Journal Logo

Research Articles

EBV-Positive Primary Large B-Cell Lymphoma: The Role of Immunohistochemistry and XPO1 in the Diagnosis of Mediastinal Lymphomas

Maracaja, Danielle L.V. MD*; Puthenpura, Vidya MD; Pels, Salley G. MD; O’Malley, Dennis P. MD; Sklar, Jeffrey L. MD, PhD*; Finberg, Karin E. MD, PhD§; Xu, Mina L. MD*

Author Information
Applied Immunohistochemistry & Molecular Morphology: November/December 2020 - Volume 28 - Issue 10 - p 725-730
doi: 10.1097/PAI.0000000000000820
  • Free


Primary mediastinal large B-cell lymphoma (PMBL) is considered a distinctive clinicopathologic subtype of diffuse large B-cell lymphoma (DLBCL). They often present as large masses (>10 cm in 60% to 70% of cases) located in the anterosuperior mediastinum and are associated with invasion of adjacent structures, airway compromise, and symptoms related to the compression and encasement of mediastinal organs. PMBL occurs predominantly in young adults with a median age of 35 years.1

The neoplastic B cells in PMBL are large with centroblasts morphology and cytoplasmic clearing, associated with various degrees of fibrosis and sclerosis.2 Occasional pleomorphic, multilobated cells, resembling Reed-Sternberg cells may be seen. The lesion is mostly compartmentalized, but it may also show focal nodularity, resembling classic Hodgkin lymphoma (CHL). CHL itself can present initially in the mediastinum, show scattered lacunar cells as well as Reed-Sternberg cells in a characteristic background rich with inflammatory cells. Signs that favor CHL include the typically scattered, rather than diffuse, distribution of neoplastic cells. Another diagnostic consideration is DLBCL with secondary mediastinal involvement. However, that entity typically presents with advanced, multifocal disease, and more aggressive behavior. Finally, mediastinal gray-zone lymphoma (MGZL) exhibits morphologic and immunophenotypic findings intermediate between PMBL and CHL.3

Pan B-cell lineage markers are typically expressed in PMBL, including CD19, CD20, CD22, and CD79a1,4 and the transcription factors PAX5, BOB1, and OCT2, whereas CD23 is expressed in ∼60% of PMBL and only in 1.3% of CHL.4 CD30 is usually present in >80% of PMBL cases, showing weak and heterogeneous expression; by contrast, CHL shows strong and homogeneous CD30 expression.1 Approximately 70% of PMBL express MAL antigen, whereas DLBCL is rarely positive for MAL.4 Strong positivity for CD79a, BCL6, and BOB1 are useful in excluding a diagnosis of CHL, especially in conjunction with the interpretation of CD15, CD20, CD23, CD30, and Epstein-Barr virus (EBV) latent membrane protein-1 (LMP‐1). In cases for which the diagnosis is still unresolved, MUM1 and p63 can also help distinguish between CHL and PMBCL. MUM1 is present in ∼67% of CHL, whereas it is positive in only 5% of PMBL.5 p63 is frequently expressed in PMBL (in>80% of cases), but it has also been reported in CHL (<5%).5

Since the diagnosis of PMBL can be particularly challenging due to the overlap of cytologic and immunophenotypic features found in other mediastinal lymphomas, the presence or absence of EBV becomes a valuable diagnostic marker. PMBL is thought to be almost always negative for EBV,1 whereas 10% to 25% of nodular sclerosis CHL, 75% of mixed cellularity CHL, <5% to 15% of DLBCL, and only rare cases of MGZL (mostly associated with CHL morphology) are EBV-positive.1

There are a few reports of EBV-positive PMBL, and to the best of our knowledge, none of these cases have undergone mutational profiling. This is the first case of EBV-positive PMBL with next-generation DNA sequencing data, which help support the diagnosis and suggest a targeted therapeutic option.


We report a case of a 19-year-old individual, who presented with a 2-week history of nonproductive cough, vocal changes, and occasional dysphagia. An ultrasonogram showed a solid mass lying below the inferior pole of the left thyroid gland, which measured 5.5 cm×2.9 cm×6.0 cm, with partial compression of the adjacent internal jugular vein that contained a substantial nonocclusive thrombus. On review of systems, the patient reported a 10-pound unintentional weight loss over 1 month. He denied having fevers, night sweats, bruising, bleeding, or any other lymphadenopathy. His physical examination was notable for tachycardia and a palpable, firm mass in the suprasternal region. No other lymphadenopathy was noted.

Chest computed tomography scan with contrast was significant for a large expanding mass that extended and enlarged the mediastinal and pericardial spaces. The mass was noted to encase all of the major mediastinal vessels including the central pulmonary artery, ascending aorta, aortic arch, and superior vena cava, as well as the trachea and esophagus (Fig. 1). A whole-body positron emission tomography scan was negative for the disease other than at the primary site.

Computed tomography scan images of the chest in the reported case. Coronal (A) and transverse (B) images of the chest highlighting a markedly widened mediastinum with a large bulky mass (5.5 cm×2.9 cm×6.0 cm) at the anterior mediastinum, originating just below the thyroid gland and encasing the esophagus, trachea, and the major mediastinal vessels.

On morphologic examination, sections of paraffin-embedded fixed tissue of the biopsy sample, stained with hematoxylin and eosin, showed a compartmentalized lesion with a diffuse infiltrate of intermediate to large cells containing predominantly round nuclei, vesicular chromatin, distinct nucleoli, and abundant clear cytoplasm. The large cells were admixed with small lymphocytes and histiocytes, separated by thin, delicate fibrous bands, creating compartmentalized sclerosis, and focal angiocentricity.

Immunohistochemical stains showed that the large neoplastic cells were positive for CD45, positive for all B-cell markers tested [CD20, CD19, CD79a, OCT2, BOB1, and PAX5 (strong)], as well as for MUM1, BCL6, CD23, and p63. There was a variable expression of CD30 (in ∼20% of the cells). Assessment for Epstein-Barr virus–encoded RNA by in situ hybridization (EBER ISH) showed strong positivity in the large tumor cells without nonspecific reactivity in small background lymphocytes, whereas EBV LMP-1 was negative by immunohistochemistry. CD15, CD10, and human herpesvirus 8 were negative. An immunostain for MAL (performed at NeoGenomics Laboratories) was positive in tumor cells. Fluorescence in situ hybridization for MYCC gene rearrangement was negative. The morphologic and immunophenotypic findings are summarized in Figures 2 and 3. A summary of the immunohistochemical stains is shown in Table 1.

Histologic and immunophenotypic findings in the reported case. Hematoxylin and eosin stain highlights a compartmentalized lesion with large lymphocytes with irregular nuclear contours, vesicular chromatin, distinct nucleolus, and a moderate amount of clear cytoplasm, separated by thin delicate fibrosis with occasional angiocentric (A–C). CD3 highlights small T cells in the background (D). CD20 highlights the large cells (E), which also express CD79a (F), and OCT2 (G). CD30 shows positive, variable staining (H). CD15 is negative in the neoplastic cells (I). The neoplastic cells demonstrate positivity for p63 (J) and MAL (K). L, The large cells show nuclear positivity for Epstein-Barr virus–encoded RNA by in situ hybridization [original magnification: A (×40), B (×200), C–L (×400)].
Hematoxylin and eosin stain (original magnification, ×600) showing the morphologic details of the lesion. The tumor cells are intermediate to large, with predominantly round to irregular nuclear contours, vesicular chromatin, distinct nucleoli, and clear cytoplasm, admixed with small lymphocytes and histiocytes.
TABLE 1 - Summary of Immunohistochemical Findings in the Reported Case
Antibodies Staining Pattern Observed Antibody Staining Pattern Observed
CD45 Positive in all B and T cells BCL6 Positive in the large cells
CD3 Highlights the small reactive T cells P63 Positive in the large cells
CD19 Positive in the large cells MAL Positive in the large cells
CD20 Positive in the large cells CD30 Variable expression in the large cells (∼20% of the cells)
CD79a Positive in the large cells EBER ISH Strongly positive in the large cells/negative in the small cells
PAX5 Strongly positive in the large cells CD23 Positive
OCT2 Positive in the large cells CD15 Negative
BOB1 Positive in the large cells CD10 Negative
MUM1 Positive in the large cells LMP-1 Negative
EBER ISH indicates Epstein-Barr virus–encoded RNA by in situ hybridization; LMP-1, latent membrane protein-1.

Mutational profiling of the lymphoma was performed by the Yale-New Haven Hospital Tumor Profiling Laboratory using the next-generation sequencing-based Oncomine Comprehensive Assay (version 3; Thermo Fisher Scientific), which interrogates 161 cancer-related genes for point mutations, deletions, insertions, amplifications, and gene fusions. The patient’s germline DNA (derived from a buccal swab) was analyzed in parallel. Variant calling, filtering, and annotation were conducted by clinically validated bioinformatics pipelines. Somatic mutations were detected in 3 genes: XPO1, SMARCB1, and MYCC. Each of these variants was detected at an allelic fraction of 7% in the tissue sample.

Subsequently, a quantitative polymerase chain reaction for EBV DNA was performed on peripheral blood and showed positive results with 22,371 copies/mL (log10=4.35). After the biopsy diagnosis was made, the patient received treatment as per Children’s Oncology Group protocol ANHL1131, which includes dose-adjusted rituximab, etoposide, prednisone, vincristine, doxorubicin, and cyclophosphamide (DA-R-EPOCH). He also received enoxaparin for his internal jugular thrombus.

Informed consent was obtained from the patient both for molecular/genetic studies and for this case report.


The diagnostic workup of primary lymphomas of the mediastinum gives rise to a differential that includes CHL, PMBL, DLBCL, NOS, and MGZL. Lymphoblastic lymphoma is another consideration at this site but is morphologically and immunophenotypically distinct. It is essential to differentiate between these lymphomas because the therapeutic approaches are different.6 Given overlapping features described in the World Health Organization classification, it has long been felt that EBV positivity of tumor cells argues strongly for a diagnosis of CHL. We report here the first case of EBV-positive PMBL in which extensive mutational profiling was performed. Our studies reveal potentially useful molecular data to assist in this diagnostic dilemma.

On the basis of the cytomorphologic features and immunohistochemistry performed, the findings most consistent with PMBL. CHL was considered unlikely given the absence of Hodgkin/Reed-Sternberg cells and typical inflammatory background, as well as the fact that the tumor cells were strongly positive for CD45 and all B-cell markers tested. In addition, the tumor cells were negative for CD15, and only variably reactive for CD30. Furthermore, the tumor also demonstrated reactivity for MAL, which is expressed by PMBL neoplastic B cell (∼70% of cases) but is rarely detected in the typical DLBCLs and Hodgkin lymphomas.4,7 The MAL protein is commonly expressed by T cells but is not generally expressed in the B-cell lineage. MAL expression has been found in thymic B cells and is likely to correlate with the cell of origin in PMBL. The function of MAL in the pathogenesis of the PMBCLs is not understood, but it may play a role in membrane microdomain stabilization, transport, and signal transduction.4

In this case, the only evidence pointing toward a diagnosis of CHL was positivity for EBV, which is typically negative in PMBL.1,2 The EBV positivity also raises the possibility of EBV-positive DLBCL, but the presentation of a solitary mediastinal mass in a young patient made this unlikely. Furthermore, the patient was found on this visit to be negative for X-linked lymphoproliferative disease by genetic testing, negative for human immunodeficiency virus, and lacked a clinical history suspicious for primary immunodeficiency.

EBV positivity has been detected in the tumors of up to 50% of patients with Hodgkin lymphoma, with a slightly higher incidence among Asians and Hispanics, when compared with the whites or African American populations.8 In DLBCL, EBV positivity usually occurs in patients above 50 years of age.1 EBV-positive DLBCL accounts for <5% to 15% of DLBCL among Asian and Latin American patients and <5% of DLBCL among European and North American patients, with no documented predisposing immunodeficiency.1 The lungs and gastrointestinal tract are commonly involved in older patients, and nodal involvement is noted in young patients. Most cases of MGZL are negative for EBV, although recent reports from European and Asian countries have suggested that almost 24% of MGZL may be EBV-positive. Given some controversy in this area, it has been suggested that a diagnosis of EBV-positive MGZL should only be considered when there is a strong expression of both CD20 and CD30 in tumor cells.9 MGZL, also known as B-cell lymphoma, unclassifiable, with features intermediate of DLBCL and CHL typically lack nodular growth pattern and fibrous bands and present as an amalgam of morphologic and phenotypic features seen in PMBL and CHL.10

Given that not all of the individual mediastinal lymphomas subtypes have undergone extensive genomic analysis to date, results from any mutational profiling studies must be interpreted in light of all other pathologic data. In the present case, the tissue sections used for next-generation sequencing were estimated by histology and immunophenotype to contain ∼15% to 20% malignant cells. In this setting, the allele fractions of the detected variants (all at 7%) are compatible with heterozygous mutations.

XPO1 encodes exportin-1 (also referred to as chromosome region maintenance 1, CRM1), a ubiquitous nuclear export receptor that transports macromolecules from the nucleus to the cytoplasm. These molecules consist of numerous RNAs and proteins, including those with tumor suppressor functions such as p53, BRCA1, survivin, nucleophosmin, APC, and FOXO.11 Because many of these proteins exert their tumor suppressor effects in the nucleus, the upregulation of exportin-1 promotes tumorigenesis and cancer progression. A number of exportin proteins have been identified in eukaryotes and have been found to be overexpressed in lymphomas, melanomas, and cancers of the pancreas, breast, colon, stomach, prostate, esophagus, and lung.12

The XPO1 variant identified in the present case results in the missense substitution Glu571Lys (p.E571K) and is the most common XPO1 variant in the Catalogue of Somatic Mutations in Cancer (COSMIC) database of all published cancer mutations. In 1 study, XPO1 mutations were identified in 24% of PMBL cases (28/117), and 26% of CHL cases (5/19), whereas they were absent or rare MGZL (0/20) and DLBCL (3/197).11,13 In addition, PMBLs defined by gene expression profiling showed an even higher prevalence (50%) of the E571K mutation.13 Although its mechanism of oncogenicity is incompletely understood, the E571K amino acid substitution has been found to alter the binding preferences of exportin-1 for certain nuclear export cargo. For example, it alters the nuclear export of nuclear factor-κB signaling.13

XPO1 has been considered a potential target for anticancer therapy because of its abundance in several cancer types. A selective inhibitor of nuclear export pharmacological agents is in phase 1 and phase 2 clinical trials for hematologic malignancies. Specifically, selinexor (KPT-330), a selective nuclear export inhibitor, has shown clinical activity in several tumor types, particularly acute myelogenous leukemia and multiple myeloma,12 and is currently under evaluation both as a single agent and in combination with chemotherapy in multiple clinical trials. In vitro studies suggest that the binding of KPT-330 to exportin-1 is not markedly altered by the presence of the E571K mutation.13

The 2 other mutations detected in this case are less contributory with respect to diagnosis. Rearrangements or point mutations in MYCC have been detected occasionally in PMBL.14 Among lymphomas, point mutations in the MYCC coding region have been found to cluster around the phosphorylation site threonine 73 (Thr73; also known as Thr58 with respect to an alternative MYCC isoform). The MYCC p.Thr73Ala missense variant detected, in this case, is located within this mutational hotspot and has been reported in 5 lymphoid neoplasms, including 3 DLBCLs, in the COSMIC database. The mutation of Thr73 (or other residues required for Thr73 phosphorylation) has been shown to stabilize the MYCC protein by impairing its proteasomal degradation.15

The third variant detected, in this case, was in SMARCB1 (SWI/SNF-related matrix-associated actin-dependent regulator of chromatin subfamily B member 1, also known as INI1). SMARCB1 encodes a core component of a chromatin remodeling complex involved in the regulation of gene activity. SMARCB1 mutations have been detected at very high frequency in malignant rhabdoid tumors, epithelioid sarcomas, and renal medullary carcinomas, as well as a lower frequency in several other tumor types.16 To date, SMARCB1 mutations have not been associated with PMBL. The SMARCB1 variant detected in the present case (p.Leu356AspfsTer4), which results from a 2-nucleotide deletion, alters the reading frame of the encoded protein and is expected to result in premature termination of the protein at amino acid 360 of 385. Thus, this variant is likely to have a damaging effect on SMARCB1 function.

EBV positivity in mediastinal lesions has been considered to be strong evidence against a diagnosis of PMBL. EBV positivity is rarely seen in PMBL, and to the best of our knowledge, has been described in only 3 prior cases in the literature. In 1 study, 141 cases of PMBLs were compared with 916 of nonmediastinal lymphomas and only 2 cases were found to be positive for EBER in the 41 cases tested.2 In another study, comparing the efficacy of pembrolizumab in EBV-positive versus EBV-negative lymphomas, one of the 4 cases of PMBL showed EBER positivity.17

The results of mutational profiling in isolation rarely provide a definitive diagnosis, and often a clinical decision must be made before receiving the results of ancillary testing, as was true in this case. Nevertheless, we demonstrate that mutational tumor profiling may still give valuable information to aid in such challenging diagnoses and additionally, as in this case, may lead to the identification of targeted therapeutics that might be considered if conventional frontline chemotherapy proves ineffective. In conclusion, this case argues for consideration of PMBL in the differential diagnosis of an EBV-positive lymphoma with mediastinal involvement and also highlights the value of mutational profiling in the characterization of these rare tumors.


1. Swerdlow SH, Campo E, Harris NL, et al. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France: IARC; 2017:314–316.
2. Cazals-Hatem D, Lepage E, Brice P, et al. Primary mediastinal large B-cell lymphoma. A clinicopathologic study of 141 cases compared with 916 nonmediastinal large B-cell lymphomas, a GELA (Groupe d’Etude des Lymphomes de l’Adulte) study. Am J Surg Pathol. 1996;20:877–888.
3. Quintanilla-Martinez L, de Jong D, de Mascarel A, et al. Gray zones around diffuse large B cell lymphoma. Conclusions based on the workshop of the XIV meeting of the European Association for Hematopathology and the Society of Hematopathology in Bordeaux, France. J Hematop. 2009;2:211–236.
4. Copie-Bergman C, Gaulard P, Maouche-Chretien L, et al. The MAL gene is expressed in primary mediastinal large B-cell lymphoma. Blood. 1999;94:3567–3575.
5. Hoeller S, Zihler D, Zlobec I, et al. BOB.1, CD79a and cyclin E are the most appropriate markers to discriminate classical Hodgkin’s lymphoma from primary mediastinal large B-cell lymphoma. Histopathology. 2010;56:217–228.
6. Traverse-Glehen A, Pittaluga S, Gaulard P, et al. Mediastinal gray zone lymphoma: the missing link between classic Hodgkin’s lymphoma and mediastinal large B-cell lymphoma. Am J Surg Pathol. 2005;29:1411–1421.
7. Quintanilla-Martinez L, Fend F. Mediastinal gray zone lymphoma. Haematologica. 2011;96:496–499.
8. Glaser SL, Lin RJ, Stewart SL, et al. Epstein-Barr virus-associated Hodgkin’s disease: epidemiologic characteristics in international data. Int J Cancer. 1997;70:375–382.
9. Sarkozy C, Copie-Bergmand C, Damotte D, et al. Gray-zone lymphoma between cHL and large B-cell lymphoma: a histopathologic series from the LYSA. Am J Surg Pathol. 2019;43:341–351.
10. Pilichowska M, Pittaluga S, Ferry JA, et al. Clinicopathologic consensus study of gray zone lymphoma with features intermediate between DLBCL and classical HL. Blood Adv. 2017;1:2600–2609.
11. Camus V, Miloudi H, Taly A, et al. XPO1 in B cell hematological malignancies: from recurrent somatic mutations to targeted therapy. J Hematol Oncol. 2017;10:47.
12. Mahipal A, Malafa M. Importins and exportins as therapeutic targets in cancer. Pharmacol Ther. 2016;164:135–143.
13. Jardin F, Pujals A, Pelletier L, et al. Recurrent mutations of the exportin 1 gene (XPO1) and their impact on selective inhibitor of nuclear export compounds sensitivity in primary mediastinal B-cell lymphoma. Am J Hematol. 2016;91:923–930.
14. Scarpa A, Borgato L, Chilosi M, et al. Evidence of c-myc gene abnormalities in mediastinal large B-cell lymphoma of young adult age. Blood. 1991;78:780–788.
15. Bahram F, von der Lehr N, Cetinkaya C, et al. c-Myc hot spot mutations in lymphomas result in inefficient ubiquitination and decreased proteasome-mediated turnover. Blood. 2000;95:2104–2110.
16. Kalimuthu SN, Chetty R. Gene of the month: SMARCB1. J Clin Pathol. 2016;69:484–489.
17. Kim SJ, Hyeon J, Cho I, et al. Comparison of efficacy of pembrolizumab between Epstein-Barr virus positive and negative relapsed or refractory non-Hodgkin lymphomas. Cancer Res Treat. 2019;51:611–622.

primary mediastinal large B-cell lymphoma; EBV-positive; EBV-positive lymphoma; mediastinal lymphoma; XPO1 mutation; E571K; exportin 1

Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.