Pathology and classification of aggressive mature B-cell lymphomas.

Abstract

The aggressive mature B-cell lymphomas are a heterogeneous group of tumors with different biological and pathological characteristics that are associated with a broad spectrum of clinical manifestations. Some of the subtypes are relatively common whereas others are less frequent and occur in particular subsets of patients. The response to therapy and the outcome are also diverse among entities. Although current therapies may cure large proportion of patients, still around 30% of them develop an incurable disease. The recent genetic and molecular studies are increasing the understanding of the mechanisms underlying the clinical and biological diversity of these tumors and are providing valuable information for new therapeutic opportunities. The recent update of the World Health Organization (WHO) classification of lymphoid neoplasms includes different entities of aggressive mature B-cell lymphomas with well-defined diagnostic criteria (Table 1).1 The most common subtype accounting for approximately 80% of all these neoplasms is diffuse large B-cell lymphoma, not otherwise specified (DLBCL, NOS), a category that includes tumors, which cannot be classified in any of the other more specific entities. The other diseases are less common but have distinctive features. T-cell/histiocyte–rich large B-cell lymphoma (THRLBCL) is a subtype in which the tumor cells are overrun by an exuberant stromal response. Some large B-cell lymphomas (LBCLs) are originated in particular topographic sites suggesting their relationship to specific cell subsets of these sites or the influence of the microenvironment in their pathogenesis. A group of LBCL are associated with Epstein-Barr virus (EBV) infection and occur in patients that may have a certain immunological impairment in controlling the viral infection. Most LBCL have a mature B-cell phenotype, but some tumors acquire a terminal B-cell differentiation phenotype loosing mature B-cell markers and expressing proteins related to plasma cell differentiation. Some of these tumors are associated with HHV8 and/or EBV infection. Burkitt lymphoma (BL) is a well-defined aggressive neoplasm that can be cured in most patients but certain aspects regarding the presence of MYC translocation in all cases still remain controversial. Finally, the updated WHO classification has revised the concept of aggressive B-cell lymphomas with features intermediate between DLBCL and BL and tumors with MYC and BCL2 and/or BCL6 translocations. This review will address the pathology of all these entities emphasizing the new aspects included in the updated WHO classification (Table 1). Diffuse large B-cell lymphoma, NOS is clinically and biologically very heterogeneous. One of the major advances understanding its diversity was the recognition of 2 molecular subtypes based on their gene expression profiling (GEP). These 2 subtypes are related to a different cell of origin either in normal germinal center B-cells (GCB) or in activated B-cells (ABC).2 In addition to the GEP, these 2 molecular subtypes differ in the activation of different molecular pathways, profile of chromosomal alterations, and somatic mutations. These biological differences translate into different outcome of the patients with most of the studies showing worse prognosis for ABC than GCB-DLBCL. Germinal center B-cell tumors rely preferentially on the activation of the PI3K pathway and BCL6 overexpression whereas ABC tumors have a constitutive activation of the nuclear factor–κB pathway through different mechanisms. Genetic alterations in GCB include BCL2 and BCL6 rearrangements and mutations in histone modifiers (EZH2, KMT2D, and CREBB) and GNA13 whereas ABC tumors have frequent mutations in MYD88 and CD79a.2, 3 The distinctive molecular pathogenesis of these tumors is expected to lead to more precise therapeutic strategies targeting the specific molecular alterations. For these reasons, the recognition of these 2 molecular subtypes by the pathologists is now recommended in the clinical practice. The considered gold standard for this molecular classification is the differential GEP detected by microarrays. However, this approach requires frozen tissues and it is not currently used in the clinics. The initial alternative approach to identify these molecular subtypes in routine practice has been the use of a small number of distinctive biomarkers by immunohistochemistry (IHC). Different algorithms have been proposed with relative good concordance with GEP classification in around 80–90% of the cases.4 However, they have also limitations related to standardization and reproducibility among pathologists. In addition, they may lack precision to identify the ABC subgroup since most IHC algorithms call for a dual classification between GCB and non-GCB tumors. However, there is a small subgroup of cases (10%) that cannot be classified by GEP (unclassified) and IHC algorithms force their inclusion in one of the 2 subtypes.4 A recent developed assay based on GEP of 20 genes (Lymph2Cx) based on Nanostring technology reliably identifies the 2 molecular DLBCL subtypes on RNA extracted from formalin-fixed and paraffin-embedded tissues with excellent concordant result when compared to the microarrays GEP.5 The assay is also highly reproducible among different laboratories and confirms the different clinical outcome of the 2 subtypes. The WHO classification recommends the identification of the cell of origin of all DLBCL, NOS and considers the IHC acceptable. However, gene expression–based assays may be a promising more precise methodology.1 T-cell/histiocyte–rich large B-cell lymphoma is an aggressive and relatively uncommon lymphoma histologically characterized by scattered, large, atypical B cells embedded in a microenvironment with abundant T cells and histiocytes. Patients are frequently diagnosed at higher clinical stages with liver, spleen, and bone marrow involvement. The aggressive behavior of these tumors has been attributed to the advanced stage and high International Prognostic Index (IPI) that the patients usually have at diagnosis. Some cases with identical morphology and phenotype may arise from nodular lymphocyte predominant Hodgkin lymphoma (NLPHL) and are considered as a THRLBCL-like transformation from NLPHL.1 The term “like” is used due to still some uncertainties regarding the relationship between de novo THRLBCL and transformed cases. However, recent studies have shown similarities in the GEP and genetic alterations of the tumor cells in THRLBCL and NLPHL supporting the idea of a closer relationship.1 Several LBCLs are classified mainly, but not exclusively, based on the topographic site of their origin (Table 1). Although these entities do not have a specific phenotype or molecular hallmark, they have particular features. Primary mediastinal large B-cell lymphoma usually presents in young females with a large mediastinal mass that may infiltrate surrounding structures.6 The tumor may affect regional lymph nodes but when it disseminates outside of the thorax, it tends to involve solid organs. Histologically, the tumor cells may have an abundant clear cytoplasm and fibrotic bands are common. Thymic remnants may be seen, as the tumor seems to originate from mature thymic B cells. Phenotypically, the tumor cells may express PDL1, PDL2, CD30, and CD23, and are usually negative for IG and HLA class I and II.6 Gene expression profiling has identified a specific signature that may be useful to differentiate primary mediastinal large B-cell lymphoma from DLBCL, NOS involving the mediastinum or to recognize these tumors in locations outside of the thorax.7 The genetic profile differs from DLBCL, NOS with frequent translocations inactivating class II major histocompatibility complex transactivator (CIITA), activation of the nuclear factor–κB, and JAK/STAT pathways due to several genetic alterations in regulatory genes such as TNFAIP3 and SOCS1, and PTPN1, respectively.6 Other large B-cell entities related to specific sites are Primary DLBCL of the central nervous system (CNS), Primary cutaneous DLBCL, leg type, and intravascular LBCL. All these tumors are uncommon. The Primary DLBCL of the CNS should be distinguished from other brain LBCL occurring in the context of immunodeficiency. Primary DLBCL of the CNS has frequent deletions and negative expression of the HLA genes and mutations of MYD88, CARD11, and CD79, which may have therapeutic implications. A certain relationship with lymphomas of the testis exists since rare cases with extracranial dissemination may involve the testes, and vice versa, some lymphomas of the testes may relapse in the brain. Primary cutaneous DLBCL, leg type is an aggressive tumor presenting in the skin of the lower extremities although it may also occur in other cutaneous locations. It has an ABC phenotype and strong expression of BCL2. Intravascular LBCL is very aggressive, usually with systemic intravascular dissemination at diagnosis. Some improvement in the outcome has been reported with immunochemotherapy. Random skin or transbronchial biopsies, even when these sites appear unaffected, have been proposed for an early diagnosis and treatment of this tumor. A group of LBCL is associated with EBV infection and has different clinical and pathological features. They occur in apparently immunocompetent patients but appear to have predisposing conditions, which decrease EBV immune surveillance. The most common is the EBV positive DLBCL, NOS.1 The term NOS refers to the fact that there are other LBCL EBV positive with distinct clinical and pathologic features. This entity was previously named EBV-positive DLBCL of “the elderly” because of the higher frequency in patients older than 50 years. However, recent studies have recognized this tumor in younger patients.8 Histologically, the large atypical B cells may have Reed-Sternberg–like features and are accompanied by a variable amount of inflammatory cells including CD8-positive T cells. Areas of necrosis are frequent. Cases in young patients resemble THCRLBCL with abundant T cells. The histological features may suggest the diagnosis and trigger the study of EBV by in situ hybridization. The virus must be present in most if not all tumor cells and usually express latency II or III antigens. The prognosis is significantly worse in elderly than young patients. Other LBCL EBV positive are DLBCL associated with chronic inflammation and lymphomatoid granulomatosis.1 The first occur in the context of long lasting inflammation, particularly in confined sites, such pleura (pyothorax), bone or joints. Lymphomatoid granulomatosis is an angiocentric and angiodestructive lymphoproliferative disease that involves extranodal sites. The lesion is composed of a variable number of EBV-infected large atypical cells intermingled with reactive T cells. The clinical behavior has been related to the proportion of tumor cells. Large B-cell lymphoma with terminal B-cell differentiation is a heterogeneous group of aggressive lymphomas characterized by a immunoblastic or plasmablastic morphology and the acquisition of a plasma cell phenotype with downregulation of mature B-cell markers (CD20 and PAX5) and expression of plasma cell antigens (BLIMP1, CD38, and CD138) (Table 1).9 Plasmablastic lymphoma mainly presents in extranodal sites, particularly of the head and neck, but the disease is usually disseminated at diagnosis.10 Some cases have bone involvement and features overlapping with plasma cell myeloma. The differential diagnosis should consider the clinical context of plasmablastic lymphoma in immunodeficient patients and EBV infection. Epstein-Barr virus is positive with latency I in approximately 70% of the cases. MYC translocations are detected in 50%, usually with an IG partner. The clinical course is very aggressive with a median overall survival of 6 to 11 months. HHV8 induces a spectrum of lymphoproliferative diseases with 2 subtypes of lymphomas, Primary effusion lymphoma (PEL) and HHV8+ DLBCL, NOS.11 PEL occurs mainly in cavities without a tumor mass. Some patients subsequently develop a solid tumor with similar pathologic features, and finally, rare cases initially present as solid tumor masses indistinguishable from PEL and have been termed extracavitary PEL. Most cases are EBV positive with latency I although some cases are negative, mainly in the elderly and in areas of endemic HHV8. HHV8+ DLBCL, NOS has similar morphology and phenotype but is EBV negative an expresses IgM, lambda. IGHV are unmutated. These tumors usually arise in HHV8+ multicentric Castleman disease but also in the absence of this disease in immunodeficient patients. Anaplastic lymphoma kinase-positive LBCL is a very aggressive tumor occurring in immunocompetent young adults.9 The tumors present with generalized lymphadenopathy and advanced stage. The plasmablastic phenotype seems induced by the anaplastic lymphoma kinase activation due to translocations mainly with Clathrin (CLTC) although other partners are also found. STAT3 seems an important downstream target. Burkitt lymphoma is a well-defined entity genetically characterized by MYC rearrangements. In the last years, Next generation sequencing (NGS) have revealed the profile of somatic mutations with frequent mutations in TCF3 and ID3 that are very uncommon in DLBCL. CCND3 is also mutated in 30% of the tumors. An unresolved issue is the existence of true BL without MYC translocation. Recent studies have identified cases with similar morphology and phenotype but negative for MYC rearrangements that have 11q alterations with proximal gains and telomeric losses. These cases have been named Burkitt-like lymphoma with 11q aberrations and have more frequent nodal presentation and complex karyotypes.1 The information on these cases is still limited. The updated 2016 WHO classification has considered high-grade B-cell lymphomas (HGBLs) as a provisional category that emphasizes the relevance of genetic information for its recognition.1 All LBCL with MYC and BCL2 and/or BCL6 rearrangements are included in the category of HGBL with MYC and BCL2 and/or BCL6 rearrangements. These cases have been also named “double hit” HGBL (HGBL-DH). The relevant criteria for the diagnosis are the genetic alterations independently of the morphology of the tumor that may be DLBCL, blastoid, or with features intermediate between DLBCL and BL. The specific morphology should be noted because it may have prognostic impact.12 Cases with blastoid morphology or with features intermediate between DLBCL and BL without translocations are considered HGBL, NOS. Diffuse large B-cell lymphoma with high expression of MYC and BCL2 protein without genetic alterations have been called “dual-expressors.”1 This dual-expressors is considered an adverse prognostic factor, but these tumors are not included in the HGBL category since the outcome does not seem so adverse, and the biological consequences of having a genetic alteration may also be different. The expansion of fluorescence in situ hybridization (FISH) studies in aggressive lymphomas is showing that the clinical impact of the DH seems more variable than suggested in the initial studies but the reasons for this diversity are not well known. The biology of HGBL-DH is complex and not yet well understood.13 The presence of MYC rearrangements is the determinant factor, but there are several additional elements that seem to modulate their biological relevance (Table 2).12, 14 The association with BCL2 translocations usually confers an adverse prognosis, but the role of BCL6 is still controversial.14, 15 Other modulators of MYC rearrangements are the translocated partner. IG-MYC rearrangements are associated with higher levels of MYC expression than non-IG partners, and this may explain the worse prognosis of the former. The MYC and BCL2 protein expression levels may be also different in cases with translocation probably due to additional phenomena such as amplification of the translocated allele, mutations of MYC, or others. The cell context may also be a modifier with DLBCL cases having a better outcome than tumors with blastoid or DLBCL/BL intermediate morphology. The clinical context of the patients seems also to be important since patients with high-risk clinical features have more adverse evolution than patients with low-risk factors.15 Therefore, further studies evaluating all these aspects are needed to better understand the significance of this new category. Independently of all these variables, the recognition of HGBL-DH is clinically relevant because most of the tumors have a very aggressive behavior and standard DLBCL treatments are considered insufficient. The importance of recognizing HCBL-DH challenges the current diagnostic practice of aggressive B-cell lymphomas since the use of FISH in all cases may confer a workload and economic burden difficult to be confronted in most centers. Although no consensus still exists, a selection of cases for FISH analysis based on morphological and phenotypic criteria seems an acceptable compromise.1 Virtually, all HGBL-DH with MYC and BCL2 translocations are of the GCB subtype and cases with these rearrangements usually have high protein expression. Therefore, a reasonable strategy would be to perform FISH in all GCB-DLBCL with high MYC and BCL2 protein expression.15 In addition, FISH should be performed in cases with blastoid or intermediate DLBCL/BL morphology given the high frequency of the translocations (approximately 50%) in these tumors. This strategy would miss some DH cases with BCL6 rearrangements that are ABC-DLBCL and DH cases with low protein expression, but the clinical significance of both situations is still controversial.14, 15 Elías Campo received research funding from Gilead, has been consultant for Celgene and Gilead is author in a patent LLMPP patent Method for selecting and treating lymphoma types PCT/US14/64161.