B-lineage lymphoblastic lymphoma is a clinicopathologic entity distinct from other histologically similar aggressive lymphomas with blastic morphology

Authors

  • Robert A. Soslow M.D.,

    Corresponding author
    1. Department of Pathology, New York Presbyterian Hospital–Weill Medical School of Cornell University, New York, New York
    • Department of Pathology, New York Presbyterian Hospital–Weill Medical School of Cornell University, Starr 1028, 525 East 68th Street, New York, NY 10021
    Search for more papers by this author
  • Rebecca N. Baergen M.D.,

    1. Department of Pathology, New York Presbyterian Hospital–Weill Medical School of Cornell University, New York, New York
    Search for more papers by this author
  • Roger A. Warnke M.D.

    1. Stanford Medical Center, Stanford, California
    Search for more papers by this author

Abstract

BACKGROUND

The authors present clinical, histopathologic, and immunophenotypic data regarding B-lineage lymphoblastic lymphoma (B-LBL), a rare entity that has not been extensively studied. To emphasize some of its unique clinical characteristics, the authors compare B-LBL with a group of histologically similar, very aggressive lymphomas, T-lineage lymphoblastic lymphoma (T-LBL) and the blastoid variant of mantle cell lymphoma (BVMCL); all were evaluated concurrently.

METHODS

Clinical data were obtained on 29 patients with very aggressive lymphomas (12 B-LBLs, 10 T-LBLs, and 7 BVMCLs) from whom paraffin-embedded material was available. The diagnoses were confirmed on review of the hematoxylin and eosin–stained slides and the immunophenotype data.

RESULTS

The mean age of patients with B-LBL was 39 years. Patients presented with both lymph node and extranodal disease, although involvement of the mediastinum and bone marrow was infrequent. Four were Stage I, 3 were Stage II, 2 were Stage III, and 3 were Stage IV. B-LBL patients were treated primarily with cyclophosphamide, hydroxydaunomycin, vincristine, and prednisone (CHOP), and one patient underwent allogeneic bone marrow transplantation. The mean follow-up time was 30 months. Seven of 11 had no evidence of disease at 48 months, whereas 4 patients were dead of disease at 5.6 months. The overall median survival was 24 months. The clinical characteristics of B-LBL patients differed significantly from those of T-LBL patients; there was more frequent bone marrow and mediastinal involvement in T-LBL cases (P = 0.03 and 0.04, respectively). T-LBL patients were also less likely to achieve a complete remission than B-LBL patients (P = 0.02). The mean age of BVMCL patients significantly exceeded that of B-LBL patients (P = 0.03).

CONCLUSIONS

The authors believe that the distinction of B-LBL from its histologic mimics, T-LBL and BVMCL, has important clinical implications. Patients with B-LBL present differently from those with the other very aggressive lymphomas studied, and they achieve complete remissions more often than T-LBL patients. Cancer 1999;85:2648–54. © 1999 American Cancer Society.

Although the clinical, morphologic, and immunophenotypic properties of B-lineage acute lymphoblastic leukemia are well known, its lymph node and soft tissue counterpart, B-lineage lymphoblastic lymphoma (B-LBL), has not been studied extensively. B-LBL is a rare neoplasm composed of immature lymphocytes that demonstrate lymphoblastic morphology and express pan-early B-cell markers, TdT and human leukocyte antigen (HLA)-DR, without surface immunoglobulin. Varying numbers of cases demonstrate reactivity for CD10 and/or CD34. Traditionally, the distinction between lymphoblastic lymphoma and acute lymphoblastic leukemia has been based on the percentage of blasts found in a staging bone marrow,1 but the revised European–American classification of lymphoid neoplasms (REAL) proposes grouping them into a single entity, precursor B-lymphoblastic leukemia/lymphoma.2 The clinical distinctiveness of the subgroup of cases that present as solid tumors without significant bone marrow or peripheral blood involvement has been reported and is acknowledged in the REAL classification.2

Because they are both neoplasms that show blastic/blastoid morphology, T-LBL and the blastoid variant of mantle cell lymphoma (BVMCL) show a great deal of histologic similarity to B-LBL; in most cases, the histologic features of B-LBL and T-LBL do not permit distinction without immunophenotyping.3, 4 We believe that the distinction between these entities merits study, especially because of the potential clinical differences between them.

T-LBL is generally a disease of late childhood or adolescence, and at least 50% of patients present with a mediastinal mass.5, 6 Both peripheral lymph nodes and extranodal sites that include Waldeyer ring, breast, skin, bone, central nervous system, gonads, liver, spleen, and bone marrow can be involved. Despite a potentially very aggressive clinical course, modern chemotherapeutic protocols have resulted in a 3-year actuarial freedom from relapse of at least 50%.7–12

The clinical characteristics of BVMCL are not well established. Blastoid morphology in a mantle cell lymphoma can occur de novo or represent a transformation from a nonblastoid mantle cell lymphoma (MCL).13, 14 With respect to mantle cell lymphoma in general, blastoid transformation is generally considered a poor prognostic feature that occurs in as many as one-third of mantle cell lymphoma patients.13–15 In general, MCL patients are men who present in late middle age or at an advanced age with disseminated disease. A significant proportion of patients with MCL and probably BVMCL can achieve a complete remission, but most patients relapse or progress within 2 years of diagnosis.14–17

The clinical features of B-LBL have not been documented thoroughly. The small series of B-LBL that have been published document clinical presentations that include prominent lymphadenopathy and bone or skin involvement.16, 18–25 Recently, Yeh et al. reported a series of 10 B-lineage lymphoblastic lymphoma patients and concluded that B-LBL is associated with a relatively favorable prognosis compared with T-cell LBL, and that, except for mediastinal involvement, the clinical presentation of both entities is similar.25 Our data here add to the literature concerning B-LBL and provide a comparison of B-LBL with T-LBL and BVMCL.

MATERIALS AND METHODS

Case Selection

We included all B-LBLs, T-LBLs, and BVMCLs accessioned in the Surgical Pathology Laboratory at Stanford University Hospital from 1984–1993 for which formalin fixed, paraffin embedded material was available. The cases were selected from data stored in a computer data base and, for the most part, represent patients referred to Stanford either for a second diagnostic opinion or for definitive therapy following diagnosis elsewhere. Following review of slides stained with hematoxylin and eosin and immunodiagnosis reports, we applied the criteria of the REAL classification to the diagnosis of B-LBL, T-LBL, and BVMCL.2 Phenotyping was performed at the time of initial presentation, using frozen tissue, if available, and formalin fixed, paraffin embedded material; these data have been published previously.3, 4 Evaluation with the following antibodies was performed subsequently and published previously: anti-CD99, CD34, TdT, and cyclin D1.3, 4 To summarize, 16 of 19 and 16 of 24 LBLs expressed TdT and CD99, respectively, and 6 of 21 expressed CD34. Eight of 8 BVMCLs expressed cyclin D1 and 5 of 8 showed coexpression of B-cell antigens with CD5 and/or CD43. None of the LBLs expressed cyclin D1 and none of the BVMCLs expressed CD99, TdT, or CD34.

Because the criteria for distinguishing lymphoblastic lymphomas from lymphoblastic leukemias is somewhat arbitrary and not universally accepted, we placed these lymphoblastic proliferations in two groups, depending on the percentage of blasts in the bone marrow, using an adaptation of the criteria described by Murphy.1 Group I neoplasms are universally accepted as lymphoblastic lymphoma: primarily lymph node or soft tissue disease with less than 25% blasts estimated in a paraffin embedded section of a staging bone marrow biopsy or aspirate. Some, but not all, investigators would consider Group II patients to have lymphoblastic lymphoma: primarily lymph node or soft tissue disease with greater than 25% blasts estimated in a paraffin embedded section of a staging bone marrow biopsy or aspirate.

Clinical Presentation and Follow-Up

The following data were obtained from surgical pathology reports and interviews with the study patients' clinicians: age at diagnosis, site of disease presentation, stage at presentation (Ann Arbor staging system),26 treatment, disease free survival, and overall survival. The Ann Arbor staging classification system is as follows: Stage I, involvement of a single lymph node region or a single extralymphatic organ or site; Stage II, involvement of two or more lymph node regions on the same side of the diaphragm or localized involvement of an extralymphatic organ or site and one or more lymph node regions on the same side of the diaphragm; Stage III, involvement of lymph node regions of both sides of the diaphragm, which may also be accompanied by involvement of the spleen or by localized involvement of an extralymphatic organ or site or both; Stage IV, diffuse or disseminated involvement of one or more extralymphatic organs or tissues, with or without associated lymph node involvement.26

Statistical Analysis

Different statistical methods were used depending on the parameters studied: analysis of variance (for age); Fisher exact test with a two-tailed P value (for gender); chi-square test (for stage and bone marrow involvement); Fisher exact test with a one-tailed P value (for lymph node vs. extranodal, mediastinal vs. not mediastinal, and DOD vs. NED); and Student t test, two-tailed (for follow-up times).

RESULTS

Morphologic Data

Morphologic data are given in Figure 1. Slides from 12 B-LBLs, 10 T-LBLs, and 7 BVMCLs were available for review. No significant differences between B-LBL and T-LBL were noted. To generalize, the cases demonstrated a diffuse growth pattern at low power and noticeable tingible body macrophages, which imparted a starry-sky appearance. High power examination demonstrated uniform, medium-sized cells, the nuclei of which contained evenly dispersed nuclear chromatin. Occasional cases demonstrated small nucleoli. The nuclear shape ranged from round to oval to convoluted. Numerous mitotic figures were appreciated.

Figure 1.

Morphologically similar aggressive blastic/blastoid lymphomas are shown. B-lineage lymphoblastic lymphoma (A) and T-lineage lymphoblastic lymphoma (B) are essentially indistinguishable on evaluation of routine sections. Lymphoblastic lymphoma (LBL) is characterized by a diffuse growth pattern at low power with noticeable tingible body macrophages that impart a starry-sky appearance (especially prominent in B). The constituent cells are mitotically active and medium-sized; the nuclei contain evenly dispersed chromatin without prominent nucleoli. The LBLs show a great deal of histologic similarity to the blastoid variant of mantle cell lymphoma (C). Subtle morphologic differences noted include the absence of numerous tingible body macrophages and a greater degree of cytologic uniformity, which is especially apparent when B and C are compared. Occasional cases showed increased perivascular collagen and numerous intermixed macrophages.

The LBLs also demonstrated significant histologic similarity to BVMCL; however, several subtle morphologic differences were noted. One case contained areas of residual nodular mantle cell lymphoma. In general, though, the growth pattern was predominantly diffuse. The cells were intermediate in size with slightly irregular nuclear contours and evenly dispersed chromatin; nucleoli were generally not apparent. Numerous mitotic figures were present. Although there were scattered macrophages, significant numbers of tingible body macrophages were not appreciated. Perivascular collagen deposition was noted in several cases. BVMCLs were slightly more uniform in nuclear size and shape than LBLs, and they contained more intermixed macrophages.

Clinical Data

Age

Clinical data are listed in Table 1. The age range of the 12 B-LBL patients studied was 8–68 years; 2 patients were children. The mean age at presentation of B-LBL was 39 years. The age range of the 10 T-LBL patients studied was 4–57 years; 3 patients were children or adolescents. The mean age at presentation of T-LBL was 26 years. The age range of the 7 BVMCLs patients studied was 44–66 years; the mean age at presentation was 58 years (P = 0.03).

Table 1. Patient Characteristics
CharacteristicB-LBLT-LBLBVMCL
  1. B-LBL: B-lineage lymphoblastic lymphoma; T-LBL: T-lineage lymphoblastic lymphoma; BVMCL: blastoid variant of mantle cell lymphoma; NS: not significant; M: male; F: female; DOD: dead of disease.

Age (yrs)392658
Range(8–68)(4–57) P = NS(44–66) P = 0.03
Gender (M:F)5:76:4 P = NS6:1 P = NS
Stage (III+IV/all patients)5/127/10 P = NS4/7 P = NS
Extranodal disease5/122/10 P = NS2/7 P = NS
Bone marrow involvement2/126/10 P = 0.033/7 P = NS
Mediastinal involvement1/125/10 P = 0.040/7 P = NS
Mean follow-up (mos)34.8 (1–108)23.5 (3–108) P = NS20.0 (4–36) P = NS
Complete remission7/111/10 P = 0.021/6 P = NS
Median overall survival (mos)2412 P = NS25 P = NS
Median survival for DOD (mos)5.616.2 P = NS25 P = NS

Sites of presenting disease

Primary sites of B-LBL involvement included peripheral lymph nodes (5 patients), mediastinum (1 patient), retroperitoneum (1 patient), soft tissue (1 patient), brain (1 patient), skin (1 patient), and breast (2 patients). Primary sites of T-LBL involvement included the mediastinum (5 patients); peripheral lymph nodes (2 patients); tonsil (1 patient); bone (1 patient); and breast, lymph node, and skin (1 patient). Five of the 7 BVMCL patients presented with disease in the peripheral lymph nodes, 1 presented in the oropharynx, and 1 in the spleen.

B-LBL patients did not differ from T-LBL and BVMCL patients with regard to lymph node disease at presentation. However, there was a significant difference between the number of B-LBL patients (1) and the number of T-LBL patients (5) with mediastinal disease (P = 0.04). B-LBL patients did not differ from BVMCL patients with regard to mediastinal involvement.

Assignment of lymphoblastic group (bone marrow involvement)

Twelve B-LBL patients' records were available for review. Ten patients were classified as belonging to Group I (less than 25% bone marrow blasts) and 2 belonged to Group II. Ten T-LBL patients' records were available for review. Four patients belonged to Group I and 6 belonged to Group II. Three BVMCL patients had bone marrow involvement, whereas 4 did not. Patients with T-LBL were significantly more likely to have bone marrow involvement than B-LBL patients (P = 0.03); BVMCL and B-LBL patients did not differ in this regard.

Stage at presentation

Four B-LBL patients were clinical Stage I at presentation (1 was extranodal), 3 were Stage II, 2 were Stage III, and 3 were Stage IV. Two T-LBL patients were Stage I at presentation (1 was extranodal), 1 was Stage II, and 7 were Stage IV. One BVMCL patient was Stage I at presentation, 1 was Stage II, 1 was Stage III, and 3 were Stage IV. These differences were not significant.

Treatment

Because many of these slides were reviewed in consultation with other institutions, chemotherapy protocols were not uniform and complete clinical summaries were not always available. Seven of the nine B-LBL patients for whom we have detailed information about treatment received CHOP chemotherapy (cyclophosphamide, vincristine, doxorubicin, and prednisone), in addition to other agents that were used for maintenance, consolidation, and/or salvage. These agents included methotrexate, bleomycin, etoposide, procarbazine, teniposide, L-asparaginase, cytarabine, and 6-mercaptopurine. One was treated with MACOP-B (methotrexate, doxorubicin, cyclophosphamide, vincristine, prednisone, and bleomycin) and one with DAP (cisplatin, cytarabine, and dexamethasone). One patient was treated with allogeneic bone marrow transplantation and stem cell infusions, three with radiation (two of whom received prophylactic cranial irradiation), and two with intrathecal methotrexate.

Six of the 10 T-LBL patients received CHOP chemotherapy; 2 received daunorubicin, vincristine, and prednisone; 1 received cytarabine, etoposide, and daunorubicin; and 1 received BACOP (bleomycin, doxorubicin, cyclophosphamide, vincristine, and prednisone) chemotherapy. Two of these patients were also treated with prophylactic intrathecal methotrexate.

Four of 6 BVMCL patients for whom we have detailed information about treatment received CHOP chemotherapy and two received CVP (cyclophosphamide, vincristine, and prednisone). Other agents used for maintenance and/or salvage included fludarabine, cisplatin, etoposide, and cytarabine.

Survival

Limited follow-up data were available on 11 B-LBL patients, 9 T-LBL patients, and 6 BVMCL patients. Follow-up periods ranged from 1 month to 9 years, and the mean follow-up time was 30 months. Seven of 11 B-LBLs were NED at an average of 4 years after presentation and 4 were DOD at 5.6 months; median survival was 24 months. With the exception of 1 T-LBL patient who was NED at 9 years, all T-LBL patients were DOD at an average of 16.2 months; the median survival was 12 months. One of 6 BVMCL patients was NED at 6 months, 4 were DOD at 25 months, and 1 was AWD; median survival was 25 months. B-LBL patients were more likely to achieve a complete remission than T-LBL patients (P = 0.02). Median survival times and the mean time to death after relapse of B-LBL versus T-LBL and BVMCL were not significantly different.

DISCUSSION

Our data show that B-LBL is a disease that usually occurs in early middle age, presents in lymph nodes or extranodal sites, and usually spares the bone marrow and mediastinum. The majority of B-LBL patients in this study achieved a complete remission. Patients who did not achieve a complete remission suffered a precipitous, downhill clinical course. Presentation in early middle age contrasted with BVMCL patients who were 25 years older, on average, than B-LBL patients; other clinical indices did not differ signficantly, although BVMCL showed a strong tendency to affect men preferentially and more frequently involved the bone marrow. BVMCL patients who did not achieve a complete remission followed a relatively prolonged course before dying of disease. T-LBL patients were more likely than B-LBL patients to have mediastinal disease and bone marrow involvement at presentation and less likely to achieve complete remissions. Although we did not identify other significant clinical differences, T-LBL patients tended to present at a younger age than B-LBL patients.

Clinicopathologic series concerning B-LBL are few, most likely because reported B-LBL cases are rare. In a large clinicopathologic study of lymphoblastic lymphomas undertaken in 1981 (before the era of routine immunophenotyping), Nathwani et al. described clinicopathologic subsets of lymphoblastic lymphoma patients, some of whom were adults, who tended to do well compared with younger patients.5 It is tempting to consider the possibility that this group of adult patients had lymphoblastic lymphomas of B-cell lineage. Several subsequent studies presented data concerning antigenically defined subgroups of lymphoblastic lymphoma, but the numbers of B-LBL patients were too small to enable the researchers to draw conclusions about the clinical significance of these subgroups.18–24 Sheibani et al., in particular, then drew attention to the possibility of the relation between lineage and survival in LBL, although data from only three patients were included in the study.23 There have also been reports that have called attention to some of the distinctive clinical features of B-LBL, compared with T-LBL. Link et al. and Sander et al. wrote about the predilection of B-LBL for cutaneous involvement,20, 21 Sander et al. described a mediastinal B-lineage lymphoblastic lymphoma,22 and Ozdemirli et al. discussed the occurrence of precursor B-lymphoblastic lymphomas that presented as solitary bone tumors.27 A recent study that evaluated the prognostic significance of B- versus T-lineage in adult lymphoblastic lymphoma presented data similar to ours.25 Our data certainly confirms these early impressions and expands on them.

Our study group of T-LBL patients compares favorably with others described in the literature.7–12 Zinzani et al., in their study of 53 adult B- and T-LBL patients, reported a mean age of 37 years, but clinical differences between patients with B- and T-lineage lymphomas were not discussed.8 In this study, 57% of the patients had mediastinal involvement, 34% had extranodal disease, 33% had positive bone marrow, and 20% had circulating blasts. Approximately one-third had Stage I or II disease, and the remainder were Stage III or IV. The complete response (CR) rate varied from 48% to 63%, depending on the chemotherapy protocol used. Fifty-five percent of the CR patients were NED at 84 months. This study confirmed that adult LBL patients fare worse than children5, 8 and that advanced age was associated with an adverse outcome. Leukemic dissemination and advanced stage at diagnosis also predicted for failure. The overall survival of the T-LBL patients reported herein was worse than that reported in more recent series.7–12 Our study differed from many other prospective, case–control series in that we analyzed retrospective data from cases reviewed at a tertiary referral center. Like the patients in the study of Zinzani et al., our T-LBL patients were predominantly adults, which might have accounted for the pessimistic clinical outcomes. Also, our survival data reflect the results of many different treatment protocols used over a period of almost 10 years.

Our clinical data are also comparable to the reported BVMCL data. Norton and Decaudin15, 17 both studied the clinical characteristics of a group of MCLs, of which BVMCLs represented a small minority (10 patients and 8 patients, respectively). These studies report occurrence in the sixth and seventh decade of life with a strong predominance of male patients. Each study documents almost universal high stage disease at presentation with frequent bone marrow involvement. Although Decaudin et al. report a 23% overall survival at 4 years, the 3-year disease free survival was nil.17 Norton reports a median survival of 36 months, but does not report on patients with blastoid morphology separately.15 Small study size may account for the inconsistent survival data. Also, several of our BVMCL patients either presented with low stage disease or had antecedent non-blastoid MCL, which might have had a favorable impact on their survival. The relatively infrequent occurrence of B-LBL and BVMCL prompted us to study these cases in the manner presented here.

We are aware of the potentially controversial contention that all of the lymphoblastic lesions studied in this report represent lymphomas, as opposed to leukemias. Eight cases demonstrating greater than 25% blasts in a staging bone marrow biopsy were considered lymphoblastic lymphoma for the purposes of this study because of significant extramyeloid disease at presentation, mostly in the absence of circulating blasts; there is a precedent for this in the literature.8 Regardless of the designation (lymphoma vs. leukemia), we show that B-phenotype lymphoblastic disease has a significantly different biologic behavior than the morphologically similar T-lymphoblastic disease. This difference holds true for those cases that are unequivocally lymphomas.

Because of the clinical differences described here and in the literature, we emphasize the importance of distinguishingB-LBL, T-LBL, and BVMCL from one another. All of these lymphomas are composed of cells with intermediate-sized nuclei that contain evenly dispersed chromatin, lack prominent nucleoli, and show very high mitotic rates; thus, evaluation of the immunophenotype is indispensable. B-LBL and T-LBL are most easily distinguished by the use of B- and T-lineage specific markers. If B-lineage has been confirmed, the use of antibodies against TdT, CD99, and PRAD-1/cyclin D1 can facilitate the distinction of B-LBL from BVMCL; B-LBL expresses TdT and/or CD99, whereas BVMCL lacks these markers and expresses PRAD-1/cyclin D1.4 Therefore, our data emphasize that the immunophenotypic subclassification of blastic and blastoid lymphomatous neoplasms predicts for several specific clinical indices and, more importantly, correlates with the achievement of complete remission.

Acknowledgements

The authors thank Roseline Russell for her help in obtaining clinical data.

Ancillary