We thank Stephen Chan for his useful discussion on this study.
Nasopharyngeal carcinoma (NPC) is prevalent in Southeast Asia. Over the last decade, plasma Epstein-Barr virus (EBV) DNA has been developed as a tumor marker for NPC. In this study, the authors investigated whether plasma EBV DNA analysis is useful for NPC surveillance.
In total, 1318 volunteers ages 40 to 60 years were prospectively recruited. Plasma EBV DNA and serology for viral capsid antigen immunoglobulin A (IgA) were measured. Participants who had detectable plasma EBV DNA or positive IgA serology underwent nasal endoscopic examination and a follow-up plasma EBV DNA analysis in approximately 2 weeks. All participants were followed for 2 years to record the development of NPC.
Three individuals with NPC were identified at enrolment. All of them were positive for EBV DNA and remained positive in follow-up analysis. Only 1 of those patients was positive for EBV serology. In 1 patient who had NPC with a small tumor confined to the mucosa, the tumor was not detectable on endoscopic examination. Because of a 2-fold increase in plasma EBV DNA on the follow-up analysis, that patient underwent magnetic resonance imaging, which revealed the tumor. Among the participants who did not have NPC but had initially positive plasma EBV DNA results, approximately 66% had negative EBV DNA results after a median of 2 weeks.
Nasopharyngeal carcinoma (NPC) is 1 of the most common cancers in Southern China and Southeast Asia.1 The incidence of NPC in this region is 20 to 30 per 100,000 men and 15 to 20 per 100,000 women.1 Like many other cancers, the most important prognostic factor for NPC is the extent of disease involvement at presentation.2, 3 In patients who have localized disease, the 5-year survival probability is as great as 90%.2, 3 This proportion drops to below 50% in patients who have locoregionally advanced disease with metastatic lymph nodes. In addition, patients with more advanced disease have far more treatment-associated morbidities.3 Unfortunately, 75% to 90% of patients with NPC already have developed local or regional spread at presentation.3, 4 Therefore, screening for early disease in asymptomatic individuals potentially may improve NPC treatment outcomes.
Recently, there has been increasing interest in the detection of tumor-associated nucleic acids in plasma for cancer detection and monitoring.5-7 It has also been proposed that the detection of cancer-derived nucleic acids may be used in the screening for early cancers.8, 9 To date, plasma Epstein-Barr virus (EBV) DNA for NPC is 1 of the most well established plasma nucleic acid-based tumor markers.4, 10-18 Plasma EBV DNA analysis is particularly useful for identifying clinically concealed residual NPC after curative-intent radiotherapy13, 14 and for predicting treatment outcomes.15 The prognostic value of circulating EBV DNA analysis is independent of the clinical disease stage of NPC. Therefore, it has been proposed that plasma EBV DNA analysis may be incorporated usefully into the staging system of NPC.4 Although a previous report suggested that plasma EBV DNA analysis may be useful for the early detection of NPC,19 there is a lack of prospective information concerning the efficacy of plasma EBV DNA analysis for NPC screening.
In contrast to the extensive investigations on plasma EBV DNA in patients with NPC, little is known about the clinical significance of detectable EBV DNA in healthy individuals. It is unclear whether the low levels of EBV DNA in apparently healthy individuals would persist over a prolonged period or would fluctuate with time. If the presence of circulating EBV DNA is only a transient phenomenon, then it would be possible to discriminate a false-positive result from a true-positive result associated with NPC by repeating the EBV DNA analysis when an initial test is positive. For patients with NPC, plasma EBV DNA would be detectable persistently, because it is continuously produced by the tumor. In contrast, the follow-up test may become negative in healthy individuals if the presence of EBV DNA is only transient. Conversely, if circulating EBV DNA persists in only a subset of the population for a prolonged period, then the clinical significance of this phenomenon in relation to the risk of developing EBV-associated cancers would warrant further investigations. In this regard, a previous study demonstrated that individuals who had positive EBV serology had a 4-fold increased risk of developing NPC.20 The presence of circulating EBV DNA and EBV antibodies may reflect the pathologic processes underlying NPC development. In the current study, we investigated prospectively whether plasma EBV DNA analysis would be useful for detecting early NPC in apparently healthy individuals, and we explored the patterns of variation of plasma EBV DNA in those without any evidence of NPC.
MATERIALS AND METHODS
In total, 1318 volunteers ages 40 to 60 years were recruited with informed consent from July to November 2008 through 12 community visits. The study was approved by the ethics committee of the institution. No participants had a history of NPC, and none were receiving steroids or immunosuppressive drugs. The demographic information of the study population is summarized in Table 1. Plasma EBV DNA analysis and immunoglobulin A (IgA) serology for the EBV viral capsid antigen (VCA) were performed for each participant. Every individual who had a positive test result (either EBV DNA or IgA-VCA) underwent an endoscopic examination of the nasopharynx and a follow-up plasma EBV DNA analysis after 2 to 4 weeks. For participants who had positive plasma EBV DNA results, the plasma EBV DNA test was repeated at 6 months and 18 months after the initial screening.
Table 1. Demographic Data of the Participants in the Nasopharyngeal Carcinoma Surveillance Program
Plasma EBV DNA concentrations were measured as described previously.11 In brief, DNA was extracted from 800 μL plasma. EBV DNA concentrations were measured using a real-time polymerase chain reaction (PCR) system targeting the BamHI-W fragment region of the EBV genome. The primers were 5′-CCCAACACTCCACCACACC-3′ and 5′-TCTTAGGAGCTGTCCGAGGG-3′, and the fluorescent probe was 5′-(FAM)CACACACTACACACACCCACCCGTCTC-3′. The amplicon size of the PCR product was 76 base pairs. Because multiple copies of BamHI-W fragments are present in each EBV genome, this assay is potentially more sensitive and less susceptible to effecting the sequence variations between different strains of EBV compared with PCR assays that target a single copy gene, for example, EBNA-1, of the EBV genome.
To investigate whether plasma EBV DNA is associated with the intact virus, additional plasma samples were collected for ultracentrifugation analysis from 4 individuals who had positive EBV DNA results on 2 occasions but no evidence of NPC. Freshly collected plasma samples were centrifuged at ×100,000 g for 30 minutes using a Biofuge Stratos centrifuge (Heraeus; Thermo Fisher Scientific, Inc., Waltham, Mass), and 800 μL of the supernatant were collected for EBV DNA analysis as previously described.21 Extracted EBV DNA and viral culture from B95-8 cells were analyzed as controls for free EBV DNA fragments and viral particles.21
Serum EBV IgA-VCA was measured by assays provided through the Microplate ELISA Systems (Euroimmun AG, Mannheim, Germany) as previously described.22 In brief, microtiter plates were coated with EBV VCA purified from P3HR1 cells. After incubating with serum samples and washing, peroxidase-labeled antihuman immunoglobulin G was added. After the addition of chromogen/substrate solution, photometric measurement was made at a wavelength of 450 nm. The ratio between the extinction values of the sample and the calibrator was determined, and a ratio >1.0 was regarded as a positive result.
Correlations between plasma EBV DNA and serum IgA-VCA were determined using the chi-square test.
Plasma Epstein-Barr Virus DNA and Epstein-Barr Virus Serology Analyses at Enrolment
At enrolment, 69 of the 1318 participants (5.2%) had detectable levels of plasma EBV DNA. Plasma EBV DNA analyses were repeated for these 69 individuals after a median of 16 days (interquartile range, 16-23 days). Twenty individuals (29%) had persistently positive results in follow-up samples. In other words, 1.4% of all participants had persistently positive plasma EBV DNA results on 2 occasions. Eighty-six of 1318 participants (6.5%) were positive for IgA-VCA. Only 7 individuals were positive for both tests. There was no significant correlation between positivity for the 2 tests (P = .317; chi-square test).
To confirm or rule out NPC, all participants who were positive for plasma EBV DNA or serum EBV IgA-VCA were assessed by an otorhinolaryngologist. Nasal endoscopy was performed for each of these individuals, and magnetic resonance imaging (MRI) studies were obtained as clinically indicated. Three individuals were diagnosed with NPC, and all 3 had positive plasma EBV DNA at screening and on follow-up. In contrast, only 1 of these individuals had positive EBV IgA-VCA results.
Clinical Information for Patients With Nasopharyngeal Carcinoma
The first patient with NPC (patient 1) was a man aged 50 years who was completely free of nasal or aural symptoms but noticed a small neck lump 3 months before joining the study. He underwent a fine-needle biopsy of the neck lump, which revealed reactive changes but no evidence of malignancy. The lump persisted without changing in size. At enrolment, his plasma EBV DNA concentration was 198 copies/mL, and he was negative for EBV IgA-VCA. Nasal endoscopy was performed 16 days later and revealed a normal nasopharynx with some lymphoid tissues. No biopsy was taken. However, the plasma EBV DNA concentration, obtained on the day of endoscopy, increased to 415 copies/mL. Because of this finding, an MRI image was obtained and revealed a small tumor in the nasopharynx together with multiple, small lymph nodes in the neck (Fig. 1). Endoscopic examination was repeated, and random biopsies were taken. Histologic examination confirmed the presence of an undifferentiated NPC. Further workup indicated that the patient had T1N2M0 disease according to the International Union Against Cancer-American Joint Committee on Cancer TNM classification. The patient received radiotherapy with concurrent chemotherapy and remained in clinical remission 2 years later with negative plasma EBV DNA in all post-treatment samples. The changes in his EBV DNA concentration are illustrated in Figure 2 (top).
The patient 2 was a woman aged 46 years who was completely asymptomatic. Her plasma EBV DNA concentration at recruitment was 28 copies/mL, and her EBV IgA-VCA test was negative. Nasal endoscopy revealed a highly suspicious lesion, which was confirmed as an undifferentiated NPC on histologic examination. Her plasma EBV DNA concentration increased to 88 copies/mL 30 days later (Fig. 2, middle). The patient was classified with T1N0M0 disease. She received radiotherapy and was in clinical remission 2 years later.
The patient 3 was a man aged 41 years who was a chronic smoker. He had a family history of NPC, and his father had died of the disease. He was completely asymptomatic. He was positive for EBV DNA on 2 occasions separated by 15 days and was also positive for EBV IgA-VCA. Nasal endoscopic examination revealed a small NPC. The patient was classified with T1N1M0 disease. He received radiotherapy and remained in clinical remission 2 years after treatment. In this patient, an initial drop in EBV DNA concentration was observed before treatment was started (Fig. 2, bottom).
The rate of plasma EBV DNA increase was determined for the 3 patients by plotting the logarithm of their plasma EBV DNA concentrations against time for samples that were obtained before treatment (Fig. 3). The slopes of the regression lines were then determined. For the third patient with NPC described above, the result of the first time point was not used to calculate the slope. The doubling time (Tdb) was then determined using the formula Tdb = log2/slope. The doubling time of plasma EBV DNA for the patients 1, 2, and 3 was 35 days, 23 days, and 27 days, respectively.
Plasma EBV DNA analysis at 6 months and 18 months
All participants who had positive plasma EBV DNA results at enrolment were scheduled for further plasma EBV DNA analyses at 6 months and 18 months to investigate fluctuations in plasma EBV DNA levels among individuals without NPC. Apart from the 3 patients who had been diagnosed with NPC, 66 participants were positive for plasma EBV DNA at enrolment. There was no significant difference in the plasma DNA level at enrolment between the 66 individuals who had false-positive results and the 3 patients who had NPC (P = .440; Student t test). At 6 months after enrolment, plasma EBV DNA levels were determined in 64 individuals, and 21 (33%) were positive. At 18 months, the test was repeated in 63 individuals, and 19 (30%) were positive. Among the 63 individuals who completed the test on the 3 occasions, 11 (17%) were positive in all 3 analyses. Each individual who had a positive plasma EBV DNA result on the follow-up visits at 6 months or 18 months underwent a follow-up nasal endoscopic examination. All of these individuals were negative for NPC. All participants in the study were interviewed 2 years after enrolment. Apart from the 3 patients who had NPC identified at enrolment, none of the participants had developed NPC.
Ultracentrifugation of Plasma Epstein-Barr Virus DNA
Extra plasma samples were collected from participants without NPC who had positive plasma EBV DNA results on 2 or more occasions for ultracentrifugation analysis. Because the freezing and thawing of plasma samples may disrupt the integrity of the viral particles and could affect the results of this experiment,21 all plasma samples that subjected to ultracentrifugation analysis were freshly collected. Although we collected plasma samples from 20 participants without NPC for the ultracentrifugation experiments, only 4 samples were subsequently identified as positive for EBV DNA without centrifugation. Thus, only those 4 samples would be informative regarding the effect of ultracentrifugation on plasma EBV DNA. In a previous study, we demonstrated that EBV DNA associated with viral particles could be pelleted down by ultracentrifugation, but the free EBV DNA fragments in patients with NPC would remain in the supernatant.21 After ultracentrifugation, EBV DNA was not detectable in the supernatant in all 4 samples that underwent ultracentrifugation. For controls, we analyzed plasma from 5 NPC patients who were recruited from the oncology clinic. After centrifugation, a median of 48% (range, 19%-83%) of the EBV DNA was still detectable in the supernatant. These findings suggest that the circulating EBV DNA in individuals without NPC probably was associated with viral particles.
In this study, we demonstrated that plasma EBV DNA analysis is sensitive for the detection of early and clinically silent NPC. For the 3 patients who had NPC identified, 2 were completely asymptomatic. Although the other patient presented with a solitary lymph node before enrolling onto the study, histologic examination of the lymph nodes revealed only benign reactive changes but no evidence of malignancy. Therefore, without participating in this surveillance program, it is likely that these 3 patients would have presented later with more advanced disease when their clinical symptoms became more overt. Plasma EBV DNA analysis appears to be more sensitive than IgA-VCA analysis for detecting early NPC. Although all 3 patients with NPC were positive for plasma EBV DNA, but only 1 was positive for IgA-VCA, the results suggest that the addition of a serology test may not be able to improve the sensitivity of NPC screening. This is consistent with previous reports that IgA-VCA is less sensitive than EBV DNA for detecting clinically confirmed, early stage NPC.12 In a previous study of a cohort that received a health check-up in a private medical center, 1 patient had NPC identified by plasma EBV DNA analysis.19 However, in that retrospective study, the reasons for referring participants for plasma EBV DNA were unclear. In addition, not all participants who had positive plasma EBV DNA results underwent nasal endoscopy examination, and those who had negative results were not followed. Consequently, the efficacy of EBV DNA in identifying early NPC could not be properly evaluated.19 In the current study, apart from the 3 patients who had NPC identified by EBV DNA analysis, we determined that all of the remaining participants were free of NPC at 24 months. This further supports the finding that plasma EBV DNA analysis at enrolment sensitively identified all patients with NPC who otherwise would have presented in the next 2 years. This may explain why the number of patients who had NPC identified was greater than expected from the prevalence of NPC in the population.
In our 3 patients with NPC, plasma EBV DNA was positive at the initial test and at all subsequent follow-up tests before treatment. In the third patient, an initial drop in the plasma EBV DNA level was observed. One possible explanation for this initial reduction is that, in addition to the contribution from the tumor tissues, the nontumor tissues also released some EBV DNA into the plasma, as in the participants without NPC. Although the contribution of the nontumor tissues was only transient, a reduction in the EBV DNA level was observed in the second sample. Because of the potential contribution from nontumor tissues, the result from the first time point was not used to calculate the doubling time of plasma EBV DNA. Overall, the plasma EBV DNA concentration increased until the commencement of treatment, with doubling time ranging from 23 days to 35 days. Because it has been demonstrated that the concentration of plasma EBV DNA reflects NPC tumor load in a linear manner, this observation suggests that the tumor load would double every 3 to 5 weeks.11, 13, 23, 24 In addition, because the presence of plasma EBV DNA in a proportion of healthy individuals is only transient, we may be able to differentiate patients with NPC from individuals who have false-positive results by repeating the plasma EBV DNA analysis. Patients with NPC would be positive in the follow-up test, but approximately 66% of the individuals with false-positive results would become negative. Using this strategy, only 1.2% would require an endoscopic examination to confirm or exclude NPC.
EBV DNA analysis can also be used to stratify patients for additional investigations. In 1 of our patients with NPC, the first nasal endoscopic examination failed to identify the tumor. However, prompted by the rising plasma EBV DNA concentrations, an MRI examination was performed and revealed a small mucosal tumor in the nasopharynx. Although it has been demonstrated that MRI is more sensitive than nasal endoscopy for detecting small and infiltrating NPC,25, 26 performing MRI on all patients with a suspicion of NPC is not practical. In this regard, plasma EBV DNA analysis would be useful for identifying high-risk individuals for MRI examination.
Using ultracentrifugation analysis, we demonstrated that plasma EBV DNA levels in a proportion of individuals without NPC were likely to be associated with viral particles instead of existing as free DNA fragments, as in patients with NPC.21 Therefore, the incorporation of strategies like ultracentrifugation of plasma may further reduce the false-positive rate. Although the presence of EBV DNA in plasma was only a transient phenomenon in the majority of our participants without NPC, 17% of individuals had persistently detectable levels of EBV DNA in plasma over an 18-month period. In a previous study, it was demonstrated that individuals who had positive EBV IgA-VCA results and the presence of neutralizing antibodies against EBV-specific DNase had an increased risk of developing NPC in the future.20 The relative risk of NPC was 4 for individuals who had 1 positive serologic marker compared with individuals who had negative results, and the risk increased to 33-fold when both markers were positive.20 Further studies will be needed to address whether individuals who have persistently positive EBV DNA results have an increased risk of developing NPC. If such a correlation is established, then plasma EBV DNA analysis also may be used to identify high-risk individuals for more regular follow-up in addition to screening for early disease.
In this study, we have demonstrated that plasma EBV DNA is useful in screening for early NPC. In addition, through our analysis of variations in plasma EBV DNA levels among individuals without cancer, we have identified ways to enhance the cost-effectiveness of carrying out NPC screening using plasma EBV DNA analysis. In endemic areas, the incidence of NPC for high-risk individuals can be up to 30 per 100,000.1, 4 The cost for each plasma EBV DNA analysis is approximately US$25, and the cost for nasal endoscopic examination is approximately US$50. Assuming that follow-up EBV DNA testing and endoscopic examination would be required in 5% and 1.5% of the screened individuals, respectively, the total cost for carrying out a screening program for 100,000 individuals would be US$2.7 million. Assuming the test has a sensitivity of 90% for detecting NPC, the cost of identifying a patient with NPC through screening would be approximately US$100,000 in high-risk individuals. Considering that NPC is relatively more prevalent in younger age groups (peak incidence is at approximately aged 50 years) and the very good prognosis of early stage disease after treatment,1, 2 NPC screening using plasma EBV DNA analysis appears to be both feasible and cost-effective in regions with a high incidence of NPC. The cost-effectiveness of screening can be enhanced further by targeting individuals who have an even greater risk of NPC, such as middle-aged men and those with a family history of NPC.
This work was supported by the Sir Michael and Lady Betty Kadoorie Cancer Genetics Research Program.
CONFLICT OF INTEREST DISCLOSURES
Y. M. Dennis Lo is a consultant to and holds equities in Sequenom.