The Epstein Barr virus DNA levels as a tumor marker in EBV-associated cancers

Abstract

The Epstein Barr virus (EBV) is causally associated to several tumors of epithelial and lymphoid origin. The cancerogenic role in other than B cells has not been proven. This virus has been considered as a target in the effective diagnosis of EBV-associated tumors. For this purpose, molecular biology methods to measure EBV DNA load in the circulation of patients suffering from EBV-related cancers have been recently developed. In this review, we discuss the role of EBV DNA determination, the technical limitations of molecular assays measuring viral load and their impact on the clinical management of patients with EBV-associated tumors arising in the immunocompetent host. Several studies have recently clarified the biological and clinical characteristics of herpesvirus-associated tumors. However, some additional issues must be clarified before introducing viral load determinations into clinical practice. Firstly, since the various EBV-related tumors have different etiopathological and clinical characteristics, the most appropriate biological samples and analytical cut off values must be clearly defined in each group of patients. Secondly, a standardization of the assay, including the definition of the gene segment to be amplified, the use of an international reference for the standard curve and disease-related cut-off values, is strongly required. Thirdly, the interpretation of laboratory data may benefit from an improved design of the studies and obtaining an aggregrate of patients from different institutions, pooling these together, in order to have a sample size that is adequate to reinforce the statistical power of the studies.

Introduction

The Epstein Barr virus (EBV) is a ubiquitous human herpesvirus infecting over 90% of the world’s population. After the acute infection, the virus establishes a life long latency that is clinically asymptomatic. However, the clinical and biological evidence indicates that EBV may disrupt this latency and become causally linked to several tumors, especially those of lymphoid and epithelial origin (Mc Sween and Crawford 2003; Cohen 2000; Anderson 2006) by a pathogenetic mechanism that is still partial and fragmentary. EBV is associated with almost all cases of undifferentiated carcinoma of nasopharyngeal type (UCNT WHO type 3) and endemic Burkitt’s lymphoma (BL). Whereas in squamous cell carcinoma (WHO type 1), non-keratinising carcinoma (WHO type 2), NK/T cell lymphoma, Hodgkin’s disease (HD) and lymphoproliferative disorders of the immunocompromised host, EBV-association is variable, ranging from 10 to 100% of the cases (Mc Sween and Crawford 2003; Cohen 2000). The EBV genome consists of approximately 100 genes, present in neoplastic cells of EBV-related tumors in an episomal form, and replicated during cellular division by host cell machinery. Of these 100 genes, only 10 are expressed in infected B cells during latency. LMP-1, LMP-2, EBNA-1, EBNA-2 and EBNA-3, are the most important genes expressed during latency. In particular, LMP-1 and LMP-2 genes are important in the process of neoplastic transformation by acting as oncogenes and inducing signalling responses or activation of the PI3K/Akt signalling pathway, respectively, and increasing resistance to apoptosis in EBV infected cells.

In addition to its cancerogenic role, this virus has been considered as a target in the effective diagnosis and monitoring of EBV-associated tumors. For this purpose, molecular biology methods to measure EBV DNA load in the circulation of patients suffering from EBV-related cancers have been recently developed (Lo et al. 1999; Kimura et al. 1999; Stevens et al. 2001; Chan and Lo 2002). Initially, the use of a qualitative PCR assay demonstrated that circulating extracellular EBV DNA was present only in 30% of UCNT patients (Mutirangura et al. 1998), but early on the use of a real time PCR assay, with increased sensitivity, showed that peripheral blood EBV DNA was detectable in 96% of patients, suggesting that viral load may be important for diagnostic and prognostic purposes in this type of cancer (Lo et al. 1999; Chan and Lo 2002). In EBV-related tumors arising in immunocompetent hosts, like UCNT, HD and NK/T cell lymphomas, peripheral blood EBV DNA is not related to viral reactivation, but probably derives from tumor cells and therefore may reflect the characteristics of the tumor. Consequently, in these cancers EBV DNA may be considered as a real tumor biomarker. A recent study elucidating the molecular nature of EBV DNA in these tumors adds additional argument to support this hypothesis (Chan et al. 2003). These authors perfected an experimental approach aiming at discriminating whether circulating EBV DNA molecules exist as fragments or as components of intact virions. For this purpose, the plasma of UCNT patients was treated with DNAse and ultracentrifugation. These studies demonstrated that circulating EBV DNA is not a part of viral particles, but is a “naked” molecule susceptible to DNAse digestion. In addition, it is not pellettable by ultracentrifugation and exists in fragment form below 181 bp in size. The former observations suggest that the virus is not actively released as an intact particle after reactivation from latency, but it may be released into circulation by an apoptotic process (Mutirangura et al. 1998; Chan et al. 2003). While several experimental data support this hypothesis in UCNT, in BL occurring in African children (a tumor that is strongly associated with EBV), the nature of the EBV viral load has not yet been elucidated (Cohen 2000). A recent investigation showed that, in populations at high risk for the development of BL like healthy Kenyan children and Ugandan children with sickle cell disease, but also in their mothers, EBV DNA was frequently detected in peripheral blood buffy coats, suggesting the presence of a high number of EBV infected circulating cells (Mbulaiteye et al. 2006). However, it is not clear whether EBV DNA detection in these subjects may be considered as a prognostic marker, since this study did not measure plasma EBV DNA, that may more precisely indicate an active viral replication. The study also omitted an adequate follow-up period to eventually link high EBV viremia to an increased risk in the development of BL. For these reasons, more biological and clinical studies on endemic and sporadic BL are needed to clarify the possible utility of EBV DNA as a diagnostic or prognostic marker in this type of cancer.

The pathogenesis of EBV-related lymphoproliferative diseases arising in the immunocompromised host, like post transplant lymphoproliferative disorders (PTLD), is different from the pathogenesis occurring in the above mentioned tumors. In these cancers, EBV DNA is actively released from circulating EBV-positive cells proliferating because of the absence of an effective T cell control (Taylor et al. 2005). For this reason, the EBV DNA level reflects the degree of immunosuppression rather than the characteristics of the tumor, and may not be considered a true tumor marker.

In this review, we shall discuss the role of EBV DNA determination, the technical limitations of molecular assays measuring viral load and their impact on the clinical management of patients with EBV-associated tumors arising in the immunocompetent host.

Technical hints on EBV quantitative assays

Although in the last few years the application of quantitative nucleic acid detection techniques has had a major impact on virological diagnosis, there are still several problems to be solved before molecular diagnostic assays for the determination of EBV DNA can be introduced in current clinical practice. In fact, differences in specimen characteristics, in performances of assay methods and difficulties in standardization and quality assurance through control programmes and misinterpretation of the analytical data may yield lack of reproducibility and/or inconsistent results (Stevens et al. 2001; Chan and Lo 2006; Niesters 2004). These considerations are particularly true when new markers are measured with innovative molecular biology assays; furthermore, in the case of EBV the interpretation of analytical data is complicated by the fact that this microorganism is widely diffuse in the human population. The major technical problems in the laboratory assays measuring EBV DNA as a tumor marker can be summarized as follows. Firstly, the amount of EBV DNA detected is influenced by the specimen type used to measure the viral load. Although various specimens have been used to quantitate EBV DNA in human diseases, the majority of the studies on EBV-related tumors arising in the immunocompetent host appropriately use plasma to measure EBV DNA (Lo et al. 1999; Lin et al. 2004; Gandhi et al. 2006; Au et al. 2004). In support of this hypothesis, Lit et al. 2004) were unable to find EBV DNA in the cellular compartment of UCNT patients. By using more specific molecular approaches, we subsequently have demonstrated that, although EBV DNA can be measured in peripheral blood mononuclear cells (PBMCs) from healthy and UCNT subjects (Pratesi et al. 2003), the sensitivity and specificity in detecting EBV DNA is much higher in the plasma and serum of immunocompetent patients with EBV-associated cancers (Bortolin et al. 2006). Hence, it is conceivable that these findings may reflect the origin of EBV DNA, that is probably released from tumor cells into the acellular compartment through mechanism(s) that are still unknown (Cohen 2000; Taylor et al. 2005). A single study considered whole blood as the best source to accurately measure EBV DNA (Stevens et al. 2001), but this study was limited to infectious mononucleosis, transplanted patients and HIV+ subjects and did not include the diseases where EBV DNA can be considered as a tumor marker. For these reasons, plasma can be defined as the preferred clinical specimen to measure EBV DNA as a diagnostic and prognostic marker in UCNT, HD, and NK/T cell lymphomas (Lin et al. 2004; Gandhi et al. 2006; Au et al. 2004).

On the contrary, in these clinical conditions the quantity of EBV reservoir in memory or in activated B cells is not modified. This situation is probably different in EBV-related tumors in the immunocompromised patients, where EBV DNA is mainly derived from the virus replicating in the B lymphocyte compartment that is not effectively controlled by cell-mediated immunity.

Extraction techniques

Quantitative analysis of nucleic acids requires that EBV DNA be extracted from biological samples. The commercially available Qiagen reagents are often used to extract EBV DNA before it is subjected to amplification (Lo et al. 1999; Taylor et al. 2005; Bortolin et al. 2006). This method is simple and less time consuming than other methods. However, monitoring the extraction process for loss or inhibition of viral genetic material has not been performed in many studies. So as to clarify this issue, we compared the efficiency of Qiagen technique with organic solvent extraction and demonstrated that the Qiagen technique underestimates viral load as compared to the organic solvent extraction system; this is more efficient in extracting viral DNA and produces higher values of EBV DNA (Bortolin et al. 2006).

Amplification technique-gene segment to be amplified

Although initial studies used qualitative PCR or competitive quantitative PCR (Mutirangura et al. 1998; Pratesi et al. 2003), these methods are not considered accurate anymore. Therefore, real time PCR represents the method of choice for EBV DNA quantitation and it is widely accepted that commercially available equipment has equivalent technical performances (Niesters 2004). A still unresolved problem concerns the most appropriate EBV gene segment to be amplified reflecting the true number of EBV genomes in the biological samples of interest. Many technical protocols include the amplification of the BamHI-W region (Lo et al. 1999; Taylor et al. 2005; Shao et al. 2004), but this region is randomly repeated a variable number of times in each subject; therefore, because it does not accurately reflect the true number of EBV genomes present in each sample, it cannot be used to compare results obtained in various patients (Stevens et al. 2001). More appropriately, other studies amplified genes that are present in the EBV genome in single copy, such as LMP-1, LMP-2, BALF-5 or Pol-1 (Lin et al. 2004; Gandhi et al. 2006; Au et al. 2004; Bortolin et al. 2006; Le et al. 2005). In this case, it has been demonstrated that the accuracy and specificity of the molecular assay increase, making the choice of cut-off values and comparison between patients more standardized. In particular, Le et al. (2005) suggested that the Pol-1 assay was the most accurate test, as compared to BAMHI-W and LMP-2. Since different EBV gene expression patterns exist between the various EBV-associated cancers, it is not clear if these differences may be important in the design of disease-tailored EBV DNA amplification assays. Besides the choice of the gene segment to be amplified, properly designed amplification assays for UCNT patients should use primers that take into account that circulating EBV DNA consists of short DNA fragments (Chan et al. 2003; Chan and Lo 2006).

Standard curve

A calibration curve for each assay aiming at measuring circulating EBV DNA must be established. One limitation of current EBV viral load quantification is the use of various cell lines and plasmids to construct the standard curve. In fact, in many studies the standard curve was calculated by using the Namalwa cell line, a diploid cell line containing two integrated EBV genomes/cell (Lo et al. 1999; Taylor et al. 2005; Bortolin et al. 2006). However, other studies included the JY or Raji cell lines (Stevens et al. 2001; Gandhi et al. 2006) or cloned EBNA-1 plasmids (Au et al. 2004). The techniques used as a gold standard for the quantification of EBV genome numbers in cell lines usually have a limited accuracy; furthermore, cellular or plasmid DNA might have different amplification efficiencies. For these reasons, the use of different sources for the construction of the standard curve may generate non-comparable results in different molecular assays. Standardization of quantitative EBV DNA measurements through the use of an internationally-defined standard curve is therefore strongly required.

EBV viral load as a screening, diagnostic or staging marker

In geographic areas where UCNT is frequent, laboratory assays are necessary to screen populations at higher risk for developing the tumor (i.e. family members of UCNT patients in Southern China). The anti viral capsid antigen (VCA) IgA antibody has been widely used for this purpose. Antibody assays are widely used as screening tests because of limited cost estimates and since their sensitivity and specificity have been generally reported to be in the order of 85–90%. Serological surveys for UCNT may also be positive in other EBV-associated diseases and are semi-quantitative, making it difficult to define an appropriate cut off value (Shao et al. 2004; Le et al. 2005; Leung et al. 2003). The use of EBV DNA molecular testing has been proposed as a screening marker in endemic areas; however, this test may be contraindicated by the technical complexity and the cost of the assay. Therefore, it seems to be more reasonable to propose the use of EBV DNA as a second level confirmatory test, when IgA-VCA results are positive or indeterminate (Leung et al. 2004). On the contrary, the incomplete association of the virus with the tumor, the availability of more effective histopathological and imaging techniques and the clinical characteristic of the tumors contraindicate the use of EBV DNA as a screening marker in EBV-associated lymphoproliferative disorders of the immunocompetent host (HD and NK/T cell lymphomas).

Besides its use as a prognostic tool, several studies reported that EBV DNA viral load may be used in the diagnosis of EBV-associated tumors. Although the diagnosis of UCNT requires a histopathological assessment of a biopsy taken from the post nasal space, this is an invasive and painful procedure which sometimes leads to non-diagnostic results. Therefore, the use of non-invasive assays may be useful in patients suspected of having UCNT based on clinical examinations or in cases where histology gave inconclusive results. In high incidence areas, plasma EBV DNA has a diagnostic sensitivity of 95% (Taylor et al. 2005; Leung et al. 2004), while in low incidence regions, sensitivity lowers down to 70–75%. (Bortolin et al. 2006). Studies available at this moment include the prospective or retrospective use of EBV DNA as a diagnostic tool in all UCNT patients, while no reports are available concerning the importance of this marker in establishing the diagnosis in patients with undetermined histology; a situation where molecular assays may be very helpful to appropriately start treatments. Some authors claimed that peripheral blood EBV DNA values negative or close to the lower detection limit are not uncommon in UCNT patients (Stevens et al. 2006; Tong et al. 2002); therefore, it was suggested that this marker may not be well suited for the primary diagnosis of this tumor. They hypothesized that a diagnosis of UCNT may be more accurately established when elevated EBV DNA levels are present at the site of the primary tumor. In support of their hypothesis, Stevens et al. (2006) and Tong et al. (2002) demonstrated that EBV DNA levels in nasopharyngeal brushings from UCNT patients were significantly higher than in brushings from healthy controls. Since normal nasopharyngeal (NP) epithelium does not contain EBV DNA, the cut off value between cases and controls can be sharply defined (Sam et al. 1993). Sensitivity, specificity, positive and negative predictive values were >91% (Stevens et al. 2006; Tong et al. 2002). It was also demonstrated that NP brushing has a false negative rate comparable to that of the diagnostic biopsy, but it is more easily feasible, it has a reduced economic cost and, in case of undetermined results, it can be repeated because of its non-invasive nature. Although these studies used an appropriate study design, proper statistical analyses and raised the possibility of detecting molecular tumor markers in NP brushings with good analytical performances, the number of patients included in these studies is too small to draw definitive conclusions.

EBV viral load as a prognostic marker

Prognostic markers may be useful to identify patients at different risk for specific outcomes and/or facilitate treatment choice. Although tumor stage was generally accepted as an important prognostic factor in UCNT patients, clinical observations suggested that a subgroup of patients with early stage disease could also have a poor prognosis (Chan et al. 2003; Chan and Lo 2006). Some evidence suggests that the measurement of EBV nucleic acids in the peripheral blood may be useful as an independent prognostic marker in UCNT and therefore may be used to further stratify patients according to multiple risk factors (Lo et al. 1999; Chan et al 2003; Chan and Lo 2006; Lin et al. 2004). In general, quantitative analysis allowed us to correlate the levels of plasma EBV DNA with disease staging, suggesting that stage I/II patients have considerably lower values of this marker as compared to stage III and IV patients (Mutirangura et al. 1998; Chan and Lo 2006; Lin et al. 2004). An additional fact was provided by recent studies, where we and others found that EBV viremia was also related to nodal invasion of the tumor (Bortolin et al. 2006; Tong et al. 2002). Thus, present data suggest that EBV DNA concentration at diagnosis seems to reflect tumor burden. The results of representative studies correlating EBV DNA levels with disease stage are summarized in Table 1.

Table 1.

Results of representative studies correlating EBV DNA levels with disease stage in undifferentiated carcinoma of nasopharyngeal type (UCNT) and Natural Killer/T (NK/T) cell lymphoma patients

Reference	Type of cancer	Stage	EBV DNA (copies/ml)
Lin et al. (2004)	UCNT	III	681
		IV	1,703
		Metastasis	29,1940
Lo et al. (1999)	UCNT	I/II	5,918
		III/IV	47,047
Bortolin et al. (2006)	UCNT	I/II	772
		III	8,417
		IV	11,495
Au et al. (2004)	NK/T lymphoma	I/II	3.2 × 107
		III/IV	8.3 × 108

In order to identify prognostic factors related to therapy, Lo et al. (2000) initially demonstrated that EBV DNA declined exponentially after radiotherapy suggesting that this decrease may reflect the in vivo radiosensitivity of tumor cells. However, since EBV DNA blood levels depend on the balance between production and catabolism, validation of this hypothesis requires proof that EBV DNA is rapidly eliminated from circulation. In fact, the same authors demonstrated that the in vivo elimination of EBV DNA from circulation after surgical removal of the tumor was very fast, with a calculated half life of 139 min (To et al. 2003).

By showing that EBV DNA is present in the plasma of UCNT patients and disappears after radiotherapy, Lo et al. (2000) were the first authors suggesting that viral load could reflect disease activity and, possibly, that it may be used to monitor this tumor.

After this pioneer work, several papers on the clinical use of EBV DNA as a monitor of EBV-associated tumors appeared in the literature. Some of these studies using viral load as a prognostic marker in areas of low and high incidence of EBV-associated tumors and its relationship with overall survival (OS) of patients suffering from EBV-associated cancers are summarized in Table 2. Although the importance of genetic and environmental background driving UCNT development in areas of high or low incidence is poorly defined at this moment, the value of EBV viremia as a prognostic marker seems to be consistent in studies including patients from both types of geographic regions. In fact, OS or relapse free survival were statistically, significantly lower in patients with elevated EBV DNA concentration. Although the studies were in general appropriately conducted (i.e. patients were enrolled prospectically, and statistical analyses were accurate) the interpretation of the data and the clinical utility of this marker are still limited by some confusing factors. For example, the choice of an appropriate cut-off value of EBV DNA copies per milliliter of plasma or serum is critical to successfully discriminate patients with short or long survival. In the case of UCNT, data comparison across studies is not easily feasible, because each study uses a different cut-off value, ranging from 5–10 copies (i.e. the lowest detection limit of molecular assays) to 4,000 copies. The appropriate timing to prospectically measure EBV viremia is not clear: in some studies pre-treatment EBV DNA levels were considered to be the best prognostic predictor (Taylor et al. 2005; Bortolin et al. 2006), while other studies consider post-treatment levels as the best predictors (Lo et al. 1999; Le et al. 2005; Lo et al. 2000). Data comparison is further complicated by the differences in the duration of follow up, ranging from 1 to 4 years. Although NK/T cell lymphoma is a rare disease and the number of patients studied is low, the data presented by Au et al. (2004) suggest that this disease is characterized by higher EBV viral load and a considerably shorter OS as compared to UCNT; therefore, diagnostic and prognostic criteria considered for UCNT may not apply to this type of tumor.

Table 2.

Viral load as a prognostic marker in areas of low and high incidence of EBV-associated tumors and its relationship with overall survival (OS) of patients suffering from EBV-associated cancers

Reference	Type of tumor	No. of patients	Type of the study	Prognosis according to EBV DNA levels or positivity	Clinical cut-off
Chan et al. (2003)	UCNT	170	Prospective	<4,000 copies OS 1 y = 93%, >4,000 copies OS = 48%	4,000 EBV DNA copies
Le et al. (2005)	UCNT	58	Prospective	EBV DNA—OS 2 ys = 94%, EBV DNA+ OS = 55%	EBV DNA+
Lin et al. (2004)	UCNT	99	Prospective	<1,500 copies OS 2 ys = 100%, >1,500 copies OS = 83.4%	1,500 EBV DNA copies
Twu et al. (2007)	UCNT	114	Prospective	EBV DNA—OS 4 ys = 100%, EBV DNA+ OS = 63%	EBV DNA+
Au et al. (2004)	NK/T cell lymphomas	23	Prospective	<6.1 × 107 copies OS = 54 months, >6.1 × 107 copies OS = 2.1 months	6.1 × 107 copies

UCNT undifferentiated carcinoma of nasopharyngeal type, NK/T Natural Killer/T, y year, ys years

Collectively, these studies suggested that EBV DNA is a useful tool for classifying UCNT patients into different prognostic subgroups. At this moment there is no indication whether stratification of patients on the basis of different EBV DNA levels may be used to adjust treatment schedules. It is conceivable that this statement may represent a further necessary step allowing the appropriate use of an EBV DNA assay in clinical practice.

Since EBV genomes are present in malignant cells in 30% up to 65% of HD patients (mainly included in the mixed cellularity type) (Cohen 2000; Anderson 2006), EBV DNA may be used as a tumor marker only in a fraction of these patients. In agreement with histopathological data, EBV DNA is detectable in the plasma of all EBV positive patients before therapy (Gandhi et al. 2006; Gallagher et al. 1999) and, in one study, this predicts response to therapy (Gandhi et al. 2006), while it is not detectable in EBV negative HD patients. The possibility to measure plasma EBV DNA only in cases where tumor cells express EBV, the very limited number of patients included in the studies, the absence of information on the possible correlation with overall or disease free survival, considerably limit the interpretation of the utility of this marker for HD patients.

Clinical studies: design, objective, sample size, statistical analysis, limitations

With the goal to define an appropriate design of clinical protocols in cancer patients, it has been suggested that predictive marker studies should be conducted with randomized trials, require a sufficient sample size and sophisticated, multiparametric-like analyses, as well as statistical methods (McShane et al. 2005). Prospective studies are considered as providing much more accurate information compared to retrospective studies, where clinical information may be incomplete and difficult to recover from data repositories. Published clinical trials exploring EBV DNA as a tumor marker are mostly prospective and, in some cases, include appropriate statistical analyses. However, the incomplete knowledge of the mechanisms linking EBV to cancerogenesis in some types of cancers, the limitation of sample size and the lack of an international standardization of the assay are major concerns limiting the importance of the published studies. However, current data are very promising, since EBV viremia cannot be used as a tumor marker in relatively small proportions of EBV-associated cancers, mostly because it is still not detectable with available techniques (10–20% of UCNT patients in areas of low incidence); otherwise because the same type of cancer probably includes disease with multifactorial mechanisms of cancerogenesis (i.e. HD cases). An additional limitation of current studies regards the length of the observation period, i.e. at present, limited to 4 years after diagnosis, while tumor relapses may occur 10 years after initial diagnosis. For this reason, future studies must prolong the observation period of the patients to include all possible cases of tumor relapse.

The clinical usefulness of EBV DNA measurements in EBV-associated cancers is complicated by the fact that it is usually difficult to establish the value of a marker from single studies, while a clearer view may emerge from scrutiny across multiple studies by means of systematic review or meta-analyses (Altman and Riley 2005). Information from such studies is still lacking and may be a key element for decision-making regarding the development of reliable diagnostic assays, and the critical evaluation and clinical use of EBV viremia.

Conclusions

Substantial evidence implicates EBV in the pathogenesis of selected tumors arising in epithelial or lymphoid tissues. In fact, EBV is present in a high proportion of UCNT patients from high and low-risk geographic regions and in a considerable percentage of HD and NK/T cell lymphoma patients. EBV DNA present in the peripheral blood is probably released from tumor cells and therefore it may be considered as a possible tumor marker in these patients. Although the biological and clinical information on the characteristics of EBV-associated tumors has been greatly improved in the recent years, some additional issues must be clarified before introducing viral load determinations into clinical practice. Firstly, since the various EBV-related tumors have different etiopathological and clinical characteristics, the most appropriate biological samples and analytical cut-off values are probably different in each group of patients. Secondly, an international standardization of the assay, including the definition of the gene segment to be amplified, the use of an international reference for the standard curve and disease related cut-off values, is strongly required; thirdly, the interpretation of laboratory data may benefit from an improved design of the studies and pooling patients from different institutions together in order to obtain a sample size that is adequate to reinforce the statistical power of the studies.

Although EBV DNA may represent a useful tumor marker in EBV-associated tumors, incomplete knowledge of the pathogenesis of these cancers and substantial technical limitations of laboratory assays detecting EBV DNA are still major hurdles; an additional effort is now required to rapidly solve the problems related to EBV DNA determination before incorporating this marker into routine clinical use.