In this study, we compared the clinical utility of whole blood EBV RQ-PCR and EBER-ISH for evaluation of patients with different lymphoma subtypes and demonstrated that EBV RQ-PCR identified more positive cases than EBER-ISH in three of four subtypes. In addition, we assessed the prognostic significance of results from whole blood EBV RQ-PCR relative to those from EBER-ISH in patients with lymphoma and found that EBV DNA positivity is a better predictor of patient survival than EBER status for multiple subtypes. Thus, our findings indicate that EBV RQ-PCR is an accurate and reliable method for detection of EBV in lymphoma patients that demonstrates high prognostic utility.

Although previous studies have investigated the impact of EBV positivity on prognosis for different lymphoma subtypes, several unanswered questions remain. In particular, the levels of concordance between results obtained from EBER-ISH and EBV RQ-PCR have varied in past reports.15Okamoto A. Yanada M. Inaguma Y. Tokuda M. Morishima S. Kanie T. Yamamoto Y. Mizuta S. Akatsuka Y. Yoshikawa T. Mizoguchi Y. Nakamura S. Okamoto M. Emi N. The prognostic significance of EBV DNA load and EBER status in diagnostic specimens from diffuse large B-cell lymphoma patients., 16Hohaus S. Santangelo R. Giachelia M. Vannata B. Massini G. Cuccaro A. Martini M. Cesarini V. Cenci T. D'Alo F. Voso M.T. Fadda G. Leone G. Larocca L.M. The viral load of Epstein-Barr virus (EBV) DNA in peripheral blood predicts for biological and clinical characteristics in Hodgkin lymphoma., 17Tisi M.C. Cupelli E. Santangelo R. Maiolo E. Alma E. Giachelia M. Martini M. Bellesi S. D'Alò F. Voso M.T. Pompili M. Leone G. Larocca L.M. Hohaus S. Whole blood EBV-DNA predicts outcome in diffuse large B-cell lymphoma., 18Kanakry J.A. Li H. Gellert L.L. Lemas M.V. Hsieh W.S. Hong F. Tan K.L. Gascoyne R.D. Gordon L.I. Fisher R.I. Bartlett N.L. Stiff P. Cheson B.D. Advani R. Miller T.P. Kahl B.S. Horning S.J. Ambinder R.F. Plasma Epstein-Barr virus DNA predicts outcome in advanced Hodgkin lymphoma: correlative analysis from a large North American cooperative group trial. In this study, we addressed this question by evaluating the agreement between EBER-ISH and EBV RQ-PCR for four different lymphoma subtypes—DLBCL, ENKTCL, PTCL, and HL. In addition, we compared the prognostic utility of the two detection methods individually and combined for these four subtypes and calculated advisable cutoff values for EBV RQ-PCR when determining EBV status.In one previous study, Hohaus et al16Hohaus S. Santangelo R. Giachelia M. Vannata B. Massini G. Cuccaro A. Martini M. Cesarini V. Cenci T. D'Alo F. Voso M.T. Fadda G. Leone G. Larocca L.M. The viral load of Epstein-Barr virus (EBV) DNA in peripheral blood predicts for biological and clinical characteristics in Hodgkin lymphoma. reported that viral DNA load in plasma displayed 90% specificity and 65% sensitivity for predicting the presence of EBER in HL cells, whereas viral DNA in whole blood showed 79% specificity and 79% sensitivity. The authors concluded that DNA load was an independent indicator of characteristics associated with worse prognosis, although it cannot be regarded as a definitive surrogate for EBER positivity. Likewise, Tisi et al17Tisi M.C. Cupelli E. Santangelo R. Maiolo E. Alma E. Giachelia M. Martini M. Bellesi S. D'Alò F. Voso M.T. Pompili M. Leone G. Larocca L.M. Hohaus S. Whole blood EBV-DNA predicts outcome in diffuse large B-cell lymphoma. did not observe significant association between the presence of EBV DNA in any blood compartment and EBER positivity in DLBCL patients. Conversely, Kanakry et al18Kanakry J.A. Li H. Gellert L.L. Lemas M.V. Hsieh W.S. Hong F. Tan K.L. Gascoyne R.D. Gordon L.I. Fisher R.I. Bartlett N.L. Stiff P. Cheson B.D. Advani R. Miller T.P. Kahl B.S. Horning S.J. Ambinder R.F. Plasma Epstein-Barr virus DNA predicts outcome in advanced Hodgkin lymphoma: correlative analysis from a large North American cooperative group trial. reported much better analytical performance for EBV RQ-PCR, with EBV DNA in plasma showing a 97% specificity and a 92% sensitivity for predicting EBER positivity in HL patients.In the present study, we found that the κ agreement between the two methods was highest for HL (Table 2). However, the negative agreement of 0.714, which is equivalent to the specificity, and the positive agreement of 0.821, which is equivalent to the sensitivity, slightly favored the results reported by Hohaus et al.16Hohaus S. Santangelo R. Giachelia M. Vannata B. Massini G. Cuccaro A. Martini M. Cesarini V. Cenci T. D'Alo F. Voso M.T. Fadda G. Leone G. Larocca L.M. The viral load of Epstein-Barr virus (EBV) DNA in peripheral blood predicts for biological and clinical characteristics in Hodgkin lymphoma. Nevertheless, given that the subject population analyzed by Kanakry et al18Kanakry J.A. Li H. Gellert L.L. Lemas M.V. Hsieh W.S. Hong F. Tan K.L. Gascoyne R.D. Gordon L.I. Fisher R.I. Bartlett N.L. Stiff P. Cheson B.D. Advani R. Miller T.P. Kahl B.S. Horning S.J. Ambinder R.F. Plasma Epstein-Barr virus DNA predicts outcome in advanced Hodgkin lymphoma: correlative analysis from a large North American cooperative group trial. (n = 116) was approximately two times larger than that analyzed by Hohaus et al16Hohaus S. Santangelo R. Giachelia M. Vannata B. Massini G. Cuccaro A. Martini M. Cesarini V. Cenci T. D'Alo F. Voso M.T. Fadda G. Leone G. Larocca L.M. The viral load of Epstein-Barr virus (EBV) DNA in peripheral blood predicts for biological and clinical characteristics in Hodgkin lymphoma. (n = 43 to 56) or the present study (n = 63), it would be prudent to not draw a definitive conclusion from this observation.In the case of DLBCL, κ agreement was substantially lower than total agreement, reflecting the proportion of agreement that could have occurred by chance. This is likely due to the fact that the proportion of EBV-positive DLBCL patients was small. Of note, among all lymphoma subtypes, the proportional difference between EBER-positive cases and EBV DNA-positive cases was the highest for DLBCL, with EBV RQ-PCR identifying more than three times as many EBV-positive patients than EBER-ISH. These findings are largely consistent with previously reported results15Okamoto A. Yanada M. Inaguma Y. Tokuda M. Morishima S. Kanie T. Yamamoto Y. Mizuta S. Akatsuka Y. Yoshikawa T. Mizoguchi Y. Nakamura S. Okamoto M. Emi N. The prognostic significance of EBV DNA load and EBER status in diagnostic specimens from diffuse large B-cell lymphoma patients.,17Tisi M.C. Cupelli E. Santangelo R. Maiolo E. Alma E. Giachelia M. Martini M. Bellesi S. D'Alò F. Voso M.T. Pompili M. Leone G. Larocca L.M. Hohaus S. Whole blood EBV-DNA predicts outcome in diffuse large B-cell lymphoma. and suggest that the relationship between DLBCL and EBV infection needs to be further elucidated. In patients with PTCL, positive agreement (ie, sensitivity) was highest, and negative agreement (ie, specificity) was lowest, among all lymphoma subtypes, indicating that the utility of EBV RQ-PCR as a screening test would be high, especially in PTCL. In the case of ENKTCL, because all patients were EBER positive, positive agreement was the same as total EBV DNA positivity rate, 0.663 (95% CI, 0.554–0.757).The potential value of whole blood EBV DNA load as a prognostic indicator of disease course has also been investigated in previous studies, including one that focused on ENKTCL.19Kim H.S. Kim K.H. Kim K.H. Chang M.H. Ji S.H. Lim D.H. Kim K. Kim S.J. Ko Y. Ki C.S. Jo S.J. Lee J.W. Kim W.S. Whole blood Epstein-Barr virus DNA load as a diagnostic and prognostic surrogate: extranodal natural killer/T-cell lymphoma. Although the authors only observed a trend toward shorter survival in EBV DNA-positive patients (P = 0.085) using the cutoff of 636.1 copies/mL, they suggested that EBV DNA load correlated well with the clinical course of ENKTCL. Consistent with these results, herein we found that EBV DNA-positive ENKTCL patients showed significantly worse OS than EBV DNA-negative ENKTCL patients (P = 0.0016). Possible explanations for the greater statistical significance in this present study include larger study population (n = 80) and use of a lower cutoff of 510 copies/mL. The prognostic value of EBV DNA in plasma (P = 0.0009) or in mononuclear cells (P = 0.006) from ENKTCL patients was also reported previously.20Suzuki R. Yamaguchi M. Izutsu K. Yamamoto G. Takada K. Harabuchi Y. Isobe Y. Gomyo H. Koike T. Okamoto M. Hyo R. Suzumiya J. Nakamura S. Kawa K. Oshimi K. Prospective measurement of Epstein-Barr virus-DNA in plasma and peripheral blood mononuclear cells of extranodal NK/T-cell lymphoma, nasal type. Our results are generally consistent with theirs and additionally show that EBV RQ-PCR analysis of whole blood can be useful for determining disease prognosis.The prognostic significance of EBV DNA in whole blood was also previously evaluated in PTCL patients.3Dupuis J. Emile J.F. Mounier N. Gisselbrecht C. Martin-Garcia N. Petrella T. Bouabdallah R. Berger F. Delmer A. Coiffier B. Reyes F. Gaulard P. Prognostic significance of Epstein-Barr virus in nodal peripheral T-cell lymphoma, unspecified: a Groupe d'Etude des Lymphomes de l'Adulte (GELA) study.,6Kim Y.R. Kim S.J. Cheong J.W. Chung H. Jang J.E. Kim Y. Yang W.I. Min Y.H. Kim J.S. Pretreatment Epstein-Barr virus DNA in whole blood is a prognostic marker in peripheral T-cell lymphoma.,21Weisenburger D.D. Savage K.J. Harris N.L. Gascoyne R.D. Jaffe E.S. MacLennan K.A. Rüdiger T. Pileri S. Nakamura S. Nathwani B. Campo E. Berger F. Coiffier B. Kim W.S. Holte H. Federico M. Au W.Y. Tobinai K. Armitage J.O. Vose J.M. Peripheral T-cell lymphoma, not otherwise specified: a report of 340 cases from the International Peripheral T-cell Lymphoma Project. Consistent with our data showing worse OS in EBV DNA-positive PTCL patients (P = 0.0013), Kim et al6Kim Y.R. Kim S.J. Cheong J.W. Chung H. Jang J.E. Kim Y. Yang W.I. Min Y.H. Kim J.S. Pretreatment Epstein-Barr virus DNA in whole blood is a prognostic marker in peripheral T-cell lymphoma. reported a significant association between presence of EBV DNA in whole blood and inferior OS for individuals with PTCL (P = 0.015). Notably, Dupuis et al3Dupuis J. Emile J.F. Mounier N. Gisselbrecht C. Martin-Garcia N. Petrella T. Bouabdallah R. Berger F. Delmer A. Coiffier B. Reyes F. Gaulard P. Prognostic significance of Epstein-Barr virus in nodal peripheral T-cell lymphoma, unspecified: a Groupe d'Etude des Lymphomes de l'Adulte (GELA) study. showed an association between EBER-ISH positivity and both OS (P = 0.01) and progression-free survival (P = 0.02), with an even stronger impact on survival for patients aged >60 years. Similarly, Weisenburger et al21Weisenburger D.D. Savage K.J. Harris N.L. Gascoyne R.D. Jaffe E.S. MacLennan K.A. Rüdiger T. Pileri S. Nakamura S. Nathwani B. Campo E. Berger F. Coiffier B. Kim W.S. Holte H. Federico M. Au W.Y. Tobinai K. Armitage J.O. Vose J.M. Peripheral T-cell lymphoma, not otherwise specified: a report of 340 cases from the International Peripheral T-cell Lymphoma Project. reported significant numbers of EBER cells as a prognostic indicator of OS (P = 0.044). Herein, however, we did not observe any significant differences in survival between EBER-positive and EBER-negative patients for any lymphoma subtype. Possible reasons for this discrepancy include the smaller subject population (n = 98) and shorter median follow-up period (16.3 months) of our study compared with previous studies,3Dupuis J. Emile J.F. Mounier N. Gisselbrecht C. Martin-Garcia N. Petrella T. Bouabdallah R. Berger F. Delmer A. Coiffier B. Reyes F. Gaulard P. Prognostic significance of Epstein-Barr virus in nodal peripheral T-cell lymphoma, unspecified: a Groupe d'Etude des Lymphomes de l'Adulte (GELA) study.,21Weisenburger D.D. Savage K.J. Harris N.L. Gascoyne R.D. Jaffe E.S. MacLennan K.A. Rüdiger T. Pileri S. Nakamura S. Nathwani B. Campo E. Berger F. Coiffier B. Kim W.S. Holte H. Federico M. Au W.Y. Tobinai K. Armitage J.O. Vose J.M. Peripheral T-cell lymphoma, not otherwise specified: a report of 340 cases from the International Peripheral T-cell Lymphoma Project. as the difference in survival rate between EBER-positive and EBER-negative patients becomes greater over time in PTCL. When only EBER-negative patients were subjected to this analysis, EBV DNA positivity retained its prognostic significance.It has also been reported that HL patients positive for EBV DNA in whole blood at pretreatment show worse event-free survival (P = 0.032) and OS (P 22Park J.H. Yoon D.H. Kim S. Park J.S. Park C.S. Sung H. Lee S.W. Huh J. Suh C. Pretreatment whole blood Epstein-Barr virus-DNA is a significant prognostic marker in patients with Hodgkin lymphoma. although EBER positivity was not effective for predicting prognosis of these patients. Similarly, Kanakry et al18Kanakry J.A. Li H. Gellert L.L. Lemas M.V. Hsieh W.S. Hong F. Tan K.L. Gascoyne R.D. Gordon L.I. Fisher R.I. Bartlett N.L. Stiff P. Cheson B.D. Advani R. Miller T.P. Kahl B.S. Horning S.J. Ambinder R.F. Plasma Epstein-Barr virus DNA predicts outcome in advanced Hodgkin lymphoma: correlative analysis from a large North American cooperative group trial. reported that EBV DNA-positive patients showed inferior failure-free survival compared with EBV DNA-negative patients, whereas no difference was observed for patients stratified by EBER positivity. In contrast, Dinand et al23Dinand V. Sachdeva A. Datta S. Bhalla S. Kalra M. Wattal C. Radhakrishnan N. Plasma Epstein Barr virus (EBV) DNA as a biomarker for EBV associated Hodgkin lymphoma. did not observe significant differences in OS or event-free survival for study subjects stratified according to either EBER positivity or plasma EBV DNA positivity. One published study by Keegan et al24Keegan T.H. Glaser S.L. Clarke C.A. Gulley M.L. Craig F.E. Digiuseppe J.A. Dorfman R.F. Mann R.B. Ambinder R.F. Epstein-Barr virus as a marker of survival after Hodgkin's lymphoma: a population-based study. did detect a significant association between EBER positivity and OS (P n = 922). Notably, this decrease in OS was primarily derived from patients with nodular sclerosis histologic subtype. Herein, we observed a trend toward shorter OS in EBV DNA-positive HL patients (P = 0.088). Also, we further performed univariate and multivariate analysis on our HL population stratified by age; however, no significant associations between EBV status and OS were observed for any group (data not shown). As borderline significance was shown (P = 0.088), and significant associations between EBV DNA positivity and survival were observed in past studies, further study on a greater population would be helpful for reaching a more definitive conclusion regarding the prognostic significance of EBV DNA in HL.Unlike the other three subtypes, for DLBCL patients in our study, EBV DNA positivity did not show significant or near-significant prognostic value, although we noted a trend toward shorter survival of EBER-positive patients (P = 0.068). Consistent with this observation, Park et al5Park S. Lee J. Ko Y.H. Han A. Jun H.J. Lee S.C. Hwang I.G. Park Y.H. Ahn J.S. Jung C.W. Kim K. Ahn Y.C. Kang W.K. Park K. Kim W.S. The impact of Epstein-Barr virus status on clinical outcome in diffuse large B-cell lymphoma. reported that EBER-positive DLBCL patients demonstrated poorer OS (P = 0.026) and progression-free survival (P = 0.018) relative to EBER-negative patients. However, in contrast to our findings, Tisi et al17Tisi M.C. Cupelli E. Santangelo R. Maiolo E. Alma E. Giachelia M. Martini M. Bellesi S. D'Alò F. Voso M.T. Pompili M. Leone G. Larocca L.M. Hohaus S. Whole blood EBV-DNA predicts outcome in diffuse large B-cell lymphoma. found that whole blood EBV DNA-positive patients showed significantly lower progression-free survival (P = 0.02) and OS (P = 0.05) versus EBV DNA-negative patients. Thus, further research is warranted to resolve this discrepancy. Interestingly, when only EBER-positive DLBCL patients were analyzed, we found that the presence of EBV DNA in whole blood was a significant predictor of poor prognosis. This is consistent with the finding that DLBCL patients whose test results from EBER-ISH and EBV RQ-PCR were both positive showed worse OS (P = 0.002) relative to other patients with DLBCL (Table 3). One possible explanation for this observation is that when pretreatment EBV DNA was identified as negative, this distinguished some EBER-ISH false-positive patients, whose samples displayed EBER signals in just a small number of bystander lymphocytes. As the minimal proportion of hybridized cells in tumor tissue required to be identified as a positive case was originally set to only 1% in our study, the EBER-ISH assay might be overly sensitive. We therefore attempted to elucidate the effect of using this more stringent cutoff by performing reanalysis after excluding the EBER-equivocal cases, as discussed below. We further note that because DLBCL patients showed a relatively lower EBER percentage-positive rate than EBV DNA percentage-positive rate (7.7% versus 28.2%, respectively), our data suggest that EBV RQ-PCR analysis would help differentiate false-negative results, which occur because of absence of EBER signal, resulting from inadequate RNA preservation, from true-negative cases, conferring additional benefits for diagnosing DLBCL related to EBV infection.Established interpretation criteria for EBER-ISH are lacking, and there is currently no consensus on what percentage of EBER-positive neoplastic cells justify a histopathologic diagnosis of EBV-related lymphoma. In particular, when EBER signal is observed in only a subpopulation of cells in the tissue, it remains unclear whether these cells are more likely to represent true neoplastic cells or just bystander lymphocytes, and in the former scenario, what mechanism might mediate this phenomenon.25Hänel P. Hummel M. Anagnostopoulos I. Stein H. Analysis of single EBER-positive and negative tumour cells in EBV-harbouring B-cell non-Hodgkin lymphomas.,26Ohashi A. Kato S. Okamoto A. Inaguma Y. Satou A. Tsuzuki T. Emi N. Okamoto M. Nakamura S. Reappraisal of Epstein-Barr virus (EBV) in diffuse large B-cell lymphoma (DLBCL): comparative analysis between EBV-positive and EBV-negative DLBCL with EBV-positive bystander cells. Accordingly, interpretation criteria for EBER-ISH have varied in previous studies (Table 7), and when different criteria were applied, the prevalence of EBV-associated lymphoma, as well as the prognostic effect of tumor tissue EBER positivity calculated on the basis of those criteria, have varied.31Wada N. Ikeda J. Hori Y. Fujita S. Ogawa H. Soma T. Sugiyama H. Fukuhara S. Kanamaru A. Hino M. Kanakura Y. Morii E. Aozasa K. Epstein-Barr virus in diffuse large B-cell lymphoma in immunocompetent patients in Japan is as low as in western countries. Thus, absence of a unified cutoff, combined with the potential for EBER signals from bystander lymphocytes that may be hard to differentiate from signal arising from true neoplastic cells, presents a major obstacle for consistent interpretation of the results from this assay.

Table 7Previous Studies on the Prognostic Impact of Tumor Tissue EBV Status in Patients with Lymphoma

EBV, Epstein-Barr virus; N/A, not applicable.

Herein, to minimize the effect of EBER-ISH interpretation criteria on our findings, we performed a reanalysis including only those patients with unambiguously positive or negative results. We found that for DLBCL and PTCL, κ agreement between the two methods was substantially lower than previously calculated (Tables 2 and 5). That said, we cannot draw a definitive conclusion from this finding, as the CIs of the κ values overlapped with those from our initial analysis. In the case of ENKTCL and HL, sizes of the excluded subpopulations were zero or minimal, and accordingly, there was no remarkable change in agreement with reanalysis for these subtypes. Notably, when only subjects with unambiguous EBER-ISH results were included, multivariate analysis using the Cox proportional hazard model revealed an association between EBER positivity and OS for DLBCL (P = 0.037). This finding highlights the importance of establishing a unified cutoff for EBER-ISH and suggests that discrepant observations from previous studies could be at least partially attributable to disparate cutoffs used in EBER-ISH interpretation. We therefore hope that future studies will establish the most effective and practical percentage cutoffs for each lymphoma subtype to ensure accurate and reproducible findings.

That being said, there are occasions where the interpretation criteria based on a single numeric cutoff would be insufficient, and both an in-depth review of the patient's clinical status and the observer's expertise in clinical pathology would have to be involved in the interpretation of EBER stains. There cannot be a sole reliance on percentage cutoffs if cells with positive signal could not be definitively categorized into either benign bystander cells or clonal malignant cells. Technical interference could also lead to ambiguous or false results in the clinical interpretation of EBER stains. Specifically, false-negative results most often result from ribonuclease-mediated loss of EBER in the tumor tissues, especially when a control probe for RNA preservation is not utilized. We are of the view that EBV viral load in blood might complement EBER-ISH in these instances as supporting evidence to help assign EBV status in a tumor.

There are several caveats to consider when making clinical decisions based on results from either of these assays or a combination of both. In particular, EBV DNA in blood may originate from temporary reactivation of circulating memory B cells latently infected with EBV, rather than EBER-positive clonal cells in tumor tissue, and there are several clinical situations that might lead to this outcome. For example, immunosuppression from cancer or treatment thereof, or immunosuppression therapy before and after organ transplant, could promote reactivation of EBV infection.32Bakker N.A. Verschuuren E.A. Erasmus M.E. Hepkema B.G. Veeger N.J. Kallenberg C.G. van der Bij W. Epstein-Barr virus-DNA load monitoring late after lung transplantation: a surrogate marker of the degree of immunosuppression and a safe guide to reduce immunosuppression.,33Lazzarino M. Orlandi E. Baldanti F. Furione M. Pagnucco G. Astori C. Arcaini L. Viglio A. Paulli M. Gerna G. Bernasconi C. The immunosuppression and potential for EBV reactivation of fludarabine combined with cyclophosphamide and dexamethasone in patients with lymphoproliferative disorders. Furthermore, as these situations all indicate systemic immunologic impairment in lymphoma patients, elevated EBV DNA in blood might independently affect patient prognosis, regardless of actual tumor tissue EBER positivity.34Whole blood EBV-DNA: a surrogate for immune dysfunction in aggressive lymphoma?. In addition, primary infection with EBV is another possible reason for elevated EBV DNA in blood.35Chan K.H. Ng M.H. Seto W.H. Peiris J.S. Epstein-Barr virus (EBV) DNA in sera of patients with primary EBV infection. Therefore, when EBV-directed therapy is considered for a patient with lymphoproliferative disease or lymphoma, the therapeutic strategy should be based on EBV positivity of malignant lymphocytes themselves, considering the expenditure and risk of possible adverse effects from these therapies.36Bollard C.M. Gottschalk S. Torrano V. Diouf O. Ku S. Hazrat Y. Carrum G. Ramos C. Fayad L. Shpall E.J. Pro B. Liu H. Wu M.F. Lee D. Sheehan A.M. Zu Y. Gee A.P. Brenner M.K. Heslop H.E. Rooney C.M. Sustained complete responses in patients with lymphoma receiving autologous cytotoxic T lymphocytes targeting Epstein-Barr virus latent membrane proteins.,37Icheva V. Kayser S. Wolff D. Tuve S. Kyzirakos C. Bethge W. Greil J. Albert M.H. Schwinger W. Nathrath M. Schumm M. Stevanovic S. Handgretinger R. Lang P. Feuchtinger T. Adoptive transfer of Epstein-Barr virus (EBV) nuclear antigen 1-specific T cells as treatment for EBV reactivation and lymphoproliferative disorders after allogeneic stem-cell transplantation.

We note that this study has several limitations, including its retrospective design. In addition, only patients with the four most prevalent lymphoma subtypes—namely, DLBCL, ENKTCL, PTCL, and HL—were analyzed. Thus, future studies that include larger populations of subjects with lymphomas having more diverse histologic classification and provide sufficient statistical power for analysis would be beneficial. A third limitation of this study is that we did not use a parallel RNA control probe for EBER-ISH to confirm adequate RNA preservation, as this probe was not available because of customs restrictions. Moreover, the primary tumor sites of study subjects were not investigated or accounted for in our study. Thus, the amount of EBV DNA shed from tumor cells may not be well represented by blood EBV DNA level in cases where the tumor was located in a unique organ comprising a restricted niche, such as the brain. Finally, we did not address the question of which peripheral blood compartment is most representative for EBV DNA positivity in lymphoma patients, as only EBV DNA loads derived from whole blood samples were available in this study.

A major strength of our study, however, is the relatively large size of our patient population. Especially for DLBCL, which is the world's most prevalent lymphoma subtype,38Lee H. Park H.J. Park E.H. Ju H.Y. Oh C.M. Kong H.J. Jung K.W. Park B.K. Lee E. Eom H.S. Won Y.J. Nationwide statistical analysis of lymphoid malignancies in Korea.,39Morton L.M. Wang S.S. Devesa S.S. Hartge P. Weisenburger D.D. Linet M.S. Lymphoma incidence patterns by WHO subtype in the United States, 1992-2001. our subject population was larger than those included in most previous studies that investi