OUP user menu

Follicular Lymphoma of the Spleen
Multiparameter Analysis of 16 Cases

Matthew T. Howard MD, Scott Dufresne MD, Steven H. Swerdlow MD, James R. Cook MD, PhD
DOI: http://dx.doi.org/10.1309/AJCPF9V8XRDYWTIR 656-662 First published online: 1 May 2009


Follicular lymphoma (FL) involving the spleen must be distinguished from reactive hyperplasia and from other lymphomas. A prior study reported that splenic FLs frequently lack BCL2 expression, further complicating diagnosis. We examined 16 cases of splenic FL, including 12 cases initially diagnosed at splenectomy. Two morphologic patterns were identified: one with architectural abnormalities (AA) and one with an extensive architectural preservation (AP) pattern. Newly diagnosed AP cases were associated with older age (P = .051) and grade 1 histologic features (P = .023). All cases displayed a CD10+/BCL2+ phenotype. Cytogenetics and FISH identified IGH/BCL2 or BCL6 translocations in all tested cases. Splenic FLs display phenotypic and cytogenetic findings similar to nodal FLs. However, splenic FLs frequently display an exclusively intrafollicular growth pattern resembling so-called in situ FL. Recognition of subtle FL with preserved architecture is important because patients may have overt FL at other sites or the FL may progress to overt nodal disease.

Key Words:
  • Spleen
  • Follicular lymphoma
  • Intrafollicular
  • In situ

The diagnosis and classification of small B-cell lympho-proliferative disorders in splenectomy specimens is frequently challenging. Splenic marginal zone B-cell lymphomas, by definition, display splenic involvement at initial diagnosis.1,2 However, many other small B-cell lymphoproliferative disorders may also be associated with splenomegaly at initial diagnosis or after dissemination of initially nodal disease.3,4 Previous studies have shown that many types of B-cell lymphomas may display morphologic features that overlap with splenic marginal zone lymphomas, including nodular expansion of the white pulp and conspicuous marginal zones.57 Establishing a precise diagnosis, therefore, requires correlation of morphologic features with the results of phenotypic studies and/or other ancillary studies.

Follicular lymphoma (FL) is typically disseminated at initial diagnosis, and there is frequent splenic involvement.8,9 A previous study of B-cell lymphomas involving the spleen reported that 6 of 11 cases of FL were negative for BCL2 by immunohistochemical analysis.5 This result, a much lower incidence of BCL2 positivity than generally reported in nodal FL,8,10 suggested that FL involving the spleen may display an unusual phenotypic profile that could further complicate appropriate diagnosis. We therefore studied the morphologic, phenotypic, and genotypic findings in FL involving the spleen.

Materials and Methods

Case Selection

Cases of FL involving the spleen were identified from the case files of the Cleveland Clinic, Cleveland, OH, and University of Pittsburgh Medical Center, Pittsburgh, PA. Available clinical data, routine H&E stains, and results of previously performed flow cytometry, immunohistochemical analysis, metaphase cytogenetics, or fluorescence in situ hybridization (FISH) studies were reviewed. Cases were classified using 2008 WHO criteria.11

Immunohistochemical Analysis

In all cases with available archived material, immunohistochemical analysis was performed using an automated immunostainer (BenchMark, Ventana, Tucson, AZ) with CC1 cell conditioning. The following antibodies were used: CD3 (polyclonal, dilution 1:600; Cell Marque, Rocklin, CA), CD20 (L26, dilution 1:50; DAKO, Carpinteria, CA), CD10 (56C6, dilution 1:5; Novocastra, Newcastle upon Tyne, England), BCL6 (PG-B6p, dilution 1:5; DAKO), and BCL2 (124, prediluted, Ventana).

FISH Studies

In all cases with available archived material, FISH was performed on intact 5-μm sections using a dual-color, dual-fusion IGH/BCL2 probe (Abbott Molecular, Abbott Park, IL) as previously described in detail.12 Briefly, slides were baked overnight at 60°C, deparaffinized, and subjected to Proteinase K treatment. Following wash treatment, 10 μL of probe solution was applied, and probe and target DNA were allowed to codenature at 73°C for 5 minutes and hybridize overnight at 37°C. Slides were counterstained with 4′-6,-diamidino-2-phenylindole, and signals were visualized on an Axioskop photomicroscope (Zeiss, Oberkochen, Germany). Atypical germinal centers were specifically targeted for analysis by comparison with serial H&E-stained sections. In each case, 200 nuclei were scored. Nuclei were considered to be abnormal when at least 1 fusion signal was identified. Analysis of 10 negative samples established a cutoff of more than 8% abnormal nuclei for interpretation as a positive result (mean plus 3 SD of negative samples).


Continuous variables were analyzed by using the Student t test and categorical variables by using the Fisher exact test or χ2 test for trend, as indicated. All statistical calculations were performed using GraphPad Prism 4 software (Graph Pad Software, La Jolla, CA).


We identified 16 cases of FL involving the spleen. In 12 cases, the diagnosis of FL was first established by splenectomy, which was performed for evaluation of abnormalities found on imaging studies (5 cases), unexplained cytopenias and splenomegaly (4 cases), or splenic rupture (3 cases). In 1 case, a benign hemangioma was also present, and 1 case was also involved by chronic myelomonocytic leukemia. In the remaining 4 cases, a diagnosis of FL had been previously established 1, 3, 96, or an unknown number of months before splenectomy. In these cases, splenectomy was performed for symptomatic splenomegaly and thrombocytopenia (2 cases), staging purposes (1 case), or incidentally as part of a distal pancreatectomy (1 case).

Morphologic Findings

FL Diagnosed at Splenectomy

Review of routine H&E-stained sections identified 2 morphologic patterns Image 1A and Image 1B . The architectural abnormalities (AA) pattern was defined by abnormal architecture with at least focal nodular regions consisting of closely packed neoplastic follicles separated by minimal bands of red pulp. The architectural preservation (AP) pattern was defined by the presence of preserved architecture with only scattered neoplastic follicles and unremarkable red pulp.

In the 12 cases with FL initially diagnosed in the spleen, 6 displayed the AA pattern and 6 the AP pattern. The clinicopathologic features associated with the AA and AP patterns in newly diagnosed FL are summarized in Table 1 . The presence of the AA pattern was associated with a trend toward younger age (mean ± SD, 63.5 ± 3.5 vs 73.5 ± 3.0 years; P = .051; Student t test). Newly diagnosed cases with an AA pattern were also associated with higher histologic grade (2/6 grade 1 vs 6/6 grade 1; P = .023, χ2 test for trend). The AA vs the AP pattern did not correlate with sex, splenic weight, or clinical stage.

Increased numbers of interfollicular B cells were identified by immunohistochemical analysis for CD20 in 5 of 6 AA cases, whereas the 6 AP cases each showed only scattered interfollicular or red pulp B cells. Large germinal centers (>1 mm) were identified in 5 of 6 AA cases and in 0 of 6 AP cases (P = .015). Prominent marginal zones (occupying >50% of the white pulp nodule diameter) were found in 6 of 6 AP cases and 0 of 6 AA cases (P = .002) Image 1C. Mantle zones were attenuated or absent in 9 cases (6 AA and 3 AP) and retained in 3, each of which showed the AP pattern with prominent marginal zones Image 1D. In 1 case with an AP pattern, hilar lymph nodes and an incidental cystic duct lymph node received with a concurrent cholecystectomy specimen each showed focal involvement by FL with a purely intrafollicular growth pattern.

FL Diagnosed Before Splenectomy

Of the 4 cases with a prior diagnosis of FL, 2 displayed an AA pattern (1 grade 1 and 1 grade 2) and 2 displayed the AP pattern (1 grade 1 and 1 grade 2). Detailed clinicopathologic features could not be further analyzed owing to the small number of cases in this subset. Hilar lymph nodes were involved in 3 cases, 2 of which displayed the AP pattern and 1 AA morphologic features. The lymph node and splenic morphologic patterns were concordant in each case.

Image 1

Histologic features of splenic follicular lymphoma. A, Scattered neoplastic germinal centers (inset) are present in this spleen sample with complete architectural preservation (AP pattern) (H&E, x40; inset, x400). B, Closely packed, often enlarged neoplastic follicles (inset) are present in this spleen sample with architectural abnormalities (AA pattern) (H&E, x40; inset, x400). C, A prominent marginal zone surrounds neoplastic germinal centers (H&E, x100). D, This case with AP demonstrates mantle zones around the neoplastic germinal centers (H&E, x100).

Phenotypic and Molecular Cytogenetic Findings for All Cases

Flow cytometric immunophenotypic studies were performed in 11 cases, 7 of which (64%) demonstrated a monoclonal B-cell population (AP, 1/4; AA, 6/7; P = .088; Fisher exact test). By immunohistochemical analysis, the neoplastic germinal centers were uniformly positive for CD10 (16/16 [100%]) and BCL6 (8/8 [100%]). Aberrant BCL2 positivity in germinal center cells was identified in every case studied (15/15 [100%]) Image 2. In each case with prominent marginal zones, the marginal zone cells were negative for BCL6 and CD10.

Metaphase cytogenetic studies were performed in 2 cases, and paraffin section interphase FISH studies for IGH/BCL2 were performed in all cases with available archived material Table 2. Together, these studies demonstrated that 10 cases contained a t(14;18)(q32;q21) IGH/BCL2 Image 3, whereas 1 case, following metaphase FISH studies, was shown to contain a cryptic 3-way translocation involving BCL6, IGH, and a region on chromosome 11q that did not include CCND1. In 2 cases with the AP pattern and prominent marginal zones, residual paraffin section FISH slides were available for detailed morphologic correlations; in each case, IGH/BCL2 fusion signals were identified within the germinal centers, whereas fusion signals seemed to be absent in surrounding marginal zones.

View this table:
Table 1

Clinical Follow-up

Clinical follow-up data were available for 13 patients who were treated by a variety of strategies (watchful waiting, 7; single agent rituximab, 3; other combination chemotherapy, 2; and allogeneic peripheral blood stem cell transplantation, 1). Overall, with a median follow-up of 20 months (range, 1–115 months), 1 patient died of disease, 2 died of other causes, and 10 were alive with stable (n = 7) or progressive (n = 3) disease. Of 7 patients with stage I disease first diagnosed at splenectomy and with follow-up information, 2 with AA patterns had progression of disease at 20 and 38 months, and 1 with an AP pattern had evidence of progression after 5 months; 4 patients with the AP pattern showed no clinical adenopathy at 1, 3, 6, and 34 months.


This study demonstrates that FLs involving the spleen may display 1 of 2 histologic patterns. The first pattern consists of aggregates of closely packed neoplastic follicles, often enlarged, that disrupt the splenic architecture and frequently show an interfollicular proliferation of neoplastic B cells. This pattern resembles typical examples of nodal FL. In contrast, the second pattern consists of only scattered neoplastic germinal centers with an entirely intrafollicular growth pattern and retention of the overall splenic architecture. This pattern was also associated with prominent marginal zones, and 3 cases displayed intact mantle zones as well. Cases of FL displaying this pattern with extensive AP may, therefore, be easily misdiagnosed as reactive lymphoid hyperplasia. It is interesting that the 2 patterns did not predict the presence or absence of previously diagnosed FL and were not associated with differences in clinical stage or splenic weights. In the cases first diagnosed at splenectomy, however, the AP pattern was associated with grade 1 morphologic features and a trend to older age of the patients.

The cases of FL with AP in this series closely resemble nodal cases previously reported as “in situ localization of follicular lymphoma.”13 Cong et al13 reported on 23 lymph node biopsy specimens containing rare BCL2+ germinal centers in lymph nodes with extensive AP. Of 18 patients with clinical follow-up data available, 5 were reported to have synchronous biopsies with typical, overt FL at other sites; in 3, overt FL developed after 3, 13, or 72 months; and in 10 patients, clinically apparent disease never developed, with a short median follow-up of 15 months. In the present study, similar findings were observed. Of the 8 patients with extensive splenic AP, 3 had a previous diagnosis of FL or clinical evidence of disease outside the spleen at the time of splenectomy, and in 1, computed tomographic evidence of retroperitoneal disease developed after 5 months. The remaining 4 cases displayed no clinical evidence of overt lymphoma after 1, 3, 6, or 34 months of follow-up.

In the light of current models of the pathogenesis of FL, the precise biologic nature of the purely intrafollicular cases described in this report is uncertain. The IGH/BCL2 translocation is currently thought to arise in a pre–B-cell precursor within the bone marrow as a result of abnormal D-J or V-DJ recombination.14,15 The resulting naive B cell containing the IGH/BCL2 translocation is then thought to circulate through the periphery and eventually enter a germinal center in which additional mutations or other abnormalities may accumulate. Following such a “second hit,” overt FL is thought to arise. In the absence of a second pathogenetic abnormality, however, the translocation-bearing cells may persist without ever developing clinical evidence of lymphoma. Indeed, sensitive polymerase chain reaction techniques have shown B cells bearing the IGH/BCL2 translocation in 12% to 47% of healthy people,1619 consistent with the proposal that an IGH/BCL2 translocation is not sufficient for development of FL. B cells containing IGH/BCL2 translocations isolated from healthy people have also been shown to have undergone class switch rearrangements,15 indicating that apparently “preneoplastic” IGH/BCL2+ cells can exit the germinal center and disseminate through the blood.

Image 2

Immunohistochemical analysis of splenic follicular lymphoma. A, BCL2 expression is seen focally in the neoplastic germinal centers in a case with extensive architectural preservation (AP pattern) (x40). The staining is much more intense than in the surrounding marginal zones. B, Many white pulp nodules in a case with abnormal architecture (AA pattern) showing strong central BCL2 expression with weaker staining in the marginal zone regions (x40). C and D, Germinal centers are uniformly positive for BCL6 (C, x100) and CD10 (D, x100). CD10+ cells outside the germinal centers predominantly consisted of granulocytes.

It is possible that at least some of the purely intrafollicular cases in this series represent the histologic counterparts of premalignant proliferations of IGH/BCL2+ germinal center cells that have not yet acquired a secondary abnormality required for development of overt lymphoma. The association of the purely intrafollicular growth pattern with grade 1 morphologic features in the newly diagnosed cases in this study is consistent with this hypothesis. Additional studies with longer clinical follow-up will be required to determine the ultimate biologic significance of this subset of cases. However, because the current report and that of Cong et al13 show that at least some patients with FL displaying a purely intrafollicular growth pattern may have concurrent overt lymphoma elsewhere or clinically overt disease may eventually develop, this architectural pattern remains important to recognize and diagnose in routine practice. In addition, as illustrated in 3 cases with extensive AP and a purely intrafollicular growth pattern in the spleen and in splenic hilar or cystic duct lymph nodes, FL with an in situ growth pattern is not necessarily restricted to one anatomic site but may be disseminated.

View this table:
Table 2
Image 3

Fluorescence in situ hybridization analysis of splenic follicular lymphoma. Analysis using a dual-color, dual-fusion IGH/BCL2 probe identified cells with 2 yellow fusion signals indicating the presence of an IGH/BCL2 translocation.

A previous study of B-cell lymphomas with prominent splenomegaly reported that 6 (55%) of 11 cases of splenic FL lacked BCL2 expression.5 More recently, the same group reported in abstract form that 13 (41%) of 32 cases of splenic FL lacked CD10 expression and 20 (63%) of 32 lacked BCL2 expression.20 Together these findings suggested that splenic FL may be less likely than nodal FL to show CD10 and BCL2 coexpression. In contrast, the cases examined in the present study universally displayed a CD10+/BCL6+/BCL2+ phenotype, similar to that expected in typical nodal FLs. The source of the discrepancy between these reports is uncertain. However, because the cases studied by Mollejo et al20 were initially selected as B-cell lymphomas with prominent splenomegaly that may mimic splenic marginal zone lymphoma, selection bias for an unusual subset of cases may contribute to these differences.

Interphase FISH and/or metaphase cytogenetic studies in 11 cases in this study demonstrated 10 IGH/BCL2 translocations and 1 complex cryptic BCL6/IGH translocation. These results are similar to those reported in previous large series of typical nodal FL.10 Furthermore, these results demonstrate the usefulness of molecular cytogenetic analysis in addressing the differential diagnosis of splenic B-cell lymphoproliferative disorders. In particular, the use of intact paraffin section interphase FISH analysis allowed for specific examination of atypical germinal centers identified in serial H&E-stained sections, even when only very focal neoplastic follicles were present. Analysis of single-cell suspensions or purified preparations of nuclei likely would not have been able to demonstrate these focal cytogenetically abnormal populations.

This study demonstrated that cases of FL involving the spleen exhibit morphologic, phenotypic, and cytogenetic findings similar to those reported in typical nodal FL. However, splenic FL may also display extensive AP and a purely intra-follicular growth pattern that may easily be misdiagnosed as reactive hyperplasia. Careful morphologic evaluation of atypical-appearing germinal centers and appropriate use of immunohistochemical stains, especially for BCL2, will assist in addressing this differential diagnosis. Cases displaying a purely intrafollicular growth pattern may have clinically overt lymphoma at other sites or eventually progress to overt nodal disease, emphasizing the clinical importance of recognition of subtle cases.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
View Abstract