Clinical Utility of Epstein-Barr Virus DNA Testing in the Treatment of Nasopharyngeal Carcinoma Patients

Abstract

Epstein-Barr virus (EBV) DNA analysis has been shown to be useful for early detection, prognostication, and monitoring of treatment response of nasopharyngeal carcinoma (NPC), and the recent literature provides growing evidence of the clinical utility of EBV DNA testing, particularly to inform treatment decisions for NPC patients. Despite the fact that NPC is a rare disease, the NRG Oncology cooperative group has successfully activated a phase 2/3 randomized clinical trial for NPC with international partners and in that process has discovered that the development of a harmonized EBV DNA test is absolutely critical for integration into clinical trials and for future deployment in clinical and central laboratories. In November 2015, the National Cancer Institute (NCI) convened a workshop of international experts in the treatment of NPC and EBV testing to provide a forum for discussing the state of EBV DNA testing and its clinical utility, and to stimulate consideration of future studies and clinical practice guidelines for EBV DNA. This review provides a summary of that discussion.

Introduction

Epstein-Barr virus (EBV) exposure is ubiquitous, with over 90% of adults worldwide infected and seropositive for EBV, which generally establishes itself in B cells (1). It is well known to cause infectious mononucleosis, a common and relatively mild chronic inflammatory syndrome, and it is also linked to other diseases and multiple cancers, including Burkitt lymphoma (a type of non-Hodgkin lymphoma [NHL]) and some B cell and T cell NHL, Hodgkin lymphoma, gastric cancer, and nasopharyngeal carcinoma (NPC). Interestingly, the incidence of NPC varies dramatically by geographic region, ethnicity, and gender (greater prevalence in male individuals). Although NPC is considered a rare cancer in the United States and Europe, with an overall incidence of 0.5 to 2.0 per 100,000 person-years, it is much more common in certain parts of Southeast Asia. The reported age-standardized incidence of NPC among male individuals in Sihui City in southern China, for example, is 29 per 100,000 person-years (2).

Inasmuch as clonal populations of transcriptionally active EBV viral DNA molecules are readily detected in cells from nasopharyngeal carcinomas, researchers investigated whether EBV DNA from brushed nasopharyngeal epithelium, serum, or plasma had potential as a biomarker for NPC. Early studies indicated a significant association between the presence of histologically confirmed NPC and an elevated level of EBV DNA in plasma (3, 4). This result was confirmed and extended by other investigators studying NPC in Asia. In 2006, Leung et al (5) demonstrated that the EBV DNA load in plasma, as measured by a quantitative polymerase chain reaction (PCR) assay, correlated inversely with prognosis (overall survival) and could be used to refine estimated survival for early stage disease at diagnosis. Specifically, when patients with stage I or II NPC were stratified using a pretreatment cutoff of 4000 copies EBV DNA/mL plasma and were followed up for 8 years, patients with EBV DNA load below the cutoff had a 5-year survival of 91%, whereas those with levels above the cutoff had a 5-year survival of 64% (P=.0003). Clinicians began to recognize that plasma EBV DNA at diagnosis had potential as a quantitative biomarker of risk of progression after curative-intent radiation or chemoradiation therapy of NPC.

Other groups have investigated the use of posttreatment plasma EBV DNA levels in patients who have undergone radiation or chemoradiation therapy, as a biomarker of the posttreatment risk of disease recurrence (6, 7). In 2 landmark studies of patients with locally advanced NPC, the level of EBV DNA in plasma in the period immediately after radiation or chemoradiation was found to be a better prognostic marker than pretreatment EBV DNA level or clinical stage for progression-free and overall survival (6, 7). This observation was subsequently validated in a 576-patient prospective trial carried out by the Hong Kong NPC Study Group (0502 trial), where the presence or lack of detectable plasma EBV DNA after treatment was established as the most significant prognostic biomarker in NPC (8). For patients with detectable levels, the 3-year relapse-free and overall survival rates were 49% and 70%, respectively. By contrast, for patients with undetectable levels, the 3-year relapse-free and overall survival rates were 86% and 95%, respectively (P<.0001 for both endpoints). Moreover, retrospective data suggested that the elevated risk of relapse and death in patients with persistently elevated plasma EBV DNA after chemoradiation could be reduced by adjuvant chemotherapy (9).

Impact of EBV DNA Testing on the Evolving Treatment Strategies for NPC

For many years, the standard treatment for locally advanced NPC involved radiation therapy alone. However, concurrent chemoradiation became the standard of care, at least in the United States, after the results of a phase 3 randomized trial (the U.S. Intergroup 00–99 Trial), which demonstrated large-scale improvements in progression-free and overall survival resulting from the addition of concurrent cisplatin to radiation, followed by cisplatin and 5-fluorouracil (5-FU) adjuvant chemotherapy (10). Subsequently, clinical practice quickly evolved toward concurrent chemoradiation therapy followed by adjuvant chemotherapy as the standard of care for NPC (11–13). However, approximately half of patients did not tolerate adjuvant chemotherapy after chemoradiation (10, 11), and reports of benefits from adjuvant chemotherapy were inconsistent, such that the value of adjuvant chemotherapy in this clinical context has been alternatively questioned (14) and supported (15). In light of these controversies, testing NPC patients for the presence or absence of persistently elevated levels of plasma EBV DNA after chemoradiation could be valuable, in that the results of that test could be used to select patients most likely to benefit from additional chemotherapy after chemoradiation.

To test the utility of EBV DNA as a biomarker to guide the use of adjuvant chemotherapy in patients with locoergionally advanced NPC, the NRG Oncology cooperative group of the National Clinical Trial Network (NCTN) has initiated NRG-HN001 trial (NCT02135042: Randomized Phase II and Phase III Studies of Individualized Treatment for Nasopharyngeal Carcinoma Based on Biomarker Epstein Barr Virus (EBV) Deoxyribonucleic Acid (DNA)). The hypothesis of this ongoing trial is that the level of posttreatment circulating EBV DNA can be used to divide patients into low-risk and high-risk groups, where the low-risk group (ie, those with undetectable posttreatment EBV DNA in plasma) may not need adjuvant chemotherapy and the high-risk group may benefit from a more aggressive adjuvant chemotherapy regimen. On the basis of this hypothesis, patients in the trial with undetectable posttreatment EBV DNA in the plasma are randomized to either observation or standard adjuvant chemotherapy (cisplatin and 5-FU) after concomitant chemoradiation, and those with detectable levels are randomized to the standard regimen of adjuvant chemotherapy (cisplatin and 5-FU) or an experimental, non–cross-resistant active regimen (paclitaxel and gemcitabine). The goal of the NRG-HN001 trial is to develop a treatment “decision tree” that eliminates unnecessary adjuvant chemotherapy for low-risk patients while maximizing progression-free survival for high-risk patients who may derive the greatest benefit from adjuvant chemotherapy. This trial is actively enrolling patients.

Because NRG-HN001 was designed as an international biomarker-driven trial, it could yield valuable data only if the assays used to quantify EBV DNA were standardized to provide accurate and concordant results in different clinical laboratories throughout the world. Thus, a collaborative study to standardize the assay for this trial was successfully undertaken by research teams at Stanford University, National Taiwan University Hospital (Taiwan), Chang Gung Clinical Lab (Taiwan), and Chinese University of Hong Kong (Hong Kong) (16). In addition, a credentialing process was developed for the assay, whereby 2 new laboratories (the National Cancer Center Singapore and Fudan University) have been credentialed. Operationally, calibrators in quantities sufficiently large to cover the entire study were synthesized, and a bank of aliquots all from a single preparation was created. New laboratories brought on board received the assay protocol and aliquots of calibrators from this single source to ensure that the laboratories were using the same PCR protocol and identical reagents. The detailed EBV DNA measurement procedure can be accessed from from https://med.uc.edu/docs/default-source/Cancer-Institute-Docs/head-and-neck/nrg-hn001_protocol.pdf?sfvrsn=2. Interlaboratory comparisons using quality control samples are performed every 3 to 6 months. Currently, there are no plans to bring on board and credential additional laboratories for this trial.

Monitoring of midtreatment plasma EBV DNA levels during the combined course of chemotherapy and radiation therapy might prove useful in designing a chemoradiation de-escalation trial aimed at reducing treatment intensity in patients showing early response. However, data on EBV DNA detection during an initial chemoradiation treatment period or after a course of induction chemotherapy are sparse and challenging to interpret. One of the earliest studies of plasma EBV DNA detection in NPC patients described its kinetics during the radiation therapy course for 15 patients (17); an early rise in plasma EBV DNA levels within the first week of radiation therapy was seen for a subset of patients, followed by a rapid decline with a half-life of decay of <4 days. Although the prognostic significance of rapid plasma EBV DNA clearance in patients with localized or locoregional disease is not known, in the metastatic setting rapid clearance is associated with a favorable prognosis (18). A prospective study of 107 individuals showed that patients with undetectable EBV DNA after 4 weeks of radiation therapy had significantly higher relapse-free and overall survival than did those with a detectable level (19). However, midtreatment EBV DNA level correlated better with distant rather than local relapse, making it difficult to argue for reducing the intensity of locally directed chemoradiation therapy based on midtreatment EBV DNA. Another prospective study of patients with advanced-stage NPC who received induction chemotherapy before chemoradiation demonstrated correlations of detectable EBV DNA after induction with progression-free and distant metastasis-free survival, but the clinical assessment of tumor response was equally prognostic for progression-free and locoregional recurrence-free survival (20).

Other Clinical Applications of EBV DNA Testing: Staging, Prognostication, Surveillance, and Screening

In addition to its potential application in guiding the use of adjuvant chemotherapy in patients with locoregionally advanced NPC, the utility of circulating EBV DNA has been studied at other clinical stages. Although the results are conflicting and the data are limited, studies have evaluated the role of pretreatment EBV DNA in guiding the use of positron emission tomography—computed tomography (PET-CT) imaging at diagnosis (21) and in fine-tuning prognostication in tumor staging (22). However, owing to the lack of assay standardization, the cutpoints have varied across reported studies, making it difficult to compare the results across studies (23, 24). Because of these uncertainties, using pretreatment EBV DNA in these applications may need to be re-examined after a harmonized assay for EBV DNA is made available.

The role of posttreatment EBV DNA level in relapse and metastasis surveillance has also been evaluated. Multiple studies have suggested that it may be useful in detecting systemic relapse months before clinical detection by current standard surveillance or follow-up regimens (25, 26), but another study suggested that the posttreatment level may fluctuate and that the plasma DNA levels can reflect transient increases in this setting (27). A study by Hsu et al (27) showed that posttreatment plasma EBV DNA was detectable in 95% of patients with metastatic disease but in only 51% of patients with localized recurrent disease.

In a cross-sectional study in Hong Kong on 29 patients after irradiation, it was demonstrated that the level of EBV in brushed nasopharyngeal samples was higher in those with recurrent disease than in those in clinical remission (28). Another study evaluating the value of assessing the EBV DNA load in direct nasopharyngeal brushing or swab specimens also supported its potential value in detecting locally recurrent disease (29). Nasal brushing is a highly sensitive and specific method to detect early stage, small, or submucosal tumors before they are visible by endoscopy or even with magnetic resonance imaging (MRI). A recent study showed that brushing had 99% sensitivity, whereas screening by nasoendoscopy had 94% sensitivity in detecting NPC (30). The brushing technique is not technically difficult (using the transoral method, with depression of the tongue using a tongue depressor) and may be valuable for detecting early stage or small tumors. Meanwhile, plasma EBV DNA measurement would be valuable for detecting distant metastases.

To be a clinically meaningful tool in surveillance, an EBV DNA test result that is positive should remain positive and rise over time if it truly reflects the burden of tumor recurrence. This clinical situation may be akin to a rising prostate-specific antigen level after surgery or radiation therapy in prostate cancer. Prospective clinical trials are needed to address the optimal timing and frequency of EBV testing after definitive therapy. Furthermore, a specific algorithm for the treatment of a patient with a positive test result should be defined to minimize the risk of over-biopsy or unnecessary treatment.

Finally, plasma EBV DNA analysis can potentially be applied in the screening of asymptomatic individuals in endemic areas for NPC. In this regard, Chan et al (31) prospectively screened 1318 apparently healthy individuals aged 40 to 60 years in Hong Kong and identified 3 cases of NPC. Among the 3 patients identified, only 1 patient was positive for immunoglobulin A for the EBV viral capsid antigen. This finding was consistent with the previous report of the superior sensitivity of plasma EBV DNA over EBV serology in the early detection of NPC (32).

Discussion on the Clinical Utility of EBV DNA Testing

In November 2015, the National Cancer Institute (NCI) held a workshop where a group of radiation oncologists, medical oncologists, head and neck surgeons, and laboratory experts who specialize in NPC treatment or EBV DNA testing were convened to discuss the clinical utility of EBV DNA testing. The goal of the workshop was to find consensus and begin the process of developing guidelines for this assay in clinical settings. Different perspectives regarding the clinical utility of EBV DNA testing were presented. Investigators based in the United States thought that it was premature to develop guidelines for EBV DNA testing in the management of nasopharyngeal cancer patients. They stressed the need for a greater quantity and a higher level of evidence from prospective clinical trials and noted that the lack of standardization of EBV DNA quantification by PCR required further assay harmonization efforts and clinical validation studies. (For the EBV DNA test used in NRG-HN001, the detection limit was established for all involved laboratories at a concentration of 5 copies/reaction, which translated to 60 copies/mL. At 5 copies/reaction using DNA from the Namalwa cell line, the coefficient of variation for the number of PCR threshold cycles was consistently less than 10%, which is normally accepted for a clinical test, for all sites. In fact, the detection limit is lower, and laboratories can consistently detect 20 copies/mL plasma.)

Investigators from endemic NPC regions, by contrast, stressed the value of EBV DNA testing and the significant clinical impact it could have. They raised the possibility that differences in the epidemiologic and clinical characteristics of target populations in the United States and Europe versus the endemic regions in Southern China and Southeast Asia might warrant different standards for clinical practice guidelines. When challenged with the fact that not all NPC patients have detectable circulating EBV DNA, which could diminish the predictive value of this biomarker, the endemic area investigators countered that in their populations, EBV DNA can be detected in virtually all NPC cases (4, 33, 34). Those investigators advocating for clinical implementation of EBV DNA testing pointed out the value of measuring posttreatment EBV DNA levels in NPC patients who have undergone radiation therapy. The irradiated nasopharynx is characterized by radiation-induced, nonspecific findings such as erythema, edema, ulcerative changes, crusting, or nonspecific asymmetric fullness of the nasopharynx that render endoscopy, computed tomography, and MRI monitoring less effective for accurate, early detection of local recurrence. Thus, persistent or rising EBV DNA in plasma or a positive EBV DNA finding in nasopharyngeal brushing, suggestive of residual disease, could be very valuable for further monitoring of disease.

Whereas there was consensus regarding the urgent need to develop a quantitative EBV DNA test for widespread clinical use, it became evident that more studies would have to be undertaken to produce the data needed to inform the formulation of clinical guidelines. To generate valid clinical data, harmonization of EBV DNA tests and the development of appropriate procedures for widespread implementation across different clinical laboratories are required. At present, there is no criterion standard assay for the quantification of EBV DNA for clinical or analytic purposes. The EBV DNA tests used in the diagnosis and management of EBV-associated diseases are predominantly laboratory-developed tests (LDTs) (with over 30 distinct LDTs reported in the literature) that use different DNA extraction methods, different instruments, different primers and probes that target different regions of the EBV genome, different quantification controls to determine the EBV DNA concentration (eg, plasmid clone of the PCR target, quantified viral DNA or virus particles, or cells containing specific copy numbers of the EBV genome), and different units for reporting the results. Given this heterogeneity, it is difficult to compare the EBV DNA measurements between different laboratories. Clinical laboratory tests are said to be “harmonized” when the results of all tests are independent of the specific assay procedures and of where and when the assays are performed, and this process requires that all procedures be directly or indirectly calibrated to a commutable reference material using a standardized unit of measure and that all procedures measure the same analyte free of interference from all other substances within the sample.

Once a harmonized assay with robust analytic performance characteristics is established, clinical validation studies would be needed to identify optimal cutoff values for different clinical settings (eg, pretreatment vs posttreatment) and intended uses (eg, diagnosis, therapy selection, monitoring) to enable clinical decision making. Clinical studies will also be needed to establish optimal cutpoints, timing/time intervals, and frequency of testing, particularly for posttreatment circulating EBV DNA, the levels of which are likely to vary with time, in such a manner that the cutoff value applied to samples collected immediately after treatment might not be valid for samples collected months after treatment.

Finally, all of the participants acknowledged that it was critical to ensure that these efforts involve an international collaboration with research partners in China and Southeast Asia, where the relevance and interest in this field are much higher than in the United States. The investigators leading the biomarker-driven NRG-HN001 trial purposely established international collaborations for this reason, and this assay may be considered the major infrastructure in place at present. The team at Stanford University performed rigorous analytical validation and submitted the data to the U.S. Food and Drug Administration (FDA), which granted investigational device exemption status for the harmonized “BamHI W” quantitative PCR assay to be used in the NRG-HN001 trial. The Stanford laboratory acts as the North American reference site for the NRG-HN001 trial, with laboratories at National Taiwan University Hospital (Taiwan), Chang Gung Clinical Lab (Taiwan), Chinese University of Hong Kong (Hong Kong, China), National Cancer Center Singapore (Singapore), and Fudan University Shanghai Cancer Center (Shanghai, China) credentialed for international testing.

Some useful information for other investigators considering international studies is that the FDA does not regulate foreign clinical laboratories that are part of a United States–led study (for testing conducted outside the United States, on specimens from non–United States patients) and is not likely to raise issues as long as the foreign laboratories are in compliance with ISO 15189 (of the International Organization for Standardization) or are accredited by the College of American Pathologists (or, eg, by the National Association of Testing Authorities of Australia). In China, there are regulatory filing requirements for studies that require long-term and large-volume testing, but there generally are no issues for tests performed within the country for investigational or research purposes, and no issues with shipping the specimens from China to specific designated research zones in Hong Kong for testing. With respect to commercial development and regulation of commercial kit products (not LDTs), the regulatory process overseen by the China Food and Drug Administration is entirely different.