The prognostic value of IgA/[EBNA1+VCA-p18] on survival of nasopharyngeal cancer patients

Department of Internal Medicine, Division of Hematology and Medical Oncology, Faculty of Medicine Universitas Gadjah Mada – Dr. Sardjito Hospital, Jl Kesehatan No. 1 Yogyakarta, DIY, Indonesia. Department of Histology and Cell Biology, Faculty of Medicine, Universitas Gadjah Mada, Yogyakarta, Indonesia Department of Ear, Nose, Throat, Dr Sardjito Hospital-Faculty of Medicine, Universitas Gadjah Mada, Yogyakarta, Indonesia. Department of Radiotherapy, Dr Sardjito Hospital – Faculty of Medicine, Universitas Gadjah Mada, Yogyakarta, Indonesia. Department of Pathology, Faculty of Medicine, Universitas Gadjah Mada, Yogyakarta, Indonesia.

.. In Indonesia, NPC is the fourth commonest cancer in men and sixth among women (Adham, 2012).More than 90% of NPC in Indonesia comprised of histology type World Health Organization (WHO) III or undifferentiated carcinoma which is strongly correlated with Epstein-barr virus (EBV) infection (Soeripto, 1997;Adham et al., 2012).
Consistent expression of EBV gene products in nasopharyngeal cancer cells, specific immune response to EBV antigen in NPC patients, as well as detection of EBV in premalignant lesion support the pathogenic role of EBV in NPC (Henle et al.,1970;Wolf et al., 1973;Henle, 1976;Ho et al., 1976;Zeng et al 1982Zeng et al , 1983;;Yeung et al., 1993;Sam et al., 1994;Pathmanathan et al., 1995;Gulley, 2001;Chien et al., 2001;Middeldorp et al., 2003).Although NPC is sensitive to radiotherapy and chemotherapy, recurrence rate of NPC during the first 2year post treatment remains high (2-year progression free survival is less than 53% with median time to progression is 17.4 month) in our centre (Taroeno-Hariadi et al., 2005 unpublished observation).Wildeman et al. (2013) reported that median overall survival of NPC patients in our centre is 21 months (95% CI = 18 to 35) from day of diagnosis.Treatment modality, tumor stage, patient performance status, viral load and viral reactivation may influence recurrence and progression (Farias et al., 2003;Twu et al., 2007;Sham and Choy, 2010;Wu et al., 2012).
IgG (and specifically IgA) response to EBV antigens is the hallmark of NPC (Henle et al., 1970;Henle and Henle, 1976;Ho et al., 1976).With the advent of polymerase chain reaction (PCR) technology, nowadays viral reactivation can be measured more directly by detecting EBV-DNA.EBV-DNA quantification has been reported as sensitive and specific method for NPC diagnosis, treatment monitoring and prognosis (Lo et al., 2000).However, this method is quite expensive to be applied in low income countries such as Indonesia.Fachiroh et al. (2006) have developed serodiagnostic tools based on enzyme-linked immunosorbent assay (ELISA) to measure IgA antibody response to combination of EBV immunodominant epitopes  in one assay to diagnose NPC.
This method had a reported sensititivity of 85.4% and specificity of 90.1%.Sensitivity, a specificity of IgA/[EBNA1+VCA-p18] is better than either IgA EBNA-1 or IgA VCAp-18 alone (Fachiroh et al., 2006).The application of this serodiagnostic tool to predict survival or recurrence in NPC requires further clinical evidence.This is the first study reporting the potential role of IgA/[EBNA1+VCA-p18] as predictor of progression or survival.

MATERIALS AND METHODS
This preliminary prospective study was held in Dr Sardjito Hospital Yogyakarta Indonesia from January, 2007 to October, 2010 and included all newly diagnosed, locally advanced NPC.Diagnosis of NPC was confirmed by histology examination and clinical staging was performed by a Multi Slice Computed Tomography scan of head and neck region for primary tumor, by abdominal ultrasonography, chest x-ray and skeletal survey for detecting metastases.Patients with stage III, IVA and IVB as designated by American Joint Commitee of Cancer (AJCC), 7th edition, aged above 18 year-old and have performance status of WHO 0, 1, 2, 3, were included in this study.Patients had normal complete blood count and blood chemistry results as requirements to receive chemoradiation with weekly low dose cisplatin (40 mg/m 2 ) for 8 cycles concurrently with radiotherapy for 70 Gy in 35 fractions.Hemoglobin was > 10 g/dl, white blood cell (WBC) count > 4,000/L or absolute neutrophil count > 1,500/L, platelet > 100,000/mmk and creatinin clearance ≥50 ml/mnt.Alanine aminotransferase (ALAT) or aspartate aminotransferase (ASAT) ≤ 2 × upper limit of normal and bilirubin ≤ 2 × upper limit of normal.
Patients who received less than 80% of planned treatment, patients with severe infection or co-morbid ilnesses were excluded from the study.Plasma samples for IgA/[EBNA1+VCA-p18] ELISA were taken at pre-treatment, at the time of tumor assessment (12weeks post-treatment) and at 12 months after treatment completion, or at the time of disease progression, whichever came first.Treatment responses were assessed at 12-weeks after treatment completion.Responses were catagorized as: complete response (CR) = disappearance of all target lesions (any pathological lymph nodes must have reduction in short axis to < 10 mm), partial response (PR) = at least a 30% decrease in the sum of diameters of target lesions, stable disease (SD) = neither sufficient shrinkage to qualify for PR nor sufficient increase to qualify for PD and progression disease (PD) = at least 20% increase in the sums of diameters of the target lesions, according to Response Evaluation Criteria in Solid Tumors (RECIST) criteria version 1.1.

IgA/[EBNA1+VCA-p18] ELISA
The method was described in previous publication (Fachiroh et al., 2006) by the use of EBNA1 and VCA-p18 peptides (Cyto-Barr, Zuidhorn, The Netherlands, kindly provided through KWF funding) in one ELISA well.All OD 450 values were normalized by subtracting the value for 1:100-diluted EBV-negative sera used in duplicate in each ELISA run.The receiver operating characteristic analysis was done to predict the cut-off value of IgA, giving best sensitivity and specificity to predict progression (Fachiroh et al., 2006).

Statistical analysis
IgA/[EBNA1+VCA-p18]-level at pre-treatment, at 12-weeks posttreatment, and at 12-months post-treatment or at the time of disease progression were calculated as mean of OD 450 ± standard deviation and grouped according to the treatment responses.The difference of mean IgA according to treatment response were analyzed with student t-test.Association of IgA reactivity and treatment response were analyzed with χ 2 test.Kaplan-Meir plots of overall survival and event-free survival were established for patients group of different serological groups.Log rank tests were performed to assess survival probabilities between patients subsets (high risk serological group versus low risk).

Ethics
The study protocol was reviewed and approved by the institutional review board of the Faculty of Medicine, Universitas Gadjah Mada, and all patients were required to fill in written informed consent before participation.

Characteristic of subjects
Forty six patients were eligible for this study as shown as Table 1.Most of them were men at productive age, with moderate performance status.Patients were characterized by larger tumor size (60.8%) and extensive neck lymphnodes involvement (56.5%).Ninety five percent subjects completed their treatment according to protocol.Pre-treatment serology data were available from all subjects, while in post treatment, serology data was missing for 13.3% patients due to early progression before the scheduled sample collection.Treatment response could be assessed in 95.6% patients.Twenty one patients achieved complete response (47.7%), 19 patients were in partial response (43.2%), 2 patients were stable (4.5%), and 2 patients (4.5%) were in disease progression.A follow-up was done during 36 months.Median time to progression was 10.81 ± 11.8 months, with 27 events of death or progression during follow up.

Pre-treatment IgA/[EBNA1+VCA-p18] plasma as prognostic marker of disease progression
To determine the OD 450 level that can define progression risk, a calculation based on receiver operating characteristic (ROC) was done and cut-off value for IgA/[EBNA1+VCA-p18] was determined at OD 450 1.44 with 81.3% of sensitivity and 58.9% of specificity.Kaplan-Meir analysis was performed to estimate survival difference based on high risk serology (IgA/[EBNA1+VCA-p18] OD 450 ≥1.4) and low risk serology (OD 450 < 1.4).Twenty seven of 46 NPC patients (58.1%) had disease progression or died during followup.The estimation of survival difference by serology yielded 30 patients at high risk and 16 at low risk.Disease progression or death were more frequent in high-risk serology group (22 of 30 subjects or 73%), with median time to progression of 13.5 month.In low risk serology group (n = 16), 5 patients had progression (31.3%) while median time of progression was not reached.There were significant differences in progression free-survival (PFS) according to serology risk (p = 0.014) as shown in Figure 2A.In high risk serology group, median overal survival (OS) was 17.3 month, whereas in low risk serology group, median OS was not reached (p = 0.114) as shown in Figure 2B.

Serology
IgA/[EBNA1+VCA-p18] changes and disease relapse The changes of serology level in this study failed to indicate a difference between those who had poor outcome (disease progression or relapse) and good outcome (remission and stable disease, without progression).During follow-up, patients with disease progression showed elevated serology, with mean elevation of OD 450 0.73 ± 1.00; whereas those without progression had mean mean elevation OD 450 0.50 ± 0.89 (p = 0.49; 95% CI: 0.45 to 0.90).Eighteen out of 28 patients (64.3%) with stable or elevated IgA within 1 year had disease progression; whereas 2 out of 7 patients (28.6%) with decreased IgA during the same period had progression (X 2 = 3.85; p = 0.14).

DISCUSSION
People infected with EBV will develop specific antibody response including IgM against viral capsid antigen (VCA) during acute primary infection, followed by IgG against VCA and EBV nuclear antigen 1 (EBNA1) that persist for life (Tsuchiya, 2002;Hess, 2004).Aberrant level of antibody response against EBV has been evident in various EBV-related malignancies (Tao et al., 2006).Nasopharyngeal cancer patients often shows increase in antibody response of IgA and IgG against VCA, EA, EBNA1 and transcription activator Zta and Rta, as well as other EBV lytic cycle protein (Henle and Henle, 1976;Fachiroh et al., 2004).Elevation of antibody responses to EBV may precede onset of clinical manifestation of NPC by 1 to 5 year (Yip et al., 1994;Ji MF et al., 2007).
Combined EBV serological biomarkers could improve diagnostic value of NPC (Neel and Taylor, 1990;Fachiroh et al., 2006;Liu et al., 2012;Ai et al., 2013;Chang et al., 2013).Furthermore, dynamic fluctuation of antibody level after treatment of NPC raised the possibilities of humoral response to be used as prognostic marker (Yip et al., 1994).Specific antibody responses to EBV proteins have become a powerful tool to detect reactivation of this virus in human body.Previous studies reported various serological biomarkers as prognostic factors with inconsistent results (Karray et al., 2005;Liu et el., 2004;Neel and Taylor, 1990;Yip et al., 1994;de Vathaire et al., 1988).Fan et al. (2004) reported the use of IgA early antigen (EA) serology to predict post treatment outcome.IgA/EA was still detected in 44% of patients and IgG/EA was detected in 94% NPC patients in remission; whilst EBV DNA became undetectable during remission.
This led to conclusion that the role of EA serology was less important than viral EBV DNA load (Liu et al., 2004;Twu et al., 2007).Similar findings was reported by Adham et al. (2013) among NPC patients in Indonesia.Adham et al. (2013) reported that there was no significant reduction at 2-months post-treatment of IgA EBV either IgA EBNA1 or IgA VCA-p18.
On the contrary, Ling et al. (2009) reported that pretreatment IgA/VCA serology test had prognostic value.Patients with higher IgA/VCA level had shorter survival.In this study, pre-treatment IgA/[EBNA1+VCA-p18] serology level could discriminate risk for progression.To our knowledge, there had not been a report of IgA/[EBNA1+VCA-p18] as a prognostic marker of NPC.This preliminary finding may help us to adjust treatment plan for high risk group for recurrence or progression.Higher dose of chemoradiation may be recommended to NPC patients with higher level of pre-treatment  IgA/[EBNA1+VCA-p18].This study measured combined EBV serological biomarkers for prognostic use, added to its diagnostic capacity (Fachiroh et al., 2006;Ai-di et al., 2009).Only pre-treatment serology IgA/[EBNA1+VCA-p18] showed prognostic role for progression.Elevation of IgA/EBV titer within 1-year post-treatment tend to be followed by disease progression, however, this did not reach the statistical significance.Some patients had disease progression or died before the serology test (missing post-treatment serology).Hence it may contribute to this result.Another possible explanation for incapability of post-treatment IgA/EBNA as prognostic factor was that EBV might be harbored not only in nasopharingeal tumor cells, but also in activated infiltrating T-cell, B lymphocytes and epithelial cells which are able to produce EBVrelated antigen.This may result in elevated IgA/EBV even after remission of disease.
Our unpublished data showed that IgA/[EBNA1+VCA-p18] level was not an independent prognostic factors for survival because of its dependency on patient's clinical performance.The prognostic role of serology IgA/[EBNA1+VCA-p18] to predict progression was strong enough (HR 3.19; 95% CI: 1.09 to 9.37; p = 0.03) without adjusting to clinical performance.Clinical performance of the patients declined continously even after treatment cessation due to difficulty in swallowing.Poor nutrition status affects the overall health status and furthermore may impair immune response.Antibody response to EBV antigen may decline after treatment and raise at the time of progression (Ai-di et al., 2009) but unfortunately this dynamic fluctuation is obscured by the declining of clinical performance, immunity and nutritional status.
Immunocompromised host responded inadequately to antigens so the reactivation of EBV might not be detected by measuring antibody response in this situation (Verschuuren et al., 2003).This may explain that only pre-treatment IgA/[EBNA1+VCA-p18] has a role to determine risk for progression.Post-treatment IgA-EBV serology failed to determine prognostic difference among NPC patients.This is consistent with the previous studies done by Adham et al. (2013) and Ai et al. (2013).Direct quantification of viral EBV DNA, therefore, might lend a hand to support the notion that viral reactivation play important role in disease progression and survival (Hassen et al., 2011;Wang et al., 2013;Yip et al., 2014).EBV DNA load from nasopharyngeal brushings and whole blood showed significant reduction at 2-month after treatment, which was not reflected by EBV-IgA serology (Adham et al., 2013).

Figure 1 .
Figure 1.Dynamic changes of OD 450 plasma IgA/[EBNA1+VCA-p18] based on time of evaluation between NPC groups that progressed (n=27) and non progressed (n=19).Evaluation was performed at baseline (pre-treatment), 3-month or 12-week post-treatment, and at 1-year posttreatment or at the time of disease progression (p<0.05).

Table 1 .
Characteristic of subjects.