Nasopharyngeal carcinoma (NPC) is the most common head and neck malignancy in southeastern China and Taiwan. Early detection of the local disease followed by timely and appropriate treatment is essential to increasing cure and survival rates. Detection of Epstein–Barr virus (EBV) genomic DNA, such as the latent membrane protein 1 gene (LMP-1), in patients postirradiation during follow-up may indicate mucosal recurrence.
Seventy-one patients with NPC underwent serial nasopharyngeal swabs for LMP-1 polymerase chain reaction assay before, during, and after irradiation. All of patients achieved a complete disease remission of the LMP-1 gene after irradiation that lasted for at least 6 months.
The median LMP-1 disease remission time after the beginning of irradiation was 4.3 weeks. Patients with early LMP-1 disease remission (≤ 4 weeks after the beginning of irradiation) and delayed LMP-1 disease remission (> 4 weeks) had 3-year local control rates of 93.5% and 76.9%, respectively (P = 0.0529). The LMP-1 gene was detected again (reexpression of LMP-1 [re-LMP-1]) in 10 patients after irradiation with at least 6 months of follow-up. Nine of 10 patients (90%) in the re-LMP-1 positive group and 2 of 61 patients (3.3%) in the re-LMP-1 negative group developed local recurrence. Mucosal recurrence developed in nine patients, and all displayed re-LMP-1. By detecting re-LMP-1 using nasopharyngeal swabs, mucosal recurrence was diagnosed with a sensitivity of 100% (9 of 9 patients) and a specificity of 98.4% (61 of 62 patients). The 3-year overall survival rate, the disease free survival rate for the entire group, and the estimated local mucosal control rates in the re-LMP-1 positive and re-LMP-1 negative groups were 86.5%, 76.5%, 19.4%, and 96.7%, respectively.
Nasopharyngeal carcinoma (NPC) is the most common malignancy among patients with head and neck carcinoma in Southeastern China, Taiwan, Hong Kong, and Singapore. Although survival in patients with NPC ranges from 40% to 70% after radiation, local failure occurs in 10–30% of patients.1, 2 Most recurrences occur within the first 5 years after treatment. Greater than 60% of recurrences occur at the primary sites.3 Patients with recurrent NPC but no repeat treatment display poor survival, although the local salvage rate can range from 30–70% with aggressive treatment.1, 4, 5 Recurrent tumors of the nasopharynx can be treated again with radiotherapy, or radiotherapy can be combined with other treatment modalities.6, 7 Recently, more advanced radiation techniques, such as three-dimensional conformal radiotherapy with or without radiosurgery and intensity-modulated radiotherapy, have been developed that can provide more accurate dosimetric distribution for recurrent nasopharyngeal tumors and can spare previously irradiated normal tissues, particularly if the recurrent lesion is small.8, 9
Epstein–Barr virus (EBV) is a large herpesvirus carried by most human adults with early infection; it generally is latent, and only rarely does it produce infectious progeny.10 B lymphocytes comprise the likely sole EBV reservoir, whereas persistent infection of epithelial cells may indicate malignant tumors like NPC.11 Numerous investigations have demonstrated a strong correlation between EBV and NPC development. Furthermore, in many serologic studies, patients with NPC showed increased antibodies to EBV antigens, generally immunoglobulin A (IgA) to viral capsid antigens (VCA) and early antigen (EA).12, 13 The latent membrane protein 1 (LMP-1) gene is one of the major transformation-related, EBV-encoded genes and is crucial for B-lymphocyte transformation. In addition, LMP-1 also transforms nonlymphoid cells.14, 15 It has been recognized that LMP-1 is expressed almost constantly in NPC, and most NPC tumor cells and preinvasive lesions of the nasopharynx contain this oncogene.16
Because of the close relation between EBV and NPC, in this investigation, we assumed that the virus would disappear after successful induction of disease remission by radiotherapy and that reappearance of the LMP-1 gene may indicate local recurrence. Given the lack of information concerning locally recurrent NPC related to the existence of EBV LMP-1, in this study, we investigated the reappearance of the LMP-1 gene in patients with NPC who had undergone irradiation and in whom LMP-1 remained undetectable for at least 6 months after treatment. Furthermore, we examined whether, because reexpression of LMP-1 (re-LMP-1) in nasopharyngeal swab specimens may indicate the development of recurrent NPC, the detection of re-LMP-1 by polymerase chain reaction (PCR) analysis can determine mucosal recurrence earlier than other conventional modalities, such as nasopharyngoscopy.
MATERIALS AND METHODS
All patients underwent nasopharyngeal swabbing after signing written informed consent. From July 1999 to June 2001, this study gathered 73 patients with newly diagnosed NPC without distant metastasis who had completed a course of local irradiation treatment (≥ 6840 centigrays [cGy]) and were followed for at least 6 months in the Radiation Oncology Department at Chang Gung Memorial Hospital. NPC histopathology was established in all patients. All patients underwent swabbing with PCR amplification to detect the LMP-1 genome before or shortly after the initiation of radiation treatment, at least every 2 weeks during radiation treatment, upon completion of radiation treatment, and at regular 2-month intervals during follow-up. Among the 73 patients, 2 patients with re-LMP-1 and positive histology on nasopharyngeal biopsy after irradiation within 6-month follow-up, indicating persistent disease, were excluded from the current study. Overall, 187 swab samples were gathered from 71 patients with NPC, and 184 of 187 swabs (98%) provided sufficient DNA for successful PCR amplification. Among this group, patients who remained in LMP-1 disease remission at their first 6-month follow-up visit were enrolled in this investigation. Table 1 summarizes the demographic and clinical characteristics of the patients. The initial disease was staged using the 1997 staging system of the American Joint Committee on Cancer.17 The median dose of external radiotherapy was 68.4 Gy (range, 68.4–70.2 Gy). Sixty-seven patients received intracavitary brachytherapy with a high-dose-rate of iridium-192 for an 800-cGy boost dose in 4 fractions.
Table 1. Characteristics of 71 Patients with Nasopharyngeal Carcinoma with the Presence of the Latent Membrane Protein-1 Gene of Epstein–Barr Virus
No. of patients (%)
WHO: World Health Organization; AJCC: American Joint Committee on Cancer.
Every patient who was enrolled in this investigation underwent a complete workup that included nasopharyngoscopy and a nasopharyngeal swab. DNA was extracted from nasopharyngeal swab samples for PCR analysis to detect LMP-1. For patients with re-LMP-1 in two consecutive samples but with normal nasopharyngoscopy results, the Epstein–Barr nuclear antigen gene (EBNA) also was examined to test for cross-contamination. If a patient was positive for re-LMP-1 in three consecutive samples and EBNA detection provided additional verification, then the patient underwent a nasopharyngoscopy-guided biopsy of highly suspicious lesions or a blind biopsy of the Rosenmuller fossa or the nasopharyngeal roof. Two experienced pathologists read the pathology reports.
All patients were encouraged to irrigate their nasopharynx regularly during or after irradiation to prevent crusting. A 15-cm cotton stick was prepared and inserted into a cut, 14-French feeding nasogastric (NG) tube (the length of cotton stick was 3.5 cm longer than that of the cut segment of the NG tube) with the cotton end kept just inside one end of the NG tube. Then, this prepared cotton-stick tube was immersed in phosphate buffered solution and was ready for use. The cotton-stick tube was inserted through the nasal cavity and advanced toward the nasopharyngeal wall. After the tip of the tube touched the posterior nasopharyngeal wall, the tube (but not the stick) was withdrawn 2 cm from the posterior nasopharyngeal wall. Simultaneously, the cotton was pressed on the posterior and lateral pharyngeal walls to sweep over the surface of the nasopharyngeal wall back and forth several times. The swabbing procedure was performed on each of the nostrils to ensure that all the nasopharyngeal space was covered. In addition, a direct fiberscope was used to evaluate the entire nasopharynx and to search for the most probable site of tumor involvement. Then, the stick was withdrawn, and the cotton end was completely inside the lumen of the NG tube. Finally, the whole cotton-stick tube was removed together from the nasopharynx and the nasal cavity. The cotton sticks were immersed immediately in phosphate buffered solution and sent to the laboratory for processing.18
Specimens gathered from nasopharyngeal swabs were immersed and washed in 2 mL phosphate buffered saline to retrieve suspended cells. After centrifugation, the supernatant fluid was discarded, and 400 μL of water were added to the pellet. A 400-μL aliquot of phenol/chloroform solution was added and extracted twice; then, 400 μL of chloroform were added and extracted once. After ethanol precipitation and centrifugation, the supernatant was discarded, and the pellet was prepared for direct use in PCR by air drying and dissolution in 50 μL of distilled water.
PCR Amplification and Gel Electrophoresis
For PCR amplification, oligonucleotide primers (sense, BN1; antisense, BN2; see below) were used to detect the presence of LMP-1 in extracted DNA. PCR amplification was performed in a total volume of 50 μL, comprised of 5 μL extracted DNA; 0.8 μM sense and antisense primers; 200 μM dNTP; 50 mM KCl; 10 mM Tris-HCl, pH 9.0; 1.5 mM MgCl2, and 2 units of Taq DNA polymerase (Promega, Madison, WI). Samples were amplified for 35 cycles using the following procedure: denaturation at 94 °C for 40 seconds, annealing at 50 °C for 1 minute, and extension at 72 °C for 90 seconds in a programmable thermal controller (PTC-100; MJ Research, Watertown, MA) without the overlay of mineral oil. Products then were examined by 1.5% agarose gel electrophoresis in 1 × Tris-boric acid-ethylenediamine tetraacetic acid solution and stained with ethidium bromide to test for the presence of a 316 base-pair PCR product.
Detection of the Presence of LMP-1
For this investigation, we amplified regions of the EBV LMP-1 gene (sense primer BN1, 5′-AGC GAC TCT GCT GGA AAT GAT-3′; antisense primer BN2, 5′-TGA TTA GCT AAG GCA TTC CCA-3′) to identify viral DNA. Amplification of a genomic region of the hemoglobin gene (sense, HEMHC-1,5′-CGT CTC CTT TCC TCC GGA-3′; antisense, HEMHC-2,5′-CAC AT GAC CTT CCC ATC-3′) served as a marker for the presence of intact genomic DNA. Negative control samples containing water were processed in parallel with all patient samples. In addition, DNA from the B95-8 cell line was used as the EBV positive control.
DNA samples from nasopharyngeal swabs were sent for direct sequencing. Purified DNA was amplified using the ABI PRISM™dRhodamine Terminator Cycle Sequencing Ready Reaction Kit (Perkin-Elmer, Oak Brook, IL). The amplicon then was purified with ethanol and sodium salt.
Follow-Up after Treatment
Regular examinations were scheduled 6 weeks after the completion of treatment, every 2 months during the first 2 years, and every 3 months thereafter. However, the follow-up interval was shortened to 2 weeks if re-LMP-1 was detected. Follow-up examinations included a physical examination and assessment of the nasopharynx by direct fiber optic endoscopy. In addition, magnetic resonance imaging (MRI) studies were performed 4 months after treatment, then twice per year, and subsequently at yearly intervals. Patients underwent nasopharyngeal biopsy if a suspicious recurrence was noted by direct fiber optic endoscopy with or without MRI changes. All patients underwent annual chest X-rays, bone scans, and liver echograms.
This investigation used hemoglobin genes as markers of adequate nasopharyngeal epithelial cell DNA. Of a total of 187 swab samples from 71 patients, 184 swabs displayed hemoglobin genes. Hence, this investigation confirmed that 98% of the swab samples obtained successfully from the nasopharynx contained sufficient DNA to amplify the LMP-1 gene.
LMP-1 Status after Primary Irradiation
Information was available for all patients on February 28, 2003. The median follow-up for the whole series, from tissue diagnosis to last known status, was 28.8 months (range, 9.8–42.3 months). LMP-1 became undetectable before the completion of radiation treatment in 63 patients (88.7%) and after treatment in 8 patients (11.3%). The median time from the beginning of irradiation to the disappearance of LMP-1 was 4.3 weeks (range, 1.3–28.0 weeks). However, two patients failed to comply with the follow-up requirements. One patient with positive LMP-1 status at the end of treatment did not visit our outpatient services until 8.1 months after irradiation, at which time LMP-1 was not detectable. The other patient had negative LMP-1 status 2.7 weeks before the completion of radiation and was lost to follow-up after 2 consecutive postradiation outpatient services; however, the patient was referred from the Ear, Nose, and Throat Department because of a local recurrence after 8.6 months.
Thirty-two patients had early LMP-1 disease remission (≤ 4 weeks after the initiation of irradiation). Among this group, 2 patients exhibited re-LMP-1 11.5 months and 16.2 months after irradiation and eventually developed local recurrences. Thirty-nine patients had delayed LMP-1 disease remission (> 4 weeks after beginning irradiation), and 9 patients in this group developed local recurrences. The outcomes of patients with different LMP-1 disease remission times were compared in terms of local control. This comparison revealed that the 3-year local control rate for patients with early and delayed LMP-1 disease remission was 93.5% and 76.9%, respectively (P = 0.0529) (Fig. 1).
Re-LMP-1 Local and Regional Status after Primary Irradiation
Nine of 71 patients (12.7%) developed local recurrences in the nasopharyngeal mucosa, with a median follow-up of 9.8 months (range, 6.3–21.8 months), and 2 patients with initial T4 disease developed local recurrences at the base of the skull at 10.6 months and 12.6 months after primary irradiation. Ten of 71 patients (14.1%) had re-LMP-1 with a median duration of 11.5 months (range, 6.1–19.0 months) after irradiation. Of those 10 patients, only 1 patient remained under local control, and the other 9 patients developed mucosal recurrences. Mucosal recurrence was confirmed by pathology verification in eight patients and by a combination of observing progressive changes in serial nasopharyngoscopies and MRI examinations in one patient. The patient who lacked tissue proof of mucosal recurrence displayed three successive positive results of re-LMP-1 and EBNA. She refused to undergo a biopsy despite nasopharyngoscopy that clearly showed an elevated mass with progression, and the patient died of local disease progression 13.4 months after the detection of re-LMP-1. Table 2 lists the detailed clinical characteristics and EBV LMP-1 status of these 10 patients.
Table 2. Clinical Characteristics and Epstein–Barr Virus Status of 10 Patients with Redetectable Latent Membrane Protein 1 Gene
Time to LMP remission after starting RT (weeks)
Time to LMP reappearance after completing RT (mos)
Time between reappearance of LMP and ABscopy (weeks)
Time to LR (mos)
LMP: latent membrane protein; RT: radiation therapy; ABscopy: abnormal finding on nasopharyngoscopy; LR: local recurrence; F: female; M: male; NED: no evidence of disease.
This patient was lost follow-up until a mass lesion was noted at the time of examination at the Ear, Nose, and Throat Outpatient Department.
It was found that, as expected, all paraffin embedded nasopharyngeal tissues from the patients with mucosal recurrent NPC in this investigation harbored EBV. Furthermore, the partial sequence of EBV re-LMP-1 was identical in specimens from swabbing and biopsy in the same patients with NPC. Thus, the presence of EBV genome in the nasopharyngeal swab indicated the presence of recurrent NPC tumor cells. The nasopharyngeal swab specimens detected mucosal recurrent NPC with a positive predictive value, sensitivity, and specificity of 90%, 100% (9 of 9 patients), and 98.4% (61 of 62 patients), respectively. Among the 9 patients who developed mucosal recurrences, 1 patient also had a simultaneous cervical lymph node recurrence, and 2 patients had lumbar spine metastases 6.6 months and 10.5 months after primary irradiation.
Nasopharyngoscopic Abnormalities and Detection of re-LMP-1
Among the nine patients with mucosal recurrences, no nasopharyngoscopic abnormalities were noted in eight patients at the time of re-LMP-1 detection. It was found that one patient, as mentioned earlier, had re-LMP-1 after pathology verification because that patient was lost to follow-up in our department.
The median duration between re-LMP-1 detection and aberrant findings by fiber optic endoscopy was 11.9 weeks (range, 2.7–27.4 weeks) in patients who developed mucosal recurrences. In two patients who had two successive, positive results for re-LMP-1 but had relatively normal nasopharyngoscopic findings, EBNA was tested to rule out the possibility of cross-contamination. It was suspected strongly that both patients had NPC, because both EBNA and re-LMP-1 were positive. Thus, nasopharyngeal biopsies were performed at the suspicious area of the nasopharynx and revealed only mild, erythematous changes without easy bleeding on touch that were strongly suspicious for benign granulation tissues caused by radiation on nasopharyngoscopic examination, displaying malignant recurrence after pathologic verification.
Retreatment for Recurrence
Of 9 patients with mucosal recurrences in the group with re-LMP-1, 7 patients received nasopharyngectomy, one received external irradiation alone at a dose of 18 Gy but died during retreatment, and 1 patient refused any retreatment and died 10 months after local recurrence. Of 7 patients who underwent with nasopharyngectomy, 2 patients with positive resection margins received postoperative radiotherapy ay a dose of 66 Gy. Table 3 lists the details of retreatment and the disease status of 9 patients who experienced mucosal recurrences in the re-LMP-1 group.
Table 3. Treatment for Local Recurrence and Status of Nine Patients with Mucosal Recurrence
Two patients with skull-base recurrences in the group without re-LMP-1 had received reirradiation at doses of 66.6 Gy and 48 Gy. One of those patients died after a follow-up of 27.4 months.
Overall Survival, Disease-Free Survival. and Local Control
Nine of 71 patients died (3 of 10 patients in the re-LMP-1 group and 6 of 61 patients in the group without re-LMP-1). Seven deaths were disease-related (three in the re-LMP-1 group and four in the no re-LMP-1 group). The 3-year overall and disease-free survival rates for the entire group were 86.5% and 76.5%, respectively. Moreover, the median survival was 28.8 months for the entire group (27 months in the re-LMP-1 group and 29.9 months in the group without re-LMP-1). At 3 years, the estimated local control rates were 19.4% and 96.7% in the re-LMP-1 group and in the group with out re-LMP-1, respectively (P < 0.0001) (Fig. 2).
The ideal method for finding EBV should be simple and reliable. Recently, highly sensitive PCR analysis has achieved the most accurate detection of EBV in human tissue and, thus, has been adopted in most analysis labs.19 Previously, we designed an effortless and noninvasive technique for nasopharyngeal swabbing and confirmed that swab samples obtained from the nasopharynx of untreated patients could provide sufficient EBV DNA to amplify the LMP-1 gene.18 That investigation showed that EBV LMP-1 detected NPC with a sensitivity of 94.7% and a specificity of 100% in the nasopharyngeal swab specimens. Sheen et al. also demonstrated that nasopharyngeal swabbing and PCR could supplement serologic screening effectively in newly diagnosed NPC.20 In addition, Sheen et al. reported a 77.8% positive LMP-1 rate in the recurrent group and a 0.0% positive LMP-1 rate in the control group; however, the number of patients in their study was limited. The current investigation is the largest study to date emphasizing the relation between local recurrence in patients who were treated with definitive radiotherapy and re-LMP-1. Several serologic studies previously demonstrated that elevated levels of EBV IgA antibodies to VCA and EA can be used to screen for NPC, but the sensitivity and specificity varied among cut-off points.21 For patients with high titers of anti-EBV VCA or EA antibodies before irradiation, the serologic markers were raised persistently, even when the tumor was in regression. Thus, detecting recurrent NPC based on a serologic method is not feasible.
In this investigation, patients achieved LMP-1 disease remission after irradiation within a median of 4.3 weeks. Patients with early LMP-1 disease remission during irradiation had marginally better outcomes in terms of local control (P = 0.0529). This finding suggests that the responsiveness of LMP-1 to irradiation may represent the radiosensitivity and malignancy of tumor cells. In addition, changes in LMP-1 status at the time of recurrence may occur earlier than any morphologic changes in the nasopharyngoscopy. In the current study, among patients with mucosal recurrences, re-LMP-1 was detected 11.9 weeks earlier than the abnormal nasopharyngoscopic findings. Furthermore, nasopharyngeal biopsies revealed mucosal recurrence, but no bizarre endoscopic findings, in two patients. Differentiating between recurrent tumor cells and radiation-induced necrotic or scarring changes at the nasopharynx using nasopharyngoscopy or conventional anatomic imaging is difficult until the macroscopic mass becomes obvious with ulceration or asymmetric mucosal change. Both MRI and CT scans have low sensitivity and moderate specificity for detecting recurrent NPC and for distinguishing recurrence from postradiation therapy changes.22 However, MRI is superior to CT scans in differentiating postradiation fibrosis from recurrent NPC, although early, distinct differentiation generally is impossible due to immature scars, which, along with well hydrated granulation tissues, generally reveal contrast enhancement.23 Recently, 18-fluoro-2-deoxyglucose-positron emission tomography (PET) or thallium-201, single-photon emission CT has been recommended for detecting recurrent NPC, particularly when MRI findings are indeterminate.24 Even if functional imaging can provide good sensitivity and specificity, numerous economic and practical problems limit its usefulness in detecting recurrence in the nasopharynx, particularly when an early diagnosis is desired.
The detection of re-LMP-1 in radiation-treated patients with NPC can enhance physician awareness, causing physicians to shorten patient follow-up intervals, pay increased attention to changes in the nasopharynx, and even perform biopsy more frequently in suspicious nasopharyngeal mucosal areas. Thus, it is reasonable that mucosal recurrence may be diagnosed earlier using detection of re-LMP-1 as one of the follow-up modalities. Moreover, salvage treatment can be more successful if NCP can be treated early. Although no randomized comparison exists, mucosal recurrence seemed to be detected earlier in this study compared with previous investigations. Lee et al. reported that 891 of 4779 patients with NPC who had achieved complete disease remission after irradiation for at least 6 months developed true local recurrences, with a median latent interval of 20.4 months.1
Tune et al. reported a sample of 32 radiation-treated patients. Of nine patients with recurrent or residual disease, only two patients were positive for EBV; therefore, scarring change after irradiation may reduce the overall ability of nasopharyngeal brushing to detect EBV positivity in patients with recurrent NPC.25 However, the fact that 98% of the swab samples obtained successfully from the nasopharynx contained sufficient DNA for the amplification of the LMP-1 gene did not appear to be a problem in this study. One reason for this phenomenon may be that radiation-treated patients were encouraged to irrigate their nasopharynx vigorously, even during radiation treatment, to prevent crust formation.
Few works have examined the relation between LMP-1 status and patients with irradiation-treated NPC; however, some investigations have examined the association between irradiation-treated patients with mucosal control of disease and the incidence of LMP-1. In the current study, the incidence of LMP-1 detection in irradiation-treated patients without mucosal recurrence was only 1.67% (1 of 60 patients). This incidence corresponds to the prevalence among the general population reported by Hao et al., with a true positive rate of 1.17% (3 of 256 patients) by testing both LMP-1 and EBNA.26 Nevertheless, some reports have revealed that the LMP-1 genome could be identified at a rate of 25–80% from throat washings, nasopharyngeal biopsies, or swabs in individuals without NPC or apparent EBV-related disease.20 Along with the possibility that cross-contamination during tissue sampling or PCR may be able to explain the divergence in the detection of the EBV genome in the nasopharynx, more effort should be made to clarify whether the divergence in the probability of detecting EBV specific genomic DNA is a consequence of cross-contamination or a more sensitive and specific assay, such as nest PCR.27 In any event, the current study tested EBNA in patients who had two successive positive results for re-LMP-1 to minimize the possibility of cross-contamination.
The detection of EBV genomic re-LMP-1 using a nasopharyngeal swab technique has a sensitivity of 81.8% and a specificity of 98.3% for predicting mucosal recurrence of NPC. A mucosal recurrence should be suspected strongly if re-LMP-1 is present in patients with treated NPC who had a latent disease remission of LMP-1 that exceeded 6 months, even if nasopharyngoscopy revealed no abnormality. In such situations, punch biopsy or imaging investigations, such as MRI or PET scans, can be performed. Undoubtedly, detection of EBV re-LMP-1 by nasopharyngeal swabbing combined with PCR assay should be part of a follow-up modality for patients with irradiation-treated NPC.