Epstein–Barr virus (EBV) is associated with some disease entities of malignant lymphomas, including Burkitt lymphoma, Hodgkin lymphoma, immunodeficiency-associated lymphoproliferative disorders (LPD), and a part of diffuse large B-cell lymphoma. We have recently identified a series of elderly patients with EBV-associated (or EBV+) B-cell LPD (B-LPD) showing similarities in many respects to immunodeficiency-associated LPD, although no evidence of underlying immunodeficiency was found. Therefore, the nosological category of senile or age-related EBV+ B-LPD has been proposed for those patients. A larger series of patients with this disease revealed that the relative ratios of such EBV+ B-LPD to all diffuse large B-cell lymphoma cases were higher with increasing with age, reaching a peak (20–30%) at ≥90 years of age, with a median of 71 years, providing additional evidence for our assertion that this disease may be related to immunological deterioration as a result of the aging process. This new disease entity is characterized pathologically by centroblasts, immunoblasts, and Hodgkin and Reed–Sternberg-like giant cells with a varying degree of reactive components, often posing therapeutic and diagnostic problems for hematologists and pathologists, respectively. The aim of the present review is to briefly summarize the overall clinicopathological profile of this newly recognized age-related (also called ‘senile’) EBV+ B-LPD and EBV+ Hodgkin lymphoma for a practical diagnostic approach. (Cancer Sci 2008; 99: 1085–1091)
Over the past few decades, human malignancies associated with EBV, including various types of malignant lymphomas and carcinomas, have been well documented and our knowledge of them has now greatly expanded. However, the epidemiology of EBV-associated or -positive (EBV+) human diseases appears to be complex, and the true contribution of EBV to their pathogenesis remains to be elucidated.
According to the World Health Organization classification of lymphoid tumors,(1,2) some disease entities such as endemic Burkitt lymphoma, lymphomatoid granulomatosis, and extranodal natural killer T-cell lymphoma of the nasal type show a high prevalence of EBV. It is also clearly involved in the pathogenesis of the majority of B-cell lymphomas arising in iatrogenically or congenitally immunosuppressed patients. However, the association of EBV is heterogeneous in the categories of DLBCL, peripheral T-cell lymphoma, and cHL. Over the last decade, accumulating evidence has supported the assertion that EBV positivity is an independent adverse factor for survival among elderly patients with DLBCL and cHL.(3,4)
We have recently highlighted the profile of senile or age-related EBV+ B-LPD, showing similarities in many respects to immunodeficiency-associated B-LPD exemplified by a marked propensity to involve extranodal sites and a morphological spectrum ranging from the precursor polymorphous proliferation of lymphoid cells to diverse types of lymphomas, although no evidence of underlying immunodeficiency was found.(5,6) It is speculated that this disease is related to immunological deterioration as a result of the aging process, that is, senescence in immunity. The recognition of age-related EBV+ B-LPD prompted us to reconsider its relationship with EBV− DLBCL and EBV+ HL, because differential diagnosis among these tumors was occasionally impossible to resolve by histological examination alone. In the present review we discuss the overall clinicopathological profile of age-related B-LPD, focusing on its relationship with and areas of differential diagnosis compared to EBV− DLBCL and EBV+ HL.
Age-related EBV+ B-LPD
Age-related EBV+ B-LPD is now defined as an EBV+ clonal B-cell lymphoid proliferation or EBV+-DLBCL,(4) which occurs in patients over the age of 40 years in the absence of any known immunodeficiency and without an underlying T-cell lymphoma.(7) Cases less than 40 years old were excluded because we could not deny the possibility that they may be associated with any primary immune disorder or chronic active EBV infection. Of note, these cases should be considered to represent overt lymphomas, with the diagnosis in most cases equivalent to that of EBV+ DLBCL occurring in someone over 40 years of age.
In 2003, we documented 22 cases named senile EBV-associated B-LPD arising in elderly patients aged >60 years without predisposing immunodeficiencies, suggesting that this disease has a relationship with immunological deterioration as a result of the aging process.(5,6) We subsequently examined the presence or absence of EBV in 1792 large B-LPD cases by Epstein–Barr virus-encoded small RNA in situ hybridization retrieved from the files of six collaborating institutions in Japan during the period from January 1990 to December 2004 (Table 1).(7) This survey revealed that 156 cases (8.7%) harbored EBV without underlying immunodeficiency-related diseases. In this larger series, 149 (8.3%) of the patients were more than 40 years of age, with a bimodal age distribution consisting of an incidence peak in the 10–19-year range and a second peak in older adults aged 70–79 years (Fig. 1a). The positive percentage of each age group for all cases examined became higher in parallel with the elder patient populations (>40 years), showing the highest peak at >90 years (Fig. 1b). These data provide additional evidence that EBV+ B-LPD without predisposing immunodeficiency occurs mainly in elderly patients, although seven patients were found to be less than 40 years of age. In taking these rare cases into consideration, the term ‘age-related’ may be more appropriate than ‘senile’ in further understanding the overall age distribution of EBV+ B-LPD without predisposing immunodeficiency. Currently, age-related EBV+ B-LPD accounts for 8–10% of DLBCL among Eastern Asian patients without a documented predisposing immunodeficiency;(4,7,8) there is little available data regarding the frequency of DLBCL patients harboring EBV in Western countries.
Table 1. Age distribution of the study population and Epstein–Barr virus (EBV)-positive cases, and the pathological diagnosis of EBV+ cases
Age at diagnosis (years)
No. all cases examined
No. EBV+ cases (% for all cases)
Autoimmune disease-associated LPD
Secondary lymphoma with prior chemotherapy
Post- transplant LPD
Pyothorax- associated lymphoma
Others without predisposing immunodeficiency (% for all cases)
HIV, human immunodeficiency virus; LPD, lymphoproliferative disorder.
Clinical findings. The clinical features of age-related EBV+ B-LPD at presentation are variable, and are briefly summarized in Table 2. Among 96 patients with available clinical data, the majority (80%) presented with lymphadenopathy with or without extranodal involvement (in 49 and 31% of patients, respectively).(7) The main sites of extranodal involvement were the skin, lung, tonsil, and stomach.(7,9) These patients with age-related EBV+ B-LPD also showed a higher age distribution and more aggressive clinical features or parameters than EBV− DLBCL: 44% with a performance status greater than 1, 58% with a serum lactate dehydrogenase level higher than normal, 49% with B symptoms, and higher rates of involvement of the skin and lung. Overall survival rate was thus significantly lower in the age-related EBV+ group than in those with DLBCL. Besides our data, Park et al. identified 34 EBV+ cases (9%) among 380 Korean patients with DLBCL, in which comparative data between EBV+ and EBV− groups were generally in agreement with ours with statistically significant differences.(4) Yoshino et al. also documented that 4 of 50 (8%) primary gastric DLBCL cases were positive for EBV, occupying half of eight patients who did not initially respond or who relapsed after chemoradiotherapy.(9)
Table 2. Clinical characteristics of patients with diffuse large B-cell lymphoma (DLBCL) according to the Epstein–Barr virus (EBV)-encoded small RNA (EBER) status
Morphology. Age-related EBV+ B-LPD showed a wide morphological spectrum similar to what is seen in post-transplant LPD.(5–7) The architecture of the involved lymph nodes or other underlying tissues was disrupted in contrast to infectious mononucleosis. Although no longer considered to be of clinical importance, including clinical features and prognosis,(7) cases were divided morphologically into polymorphous and large-cell lymphoma subtypes, both of which included many large transformed cells and immunoblasts, and HRS-like giant cells (Fig. 2a). The polymorphous subtype exhibited a broad range of B-cell maturation and a variable component of reactive elements such as small lymphocytes, plasma cells, histiocytes, and epithelioid cells. In the large-cell lymphoma subtype, most of the cells appeared to be transformed. Both subtypes sometimes demonstrated large areas of geographic necrosis, and the histology was often variable from area to area, indicating a continuous spectrum between the two subtypes. Indeed, large-cell lymphoma cases showed areas of polymorphous-subtype lesions in the same or other tissues. Both subtypes are classified as DLBCL according to the current World Health Organization classification.
Immunophenotype. According to the disease definition, EBV+ B-cells were usually positive for CD20 or CD79a (Fig. 2b,c), and showed light chain restriction. Cases bearing cytopathological findings of an immunoblastic lymphoma to plasmacytoid features or a plasmablastic lymphoma lacked CD20 expression and had detectable cytoplasmic immunoglobulin. LMP1 and EBNA2 were positive on large atypical cells in 94 and 28% of the tested cases, respectively.(5–7) The cells were stained for CD30 in 75% of cases, but not CD15. Also, a comparison of adjacent sections often disclosed an overlapping staining pattern of LMP1 and CD30. These positive ratios of LMP1, EBNA2, and CD30 are mostly in keeping with those documented by Kuze et al. in their series of 13 cases.(8)
Genotype. Clonality of the immunoglobulin genes and EBV was usually detected by molecular genetic techniques in both morphological subtypes.(5,7) These studies were helpful for distinguishing polymorphous age-related B-LPD from infectious mononucleosis and EBV+ reactive B-LPD in the middle-aged or elderly patients, the atypical clinical findings of the latter having been reported recently by Kojima et al.(10) The cytogenetic features of this disorder and whether there are oncogenic abnormalities remains to be elucidated.
Treatment, prognosis, and predictive factors. Treatment with chemotherapeutic regimens containing anthracycline achieved complete remission in 66% of patients with age-related B-LPD with initial therapy, suggesting that this disease is significantly more refractory to initial chemotherapy compared with EBV− DLBCL.(7)
Age-related EBV+ B-LPD has strikingly inferior survival to EBV− DLBCL (Fig. 3). Most patients with age-related EBV+ B-LPD showed an aggressive clinical course with a median survival of approximately 2 years.(7) Neither the international prognosis index (IPI) score nor the histopathological subtype affected the prognosis. The presence of B symptoms and an age greater than 70 years appeared to be reliable prognostic factors in univariate and multivariate analyses. A prognostic model using these two variables defined three risk groups: low risk (no adverse factors), intermediate risk (one factor), and high risk (two factors). They showed median overall survival times of 56, 25, and 9 months, respectively.(7)
Epstein–Barr virus-associated HL
Hodgkin lymphoma is a unique clinicopathological disorder in which HRS cells constitute only a minority (1–2%) of the total tumor mass. In the World Health Organization classification,(1,2) HL is classified into two distinct entities: the nodular lymphocyte-predominant HL, and cHL, which can be further divided in the morphological subtypes of NSHL, MCHL, lymphocyte-depleted HL, and lymphocyte-rich HL. HL is a heterogeneous condition that most probably comprises more than one etiological entity. Recently, a few authors shed some light on the distinctive features of EBV+ HL by undertaking epidemiological studies,(3,11,12) suggesting that EBV-associated and -non-associated cases may represent two distinct etiological entities. This implies that the biological properties, such as an EBV association, may precede the morphological evaluation for a greater understanding of the complete clinicopathological profiles of HL. However, recent observations indicate that HL and non-HL may be more closely related than previously believed, achieving a general consensus regarding the B-cell derivation of HRS cells in most HL cases.(12–14) A biological overlap between HL and non-HL may be indicated based on the fact that the prototypic cytomorphology of HRS cells is shared by transformed large B lymphocytes, especially those harboring EBV. Biological interfaces are thus assumed between EBV+ HL and EBV+ B-cell lymphoma.
Classical Hodgkin lymphoma in Japan. We recently documented the clinicopathological features of 324 cHL patients, which were retrieved from the slide files of Aichi Cancer Center during the period from April 1982 and March 2005 with a histological review.(15) This series included 132 patients with NSHL (median age, 31 years; range, 12–84 years; male-to-female ratio, 1.36) and 157 with MCHL (median age, 57 years; range, 4–89 years; male-to-female ratio, 3.36). The age distribution of the NSHL patients was bimodal, with the first peak in the 20–29-year age range and the second peak in the 60–69-year age range, the overall profile of which was consistent with most marked bimodalities for NSHL occurring in Western countries.(3,16) The MCHL patients showed a peak in older adults aged 60–69 years, providing evidence that the age-specific incidence rate for MCHL is more comparable across races.
Epstein–Barr virus status in HL patients. It is well known that EBV is more often associated with MCHL than NSHL, and that HL in children and older adults is more likely to be EBV-associated than in young adults.(16–24) Indeed, in our series, EBV presence was detected in 149 (47%) of 314 patients with 74% of MCHL and 14% of NSHL cases being EBV-associated (P < 0.0001 by the χ2-test).(15) The clinicopathological findings of EBV+ and EBV− cHL groups are summarized briefly in Table 3 with their age-specific distributions in Fig. 4. A bimodal curve with an incidence peak in young adulthood (20–29 years) and a second peak in older adults (60–69 years) was evident for EBV− cHL. In contrast, the curve for EBV-associated cHL had a peak in the older adult age group (70–79 years). There was also a male excess within EBV+ HL (male-to-female ratio, 107:30) but not EBV− HL (male-to-female ratio, 95:60) (P = 0.0019 by the χ2-test). As is described in Western countries, the present data for Japanese patients indicates that the proportion of EBV+ cHL cases is generally higher in childhood (especially <10 years of age) and in older adults (>40 years) than in younger adults (10–39 years).(25) Comparison of these age-specific distribution patterns for EBV+ and EBV− cHL may provide additional support for the hypothesis that these form two distinct etiological entities.(3,16,17)
Table 3. Clinical characteristics of patients with classical Hodgkin lymphoma according to Epstein–Barr virus-encoded small RNA (EBER) status
P-value by χ2-test for independence, or Fisher's exact probability test, EBER+ classical Hodgkin lymphoma versus EBV− classical Hodgkin lymphoma. LDH, lactate dehydrogenase; PS, performance status.
Median age (years) (range)
Age > 45 years
Age > 60 years
PS > 1
Extranodal > 1site
LDH > normal
Median survival (months)
Diagnostic accuracy of EBV+ cHL in the elderly. Although accuracy in the diagnosis of HL is generally good, certain subgroups of cases and the elderly continue to present diagnostic difficulties. Jarrett et al.(16) and Glaser et al.(25) recently indicated that, in their expert reviews of HL histology, the diagnosis of HL in the elderly and classification of MCHL was less reliable, although the diagnosis of NSHL had a very high degree of reliability. Positive predictive values of the original diagnosis (the percentage of referral diagnoses confirmed on review) differed for the different histological subtypes, being good for NSHL (88–95%) but poorer for MCHL (58–59%). Interestingly, Jarrett et al. indicated that diagnostic reliability was significantly lower (P < 0.001) for the older adult age group (>50 years), with more misdiagnoses, subtype changes, and uncertainties, in comparison with the other age groups (16–34 years and 35–49 years).(16) This combination of MCHL and elderly patients may in part represent diagnostic difficulties between EBV+ HL and age-related EBV+ B-LPD in addition to the increased incidence of the latter in older patients.
Prognostic impact of EBV on cHL patients. The clinicopathological significance of EBV as a prognosticator in HL patients is still controversial, and has been reviewed concisely by Gandhi et al.(3) Interestingly, two population-based studies without selection bias recently documented a marked survival disadvantage in older EBV+ cHL patients compared with EBV− cHL cases, contrasted with no effect of the EBV status on the clinical outcome of HL patients enrolled selectively in clinical trails with a tendency of a relatively younger age distribution. The studies of Clarke et al.(20) and Stark et al.(26) demonstrated that EBV positivity in HRS cells was correlated with significantly poorer survival in patients aged 45–79 years of age and those aged 60 years and above, respectively. Starke et al. further found that EBV status had an adverse influence on clinical outcomes especially in elderly patients (70 years of age and above) compared with a younger elderly group (60–69 years of age).(26) Enblad et al. also reported that EBV+ cHL patients were more likely to be older, have more B symptoms, and advanced disease, and showed a tendency toward reduced survival compared with EBV− cases.(27) However, some authors reported that EBV+ HL is associated with improved disease-free survival in young adults.(28,29) As the interpretation for this age-related influence of EBV on the clinical outcome of HL patients, Ganhdi et al. stated the following: ‘it may be that the differential impact that EBV HRS status has on outcome between young adult and older adult age groups is either a consequence of biologically distinct diseases, or alternatively a decline in EBV-specific cellular immunity with age’.(3) More recently, Jarrett et al. also indicated that, among patients 50 years or older with cHL, EBV positivity was associated with a significantly poorer outcome, suggesting that an impaired immune status may contribute to the development of EBV+ cHL in older patients.(17) These standpoints were interesting to further our understanding of the background of the pathogenesis of both EBV+ cHL and age-related EBV+ B-LPD. However, in our series of these diseases, we could not identify any statistical difference in prognosis between EBV+ and EBV− groups among the elderly patients >60 years with cHL, which were significantly superior to that of age-related EBV+ B-LPD (N. Asano, K. Yamamoto, S. Nakamura, unpublished data, 2008).
Epstein–Barr virus-positive cHL and age-related EBV+ LPD as B-cell neoplasms
The exact origin of HRS cells has long been a topic of intense debate.(1) The phenotype of HRS cells has no obvious normal cellular counterpart. However, elegant studies by Küppers and colleagues have produced compelling evidence to suggest that these malignant cells are predominantly lymphocytes of B-cell lineage.(30,31) Biological interfaces or overlaps are thus identified between HL and diverse subtypes of B-cell lymphoma. Because the range of histological appearances of both EBV+ HL and age-related EBV+ B-LPD is so wide, it is not surprising that the distinction between them may be difficult or impossible in occasional cases. The presence of an appropriate background of variable numbers of reactive components such as small lymphocytes, eosinophils, histiocytes, and plasma cells is indicated to favor the diagnosis of HL. However, the degree of the reactive infiltrate was often shared by age-related EBV+ B-LPD, especially in the polymorphic subtype, although necrotic lesions were less commonly found in EBV+ cHL. Because the differential diagnosis may sometimes be impossible to resolve by histological examination alone, the key to the distinction between these two EBV+ diseases now depends on recognition of the degree of expression of B-cell markers such as CD20 and CD79a on tumor cells. Cases were categorized as age-related EBV+ B-LPD if 50% or more of the EBV+ tumor cells with CD20, CD79a, or light chain restriction were observed, or cHL if less than 50% (N. Asano, K. Yamamoto, S. Nakamura, unpublished data, 2008).(7) Expression of CD15 is also useful for distinguishing two diseases: age-related EBV+ B-LPD is negative for CD15, whereas cHL is positive.(7) Clinically, EBV+ B-LPD often show extranodal involvement,(5,7) whereas cHL does not usually.
Single-cell analysis of HRS cells showed that immunoglobulin gene rearrangements took place with ‘crippling’ somatic mutations in 25% of cHL cases, resulting in the absence of the immunoglobulin gene transcription.(32,33) These ‘crippling’ mutations are indicated to be normally incompatible with B-cell survival, suggesting that a transforming event, such as EBV infection, has taken place. Interestingly, Timms et al. recently documented similar ‘crippling’ immunoglobulin mutations of 11 post-transplant LPD cases,(34) which is typically driven by the uncontrolled proliferation of EBV-transformed B cells due to the diminished EBV-specific cytotoxic T lymphocyte response. Although there are currently no available data on ‘crippling’ mutations in our series of age-related EBV+ B-LPD, this disease seems to bear an immunodeficiency-related aspect probably caused by immune senescence. Indeed, Matsuoka's group recently reported that, using the real-time polymerase chain reaction, EBV DNA in peripheral blood mononuclear cells was more frequently detected in healthy individuals older than 70 years of age (eight of nine, 89%) than in those less than 70 years (1 of 11, 9%).(35) Yanagi et al. also demonstrated that EBNA2 IgG antibodies generated in young children by asymptomatic primary EBV infections remain elevated throughout life in their age-related analysis using sera, suggesting the intervention of reactivation of latent or exogenous EBV superinfection.(36) These phenomenon may constitute the background in the pathogenesis of EBV-associated cHL and DLBCL, both of which have a propensity to occur more frequently in the elderly >40 years of age. Interestingly, their age distribution patterns appeared to be distinct from each other. However, these definite and conceivable B-cell neoplasms (DLBCL and cHL) showed the highest incidence at 70–79 years of age as a total group (Fig. 5a). The relative ratio of EBV+ DLBCL to EBV+ cHL became higher, from 0.75 (50–59 years of age) to 2.7 (80–89 years of age), in parallel with the older patient populations (≥50 years) (Fig. 5b).
It is tempting to speculate that EBV+ cHL and EBV+ B-LPD may be related in their pathogenesis or that a true gray zone may lie between these two diseases.(37) However, much still needs to be learned regarding the detailed clinicopathological profiles, immunology, molecular biology, and therapeutic strategies of these diseases.
The authors are grateful to Dr Masao Seto for his continuous encouragement to prepare this manuscript. This work was supported in part by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science, a Grant-in-Aid for Cancer Research from the Ministry of Education, Culture, Sports, Science, and Technology, Japan, a Grant-in-Aid for the ‘Delineation of the molecular biological profile of refractory lymphoid malignancy and development of its tumor type-specific management’ from the Ministry of Health, Labor, and Welfare, Japan, and a grant (07-23906) from the Princess Takamatsu Cancer Research Fund.