Castleman disease (CD) was first described by Benjamin Castleman in 1954 and consists of a heterogeneous group of disorders that share similar histomorphologic features and are clinically separated into two broad categories: unicentric Castleman disease (UCD) and multicentric Castleman disease (MCD). UCD patients present with a single enlarged lymph node or a tightly localized group of lymph nodes, while MCD patients typically present with multifocal lymphadenopathy associated with systemic symptoms including fatigue, fever, night sweats, and weight loss. UCD can further be divided into two or three morphologic variants: the hyaline vascular variant and plasma cell variant, with some texts including a mixed variant. Patients with the hyaline vascular variant of UCD typically lack systemic symptoms, while those with the plasmacytic variant often present with systemic symptoms similar to those in MCD, but in contrast to MCD, symptoms resolve with excision of the involved lymph node(s). Patients with UCD showing mixed hyaline vascular and plasmacytic morphology are not as well characterized, but their existence is considered evidence that the hyaline vascular variant and plasma cell variant are two ends of a spectrum rather than two distinct entities.
Similar to UCD, MCD can present with hyaline vascular-like morphology, plasmacytic morphology, or mixed morphology, although the plasmacytic morphology is more common in MCD. Unlike UCD, however, MCD is subtyped based on the presence or absence of an identifiable cause for the clinical and pathologic findings. The major known cause of MCD is infection by human herpesvirus-8 (HHV8), also known as Kaposi sarcoma associated herpesvirus (KSHV), typically occurring in immunocompromised patients. Many HHV8-associated MCD patients are also infected by the human immunodeficiency virus (HIV), but patients infected with HHV8 appear to have a similar clinical course regardless of HIV coinfection. MCD with no known etiology is termed idiopathic MCD (iMCD), which currently includes all non-HHV8 infected MCD patients. Within this category is a group of patients with a combination of findings including thrombocytopenia, anasarca, fever, fibrotic bone marrow, and organomegaly, which has come to be known as the TAFRO syndrome. Another entity called POEMS (polyneuropathy, organomegaly, endocrinopathy, monoclonal plasma cell disorder, skin changes) syndrome displays significant overlap with MCD and can be present in both HHV8+ and HHV8- MCD. It is important to note that several unrelated diseases can cause Castleman-like histopathology, such as autoimmune diseases, malignancy, and infections; it is currently recommended that such disorders be excluded before a diagnosis of CD is rendered (Fajgenbaum et al. 2017).
The incidence of CD is not well established but has been estimated to be approximately 7000 patients in the United States per year: approximately 77% UCD and 23% MCD (Munshi et al. 2015). Approximately 85% of UCD cases are of the hyaline vascular variant (Wang et al. 2016). Of the MCD cases, roughly one half to two-thirds of cases are HHV8 associated, with the remainder falling into the iMCD category (Liu et al. 2016). Approximately, 34% of MCD patients and 2% of UCD patients have associated POEMS syndrome (Dispenzieri et al. 2012). See Fig. 1 for an estimate of the relative incidences of specific CD subtypes.
There is a wide age distribution among CD patients, with young adults affected most frequently in UCD, and MCD displaying an older median age (fifth to sixth decade) (Casper 2005). HIV+ patients with MCD are younger than HIV- MCD patients, with an age range similar to the general HIV+ population.
- UCD: There is no sex predilection.
MCD: Males are slightly more frequently affected than females in some studies.
UCD affects in order of frequency: lymph nodes of the chest, neck, abdomen and retroperitoneum, followed by axilla, groin, and pelvis (Talat et al. 2012). Primary organ involvement is infrequent but most commonly seen in the spleen and parotid gland. MCD involves similar sites, although biopsies are more frequent in the axilla, neck, abdomen, and groin. By definition, MCD involves more than one site; however, a dominant focus of disease can be present with relatively mild involvement at other sites. In addition to lymphadenopathy, MCD patients often present with splenomegaly, hepatomegaly, pulmonary symptoms, bone marrow plasmacytosis (also reported in UCD), edema, ascites, and/or pleural effusions. TAFRO syndrome patients additionally display reticulin fibrosis in the bone marrow.
UCD: The treatment of choice for UCD is surgical excision of the affected lymph node(s)/tissue. In cases of unresectable disease, radiotherapy or therapy similar to that in MCD may be considered.
MCD: Optimal treatment strategies for patients with MCD are not clear due to the paucity of clinical trials and rarity of the disease. The antiviral ganciclovir is recommended in HHV8 positive patients, along with antiretroviral therapy for those concurrently infected with HIV. Since the discovery of the role for IL-6 in CD (see molecular features section below), anti-IL-6 antibodies have shown efficacy in treating a significant subset of patients with iMCD, and anti-gp80 (IL-6 co-receptor) antibodies have shown promising results. For both iMCD and HHV8-associated MCD patients, additional combinations of agents including anti-CD20 antibodies and steroids may be helpful. For patients refractory to other therapies and/or with severe symptoms, single agent cytotoxic chemotherapies such as etoposide or vinblastine, or multi-agent chemotherapy such as cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) may be utilized. Because therapies are constantly evolving as new evidence is collected, the most recent literature should be reviewed before determining the most appropriate treatment. Enrolment in a clinical trial is highly encouraged.
The majority of UCD patients are cured after complete excision, although mortality has been reported in a subset of patients. One study showed an overall 5-year survival of 91% (Dispenzieri et al. 2012), while another reported 10-year mortality of approximately 5% (~95% 10-year survival) associated with involvement of the following sites: retroperitoneum (11% mortality, n = 36), abdomen (2.4% mortality, n = 41), neck (1.8% mortality, n = 51), axilla (0% mortality, n = 13), groin (0% mortality, n = 4), and pelvis (0% mortality, n = 5) (Talat et al. 2012). The clinical course of patients with MCD is variable with a spectrum ranging from rapidly fatal to years of being relatively asymptomatic with only occasional flares. In one study, 5-year and 10-year overall survival in MCD was 65% and ~40%, respectively (Dispenzieri et al. 2012), while another reported 10-year mortality of approximately 38% (~62% 10-year survival) among MCD patients undergoing resective surgery involving the following anatomic sites: mediastinum (57.1% mortality, n = 7), groin (50% mortality, n = 2), abdomen (48% mortality, n = 25), pelvis (33.3% mortality, n = 3), retroperitoneum (16.7% mortality, n = 6), axilla (0% mortality, n = 4), and neck (0% mortality, n = 3) (Talat et al. 2012). There was no significant difference in mortality between MCD patients undergoing resective surgery and MCD patients without resective surgery in this study. Among MCD patients, those with POEMS syndrome without osteosclerotic bone lesions have been associated with worse prognosis (27% 5-year overall survival), while those with POEMS syndrome and the presence of osteosclerotic bone lesions had 5-year survival rates similar to UCD (Dispenzieri et al. 2012). MCD associated with HHV8 generally has a poor prognosis (median survival of approximately 1 year). For both UCD and MCD, hyaline vascular morphology is associated with a slightly better prognosis than plasma cell variants. HHV8+ MCD patients have increased risk for Kaposi sarcoma and primary effusion lymphoma, both of which are HHV8-associated malignancies. There is also a threefold increased risk for malignancy in patients with iMCD, though the temporal and causal relationship between iMCD and the various malignancies is unclear.
Involved lymph nodes are typically enlarged.
Hyaline Vascular Morphology
Sheets of plasma cells are seen in addition to prominent hyaline vascular morphology.
In HHV8-associated MCD, LANA1 antibody stains HHV8 infected cells, which can include B cells, endothelial cells, follicular dendritic cells, monocytes, and often plasmablasts residing in the mantle zones (Fig. 4b). Plasma cells in CD are usually polytypic by kappa and lambda immunohistochemistry or in situ hybridization but can be monotypic, even in the absence of POEMS syndrome (Figs. 4c, d). Of note, plasmablasts in HHV8-associated MCD are usually lambda restricted in protein expression but are polyclonal by PCR analysis. T cell and B cell markers reveal relatively normal distribution and immunophenotype of T cells and B cells, with decreased lymphocytes in regressed follicles. CD21, CD23, and CD35 highlight prominent follicular dendritic cells in regressed follicles.
CD has not been associated with specific recurrent molecular genetic aberrations, although the disease is difficult to study because of a combination of low prevalence and disease heterogeneity. Some studies suggest clonal proliferation of stromal or follicular dendritic cells in hyaline vascular CD, and occasional CD cases (more frequently PC variant) are positive for B cell clones by PCR analysis of the immunoglobulin heavy chain locus. The identification of HHV8 as a cause for CD in a significant subset of CD patients has provided a nidus for investigation into the pathogenesis of both HHV8+ and HHV8- CD. HHV8 is expert at promoting inflammatory cytokine release, and even encodes a homolog of human interleukin-6 (IL-6). Substantiating a pathogenic role for IL-6 in CD is the fact that many HHV8+ and HHV8- MCD patients respond to therapy inhibiting IL-6 signaling. A potential genetic cause of increased IL-6 was reported in a MCD patient with significantly elevated IL-6 levels and translocation t(7;14)(p22;q22) involving the IL-6 locus, raising the possibility that at least some iMCD patients may have genetic aberrations leading to CD (Nakamura et al. 1993). Approximately two-thirds of iMCD patients do not respond to IL-6; however, highlighting the need to search for other mechanisms of pathogenesis will hopefully point to new therapies.
Many disease entities can display variable aspects of Castleman-like features, including autoimmune/inflammatory disorders, infections, and neoplastic conditions (Fajgenbaum et al. 2014). Many of these entities can display some of the features seen in hyaline vascular change, although typically not as prominently as in true CD. Any disease leading to increased plasma cells can mimic the plasmacytic morphology in CD, including plasma cell neoplasms, B cell neoplasms with plasmacytic differentiation, IgG4 related disease, autoimmune disorders such as rheumatoid arthritis, and infectious organisms such as syphilis. Most of these entities, if present at the time Castleman-like features are discovered, should exclude a diagnosis of CD. Some entities have been reported to co-occur with CD (especially autoimmune diseases and POEMS syndrome) and thorough evaluation of clinical, laboratory, and histologic features utilizing diagnostic criteria recently published (Fajgenbaum et al. 2017) will be required to determine whether a concurrent diagnosis of CD is warranted and CD targeted therapy is justified.
References and Further Reading
- Fajgenbaum, D. C., Uldrick, T. S., Bagg, A., Frank, D., Wu, D., Srkalovic, G., Simpson, D., Liu, A. Y., Menke, D., Chandrakasan, S., Lechowicz, M. J., Wong, R. S., Pierson, S., Paessler, M., Rossi, J. F., Ide, M., Ruth, J., Croglio, M., Suarez, A., Krymskaya, V., Chadburn, A., Colleoni, G., Nasta, S., Jayanthan, R., Nabel, C. S., Casper, C., Dispenzieri, A., Fosså, A., Kelleher, D., Kurzrock, R., Voorhees, P., Dogan, A., Yoshizaki, K., van Rhee, F., Oksenhendler, E., Jaffe, E. S., Elenitoba-Johnson, K. S., & Lim, M. S. (2017). International, evidence-based consensus criteria for HHV-8-negative/idiopathic multicentric Castleman disease. Blood, 129(12), 1646–1657.CrossRefPubMedPubMedCentralGoogle Scholar
- Liu, A. Y., Nabel, C. S., Finkelman, B. S., Ruth, J. R., Kurzrock, R., van Rhee, F., Krymskaya, V. P., Kelleher, D., Rubenstein, A. H., & Fajgenbaum, D. C. (2016). Idiopathic multicentric Castleman’s disease: A systematic literature review. Lancet Haematol, 3(4), e163–e175.CrossRefPubMedGoogle Scholar