**3.1 Plasma EBV protein in the diagnosis of NPC**

84 Carcinogenesis, Diagnosis, and Molecular Targeted Treatment for Nasopharyngeal Carcinoma

patient's geographical origin and is observed in World Health Organization (WHO) types II and III. However, the association of NPC type I with EBV has long been a matter of controversy. It is now clear that WHO type I tumors are frequently associated with EBV in endemic regions, but not in non-endemic regions, where they often result from tobacco and alcohol abuse (Nicholls et al., 1997). Types II and III may be accompanied by an

inflammatory infiltrate of lymphocytes, plasma cells, and eosinophils.

Fig. 1. EBV in the pathogenesis model for nasopharyngeal carcinoma

More than 95% adults in all ethnic groups across the world are healthy carriers of EBV (Ooka et al., 1991). This means that NPC oncogenesis is not simply a consequence of EBV infection. It probably results from a form of viral reactivation in combination with other events, such as cellular genetic lesions due to environmental carcinogens and/or some form of immune defects. EBV-encoded RNA signal (EBER) has been shown, by in-situ hybridization, to be present in nearly all tumor cells, whereas EBV-encoded RNA is absent from the adjacent normal tissue, except perhaps for a few scattered lymphoid cells. Premalignant lesions of the nasopharyngeal epithelium have also been shown to harbor EBV, which suggests that the infection occurs in the early phases of carcinogenesis (Gulley, 2001). Consistent with this hypothesis, is the fact that NPC generally occur several years after EBV primary infection. The expression of EBV latent genes provides growth and survival advantages to these infected cells, ultimately leading to the development of NPC. Further genetic and epigenetic alterations post-NPC development can occur, which may result in a more metastatic disease. Because of it takes years for premalignant lesion after Serological studies have shown that the clinical onset of NPC is preceded by the appearance of a high titer of various EBV antibodies such as viral capsid antigens (VCA) IgA(Li et al., 2010; Ng et al., 2010; O et al., 2007), anti-EBV DNase(Chien et al., 2001) and combined EBV EA(early antigen) + EBNA-1 (Nuclear antigen 1) IgA test(Chang et al., 2008)( **Table 1.**).


IF: immunofluorescent assay; ELISA: enzyme-linked immunosorbent assay

Table 1. Reported plasma EBV antibody in NPC diagnosis and risk assessment of NPC

For NPC serodiagnosis, cell-based indirect immunofluorescent assay (IF) methods are widely considered the gold standard (Paramita et al., 2009). IF involves the separate analysis of antibody responses to VCA, EA, and EBNA, and requiring different cell lines for specific analysis. However, this method shows considerable variation among laboratories and is time-consuming, subjective, and not suitable for large-scale automatic handling. Enzymelinked immunosorbent assay (ELISA) techniques are increasingly used recently and have shown a better sensitivity and specificity compared to IFA and which are suitable for largescale application. Stage is one of the most important prognostic predictors of NPC. NPC in its early stages are highly curable with radiotherapy. Screening may change the distribution of stage and prognosis. In a cohort study undertaken in Wuzhou (Guangxi province, China) in the early 1980s (Zeng et al., 1985), total 1136 individuals identified as positive for Ig A

Epstein-Barr Virus Serology in the Detection and Screening of Nasopharyngeal Carcinoma 87

CI: 0.97–0.99, **Fig. 3.**). There are limitations of this meta-analysis: (1) All of the included studies were Chinese articles, which would lead to language bias. (2) Significant heterogeneity exists in this metaanalysis, even with random effect model, pooling is not a

Fig. 3. Forest plot of estimates of sensitivity and specificity for VCA-IgA in the diagnosis of NPC. The point estimates of sensitivity from each study are shown as solid squares.

Sustained elevation of EBV antibody had a high possibility of NPC formation. Ji monitored serologically and clinically 39 cases for different periods of up to 15 years before NPC was diagnosed, and assessed the preclinical serologic status of another 68 cases(Ji et al., 2007). The results identify a serologic window preceding diagnosis when antibody levels are raised and sustained. This window can persist for as long as 10 years, with a mean duration

The presence of circulating DNA was first reported by Mandel and Metais in 1948 (Chan and Lo, 2002). They demonstrated that extracellular DNA and RNA could be detected from the blood of healthy as well as sick individuals. The blood plasma EBV DNA load was shown to be proportionately related to the presence of malignant disease (Tan et al., 2006). While the EBV copy number in untreated NPC patients had a median of 2,043 copies/ml, viral load in plasma of healthy controls was significantly lower (median of 0 copy/ml). The

proper method.

Error bars are 95% *CI*.

estimated to as 37 months.

**3.2 Circulating EBV DNA in the diagnosis of NPC** 

against VCA received regular clinical examinations of the nasopharynx and neck for 4 years. During this follow-up period, 35 cases of NPC were detected, most of which (92%) were diagnosed early at either stage I or stage II. The annual detection rate of NPC for this group was 31·7 times higher than for the population as a whole. Distribution of different stages of screen-detected and symptomatic NPC in Hong Kong (Ng et al., 2010) also revealed early diagnosis of NPC in screen-detected patients, comparing to symptomatic NPC (**Table 2.**).


Table 2. The comparison of the stage distribution between screen-detected and clinically detected based on published data on NPC screening during 1979–1992 (Ng et al., 2010).

Li conducted a systematic review of studies in Chinese on the accuracy of VCA-IgA concentrations in the diagnosis of NPC using random effects models (Li et al., 2010). Twenty studies met the inclusion criteria for the meta-analysis. The summary estimates (**Fig. 2.**) for VCA-IgA in the diagnosis of NPC were: sensitivity 0.92 (95% confidence interval (CI): 0.89– 0.95), specificity 0.98 (95% CI: 0.95–0.99), positive likelihood ratio 38.5 (95% CI: 19.0–78.0), negative likelihood ratio 0.08 (95% CI: 0.05–0.12) and diagnostic odds ratio 487 (95% CI: 224– 1059). The area under the summary receiver operating characteristic curves was 0.98(95%

Fig. 2. Summary receiver operating characteristic curves for VCA-IgA in the diagnosis of NPC. Each solid circle represents each study in the meta-analysis.

against VCA received regular clinical examinations of the nasopharynx and neck for 4 years. During this follow-up period, 35 cases of NPC were detected, most of which (92%) were diagnosed early at either stage I or stage II. The annual detection rate of NPC for this group was 31·7 times higher than for the population as a whole. Distribution of different stages of screen-detected and symptomatic NPC in Hong Kong (Ng et al., 2010) also revealed early diagnosis of NPC in screen-detected patients, comparing to symptomatic NPC (**Table 2.**).

Screen-detected 41.2% 17.6% 35.3% 5.9% Clinically detected 0.7% 23.2% 36.3% 39.5% Table 2. The comparison of the stage distribution between screen-detected and clinically detected based on published data on NPC screening during 1979–1992 (Ng et al., 2010).

Li conducted a systematic review of studies in Chinese on the accuracy of VCA-IgA concentrations in the diagnosis of NPC using random effects models (Li et al., 2010). Twenty studies met the inclusion criteria for the meta-analysis. The summary estimates (**Fig. 2.**) for VCA-IgA in the diagnosis of NPC were: sensitivity 0.92 (95% confidence interval (CI): 0.89– 0.95), specificity 0.98 (95% CI: 0.95–0.99), positive likelihood ratio 38.5 (95% CI: 19.0–78.0), negative likelihood ratio 0.08 (95% CI: 0.05–0.12) and diagnostic odds ratio 487 (95% CI: 224– 1059). The area under the summary receiver operating characteristic curves was 0.98(95%

Fig. 2. Summary receiver operating characteristic curves for VCA-IgA in the diagnosis of

NPC. Each solid circle represents each study in the meta-analysis.

Stage I Stage II Stage III Stage IV

CI: 0.97–0.99, **Fig. 3.**). There are limitations of this meta-analysis: (1) All of the included studies were Chinese articles, which would lead to language bias. (2) Significant heterogeneity exists in this metaanalysis, even with random effect model, pooling is not a proper method.

Fig. 3. Forest plot of estimates of sensitivity and specificity for VCA-IgA in the diagnosis of NPC. The point estimates of sensitivity from each study are shown as solid squares. Error bars are 95% *CI*.

Sustained elevation of EBV antibody had a high possibility of NPC formation. Ji monitored serologically and clinically 39 cases for different periods of up to 15 years before NPC was diagnosed, and assessed the preclinical serologic status of another 68 cases(Ji et al., 2007). The results identify a serologic window preceding diagnosis when antibody levels are raised and sustained. This window can persist for as long as 10 years, with a mean duration estimated to as 37 months.

#### **3.2 Circulating EBV DNA in the diagnosis of NPC**

The presence of circulating DNA was first reported by Mandel and Metais in 1948 (Chan and Lo, 2002). They demonstrated that extracellular DNA and RNA could be detected from the blood of healthy as well as sick individuals. The blood plasma EBV DNA load was shown to be proportionately related to the presence of malignant disease (Tan et al., 2006). While the EBV copy number in untreated NPC patients had a median of 2,043 copies/ml, viral load in plasma of healthy controls was significantly lower (median of 0 copy/ml). The

Epstein-Barr Virus Serology in the Detection and Screening of Nasopharyngeal Carcinoma 89

Fig. 5.Summary receiver operating characteristic curves for VCA-IgA in the diagnosis of

With the demonstration of the presence of extracellular EBV DNA in the circulation of NPC patients and its disappearance following radiotherapy, the levels of plasma/serum EBV DNA in patients with NPC recurrence were much higher than the levels of those who remained in continuous clinical remission. A significant decrease in EBV load was observed in patients who had undergone radiotherapy while a high viral load indicated in patients correlated to tumor relapse and presence of distant metastasis upon clinical investigation(Lo et al., 1999; Tan et al., 2006). The median EBV DNA concentration in the relapsed patients was 32, 350 copies/ml, whereas that in patients in remission was 0 copy/ml (Lo et al., 1999). Plasma EBV DNA is superior to serum EBV VCA antibodies in prognostic predictions and monitoring for NPC (Twu et al., 2007). Relapsed patients had significantly higher pretreatment EBV DNA concentration than patients without relapse (p<0.05). No associations of VCA-IgA (p =0.97) or VCA-IgG (p=0.61) were observed between patients

**3.4 Comparison of DNA and serology in the diagnosis and follow-up of NPC** 

Plasma EBV DNA detection has a similar sensitive and specific as the serum IgA antibody titer for the diagnosis of patients with NPC. In terms of real costs, however, the IgA test costs approximately \$20, whereas the DNA test ranges from \$100 to \$200 (5-10 times more)(O et al., 2007). For screening purposes, the serologic assays may be made first, so that more patients can be evaluated. If a patient is assessed to be at high risk, then EBV DNA

NPC. Each solid circle represents each study in the meta-analysis.

**3.3 EBV DNA in the monitoring of NPC** 

with and without relapse (Twu et al., 2007).

demonstration of EBV DNA in the plasma/serum of patients suffering from NPC has provided us with a new tool for NPC detection and monitoring.

Liu reported a systematic review of 15 studies in English on the accuracy of EBV-DNA in the diagnosis of NPC (Liu et al., 2011). NPC could be diagnosed by detecting plasma or serum EBV DNA. Pooling using random effects models showed that EBV DNA detection was also highly sensitive and specific for cancer detection, and can possibly help the clinician to diagnose NPC. The summary estimates (**Fig. 4.**) for DNA in the diagnosis of NPC were: sensitivity 0.92 (95% confidence interval (CI): 0.82–0.96), specificity 0.88 (95% CI: 0.78–0.94), positive likelihood ratio 7.7 (95% CI: 4.1–14.6), negative likelihood ratio 0.09 (95% CI: 0.05–0.15) and diagnostic odds ratio 89 (95% CI: 44–181). The area under the summary receiver operating characteristic curves was 0.96(95% CI: 0.94–0.97, **Fig. 5.**).

Fig. 4. Forest plot of estimates of sensitivity and specificity for EBV DNA in the diagnosis of NPC. The point estimates of sensitivity from each study are shown as solid squares. Error bars are 95% CI.

The presentation with "occult primary" of NPC is a diagnostic challenge for the clinician. Detection of EBV DNA by PCR in metastatic neck nodes has a good diagnostic rate (97.1%) (Yap et al., 2007). PCR is an ideal tool for suggesting occult primary NPC and guiding the diagnostic workup, facilitating earlier diagnosis and reducing morbidity and mortality. It is, therefore, expected that this promising molecular tumor marker would soon be incorporated into routine clinical use.

Fig. 5.Summary receiver operating characteristic curves for VCA-IgA in the diagnosis of NPC. Each solid circle represents each study in the meta-analysis.

#### **3.3 EBV DNA in the monitoring of NPC**

88 Carcinogenesis, Diagnosis, and Molecular Targeted Treatment for Nasopharyngeal Carcinoma

demonstration of EBV DNA in the plasma/serum of patients suffering from NPC has

Liu reported a systematic review of 15 studies in English on the accuracy of EBV-DNA in the diagnosis of NPC (Liu et al., 2011). NPC could be diagnosed by detecting plasma or serum EBV DNA. Pooling using random effects models showed that EBV DNA detection was also highly sensitive and specific for cancer detection, and can possibly help the clinician to diagnose NPC. The summary estimates (**Fig. 4.**) for DNA in the diagnosis of NPC were: sensitivity 0.92 (95% confidence interval (CI): 0.82–0.96), specificity 0.88 (95% CI: 0.78–0.94), positive likelihood ratio 7.7 (95% CI: 4.1–14.6), negative likelihood ratio 0.09 (95% CI: 0.05–0.15) and diagnostic odds ratio 89 (95% CI: 44–181). The area under the summary

Fig. 4. Forest plot of estimates of sensitivity and specificity for EBV DNA in the diagnosis of NPC. The point estimates of sensitivity from each study are shown as solid squares. Error

The presentation with "occult primary" of NPC is a diagnostic challenge for the clinician. Detection of EBV DNA by PCR in metastatic neck nodes has a good diagnostic rate (97.1%) (Yap et al., 2007). PCR is an ideal tool for suggesting occult primary NPC and guiding the diagnostic workup, facilitating earlier diagnosis and reducing morbidity and mortality. It is, therefore, expected that this promising molecular tumor marker would soon be

bars are 95% CI.

incorporated into routine clinical use.

provided us with a new tool for NPC detection and monitoring.

receiver operating characteristic curves was 0.96(95% CI: 0.94–0.97, **Fig. 5.**).

With the demonstration of the presence of extracellular EBV DNA in the circulation of NPC patients and its disappearance following radiotherapy, the levels of plasma/serum EBV DNA in patients with NPC recurrence were much higher than the levels of those who remained in continuous clinical remission. A significant decrease in EBV load was observed in patients who had undergone radiotherapy while a high viral load indicated in patients correlated to tumor relapse and presence of distant metastasis upon clinical investigation(Lo et al., 1999; Tan et al., 2006). The median EBV DNA concentration in the relapsed patients was 32, 350 copies/ml, whereas that in patients in remission was 0 copy/ml (Lo et al., 1999). Plasma EBV DNA is superior to serum EBV VCA antibodies in prognostic predictions and monitoring for NPC (Twu et al., 2007). Relapsed patients had significantly higher pretreatment EBV DNA concentration than patients without relapse (p<0.05). No associations of VCA-IgA (p =0.97) or VCA-IgG (p=0.61) were observed between patients with and without relapse (Twu et al., 2007).

#### **3.4 Comparison of DNA and serology in the diagnosis and follow-up of NPC**

Plasma EBV DNA detection has a similar sensitive and specific as the serum IgA antibody titer for the diagnosis of patients with NPC. In terms of real costs, however, the IgA test costs approximately \$20, whereas the DNA test ranges from \$100 to \$200 (5-10 times more)(O et al., 2007). For screening purposes, the serologic assays may be made first, so that more patients can be evaluated. If a patient is assessed to be at high risk, then EBV DNA

Epstein-Barr Virus Serology in the Detection and Screening of Nasopharyngeal Carcinoma 91

subsequent NPC development. Screening for high risk multiplex family may be more cost-

Screening in Hong-Kong with positive NPC family history was conducted since 1994 (Lee et al., 2005; Ng et al., 2010; Ng et al., 2005). Participants in this screening program for NPC were all first degree relatives of patients with NPC and they were all 18 years of age or older. Participants were offered annual assessment serological test of EBV and endoscopic examination of the nasopharynx. Between 1994 and 2005, total 1,199 asymptomatic family members of NPC patients were recruited and reported (Ng et al., 2010). Eighteen participants of the screening program developed NPC; 16 of them were detected in the screening. Stage distributions and survival outcomes of the 17 cases were compared with that of 1,185 consecutive symptomatic patients diagnosed in the same period through general referral. It was found that the screening program resulted in early detection of cancer (Table 2), with 59% presenting at early stage (stage I: 41%, stage II: 18%) compared to 24% (stage I: 1%, stage II: 23%) of symptomatic cancers (P=0.001), and a significant improvement in disease-free survival (P = 0.04). The cancer specific survival and overall

survival rate at 5-year are also higher (92 vs. 77% and 92 vs. 70%, respectively).

screening of family members of NPC patients.

Reduction in disease

\* Relative to strategy A Strategy A. Annual screening Strategy B. Biennial screening Strategy C. Triennial screening

(Choi et al., 2011).

participants are tested EBV positive

**4.2 Screening strategies of NPC among individual from high risk NPC families** 

Total screens 77,652 41,837 29,898 44,618 - Positive EBV test 14,962 8,071 5,772 11,413 - Screen detected case 47 42 38 47 - Disease pick up rate 88.2% 78.6% 70.8% 87.4% -

pick up rate\* - 9.6% 17.3% 0.8% - Screen missed cases 7 12 16 7 - 5-year overall survival 80.1% 78.8% 77.7% 80.0% 67.9%

Strategy D. Triennial screening for participants tested EBV negative and annual screening once the

Table 3. Simulated screening result with the four strategies based on an imaginary population of 6,000 participants follow up for 12 years with family history of NPC

The efficacy of any screening strategy should be evaluated before putting it into practice. Based on Markov chain models, Choi simulated and compared the outcomes of 4 screening strategies over a period of 12 years: (A) Annual screening, (B) biennial screening, (C) triennial screening, and (D) triennial screening for participants tested EBV negative and annual screening once the participants are tested EBV positive (Choi et al., 2011) . The result is summarized at **Table 3**, strategy A (screening annually) yields the maximum disease pickup rate, but strategy D (triennial screening for participants tested EBV negative and annual screening once the participants are tested EBV positive) offered the highest efficacy for NPC

Strategy A Strategy B Strategy C Strategy D No screening

effective.

PCR can be performed in a series fashion (O et al., 2007). If this work-up is positive; then the patient should be referred to a specialist for further work-up including fiberoptic nasopharyngoscopy or MRI examination. The economic cost of providing this screening is reduce by prescreening to exclude the low-risk patients and performing series testing on the at-risk patients.
