the patients with anaplastic large cell lymphoma had marrow involvement by conventional morphologic criteria, while immunostaining for CD30 or epithelial membrane antigen showed occult malignant cells in 23% of patients with negative marrow on histologic examination. Likewise, the morphologic identification of neoplastic involvement of marrow by nasal-type NK/T cell lymphoma is difficult unless the involvement is extensive. Isolated tumor cells can be easily overlooked, especially because the neoplastic cells are often small to medium-sized.
In NK/T cell lymphoma, CD56 is a sensitive marker expressed in the majority of the cases. However it is of no use in detecting tumor cells in the bone marrow biopsy specimens, because decalcification process with acid destroys CD56 antigen on the cell surface. EBV ISH is a sensitive technique because normal marrow rarely contains EBV-positive cells (2
). In the present study, we determined the frequency of lymphoma involvement of the marrow by nasal-type NK/T cell lymphoma with the use of EBV ISH. The results confirm that marrow involvement by nasal type NK/T cell lymphoma is uncommon at diagnosis and EBV in situ hybridization is a useful tool in detecting minimal infiltration of EBV infected tumor cells.
Two patients with bone marrow involvement had disseminated disease proven by clinical and radiologic studies at diagnosis, which suggests that bone marrow involvement in nasaltype NK/T cell lymphoma is a late event in the course of disease. Another patient with bone marrow involvement had a colonic disease and was initially diagnosed as stage I by Ann Arbor Staging System (16
). It is noteworthy that the lesion in the colon involved the entire colonic segments diffusely, which indicates higher tumor burden than expected in the estimated stage.
Ann Arbor Staging System is based on nodal lymphomas such as Hodgkin's lymphoma and B-cell lymphoma. In contrast to extranodal B-cell lymphoma which usually forms a single mass, multifocal or diffuse ulcerative involvement of one or two anatomical segments of the intestine is a common finding in NK or T cell lymphoma (17
). In Ann-Arbor Staging System, diffuseness or multifocality of the lesion in NK or T-cell lymphoma are not taken into account, which can lead to underestimation of the tumor stage.
It is difficult to establish the definite prognostic significance of marrow involvement in nasal type NK/T cell lymphoma because of the aggressive nature of the disease with very short survival time in the majority of patients and the small number of cases with marrow involvement. In this study, all three patients with bone marrow involvement died of the disease.
In conclusion, minimal involvement of bone marrow is infrequent in nasal-type NK/T cell lymphoma. And EBER ISH is a useful technique for identifying minimal involvement of bone marrow by nasal-type NK/T cell lymphoma, which is easily overlooked but closely related with poor prognosis.