A randomized clinical trial comparing prophylactic upper versus whole-neck irradiation in the treatment of patients with node-negative nasopharyngeal carcinoma




This study sought to compare the clinical outcomes of upper versus whole-neck prophylactic irradiation in the treatment of patients with node-negative nasopharyngeal carcinoma (NPC).


Between November 2005 and June 2012, 301 patients with node-negative NPC were randomly assigned to receive primary plus prophylactic upper neck irradiation (UNI, 153 patients) or primary plus whole-neck irradiation (WNI, 148 patients). Patients in both groups received irradiation to the primary tumor and the upper neck nodal regions, and patients in the WNI group also received irradiation to the lower neck. The main endpoint of the study was to compare the lower neck control rate between the 2 groups.


With a median follow-up period of 39 months (range, 6-84 months), no patient in either group had a cervical node relapse. The overall survival at 3 years was 89.5% (95% confidence interval [CI] = 84.1%-95.0%) in the UNI group and 87.4% (95% CI = 81.4%-93.5%) in the WNI group (hazard ratio [HR] = 0.866, 95% CI = 0.41-1.82; P = .70). The 3-year relapse-free survival rate was 89.8% and 89.3% (95% CI = 84.2%-95.3% and 83.7%-94.8%, HR = 0.914, 95% CI = 0.42-2.00; P = .82), and the 3-year metastasis-free survival rate was 91.7% and 90.9% (95% CI = 87.0%-96.5% and 85.7%-96.1%) for the UNI and WNI groups, respectively (HR = 1.007, 95% CI = 0.44-2.32; P = .99).


Prophylactic upper neck irradiation is sufficient for patients with node-negative NPC. Cancer 2013;119:3170–3176. © 2013 American Cancer Society.


Radiotherapy is the mainstay of treatment for patients with nasopharyngeal carcinoma (NPC).[1] Because of the rich submucosal lymphatic network of the nasopharynx, NPC has a propensity for early cervical node metastases.[2] A retrospective study showed that up to 40% had nodal recurrence if the neck irradiation was withheld in patients without clinically palpable cervical nodes. Although most of the nodal relapses could be successfully salvaged by therapeutic irradiation, these patients had significantly more distant metastases than those without nodal recurrence.[3] Therefore, it has long been accepted that the bilateral cervical regions should be electively irradiated for potentially metastatic nodes.[4, 5] Currently, most clinical protocols, such as RTOG-0225 (Radiation Therapy Oncology Group-0225), recommend the routine elective irradiation of node groups II to V and the supraclavicular nodal areas, regardless of the nodal status.[6] However, this recommendation is based on retrospective data where the evaluation of node status was largely based on clinical palpation alone. With the routine use of modern imaging in the staging of NPC, the exact volumes that need to be irradiated to obtain the optimal outcomes became controversial, especially for patients with node-negative disease.[7, 8]

It has long been documented that metastases to the cervical lymph nodes from NPC follow an orderly pattern, from the upper neck inferiorly to the lower neck and then the supraclavicular fossa.[9-11] Skip metastases to the IV, Vb, and supraclavicular fossa regions without corresponding higher level involvement are relatively rare, with an incidence that ranges from 0.5% to 7.9%.[12] In addition, computed tomography (CT) and magnetic resonance imaging (MRI) scans have been commonly used in the evaluation of nodal spread in patients with NPC, and the sensitivity of such imaging for the detection of cervical nodal metastases appears to be higher than that of clinical palpation.[13, 14] More recent studies, including one prospective phase 2 study based on CT or MRI, show that the relapse rate in the neck after upper neck irradiation alone is acceptably low, and recurrence outside the irradiated lower neck is extremely rare.[9, 15]

Therefore, we have conducted this randomized clinical trial to verify our hypothesis that omitting the lower neck during prophylactic node irradiation will not significantly increase the rate of lower neck nodal relapse.


Patient Eligibility

This study was approved by the ethics committee of our hospital, and informed consent was obtained from each patient before their entry into the study. Patients with biopsy-proven NPC underwent a complete medical history, physical examination, MRI (or CT scan if MRI was contraindicated) with contrast from the top of head to 3 cm below the clavicles, in addition to a chest x-ray or chest CT, abdominal sonography, and laboratory investigations including a routine full blood count and biochemistry. The disease was then staged according to the American Joint Committee on Cancer (AJCC)/International Union Against Cancer (UICC) 2002 staging classifications.[16]

Patients who fulfilled all of the following criteria were considered eligible for this study: World Health Organization type II or III NPC, no regional nodal metastases based on MRI or CT scans, no distant metastases on clinical work-up, adequate hematological function with total leukocyte count ≥ 4000/μL and platelets ≥ 100,000/μL, sufficient renal function with creatinine clearance ≥ 60 mL/minute, satisfactory performance status with Karnofsky scores ≥ 80, and age between 18 to 70 years old. The exclusion criteria included 1 or more of the following: keratinizing carcinoma or adenocarcinoma, neck dissection or cervical nodal biopsy, pregnancy or lactation, history of prior malignancies other than adequately treated basal cell carcinoma of the skin or carcinoma in situ of the cervix, chemotherapy or radiotherapy prior to the commencement of the study, or relapsed and/or metastatic NPC.

The diagnosis of cervical node metastasis was made by a radiologist with particular expertise in head and neck cancer imaging, and followed the criteria reported by van den Brekel et al.[17] A suspicious metastatic node was defined in this study as any node that did not fulfill the above diagnostic criteria but had a shortest axial diameter of 8 mm, or grouped nodes in which at least 1 had a minimal dimension of 5 mm. The presence of suspicious nodes did not change the node staging, but the corresponding lymphatic drainage level received an additional boost dose of 10 Gy to account for any diagnostic uncertainty.

Eligible patients were randomly assigned to receive radiotherapy to the primary tumor plus prophylactic irradiation to either the whole neck (whole-neck irradiation arm, WNI group) or to the upper neck (upper neck irradiation arm, UNI group). The upper and lower neck was divided at the lower edge of the cricoid bone.


Patients were treated by either conventional or intensity-modulated radiotherapy (IMRT). Concurrent chemotherapy using weekly cisplatin (40 mg/m2) was also recommended for patients with T3 or T4 disease.[18] For patients in both groups, the primary tumor, the potentially involved surrounding tissues, and the upper neck lymph drainage areas were irradiated. We used the technique reported by Yi et al[19] if the patients were treated by conventional radiotherapy, and the inferior limit of the irradiation field was set at the lower margin of the cricoid bone. The prescribed dose was 70 Gy in 35 daily fractions to the nasopharynx and the surrounding risk tissues, and 50 Gy in 25 daily fractions to the upper neck. When the patients were treated by IMRT, the RTOG-0225 standard was used to contour the target volumes.[6] The total doses were 69 Gy to the primary tumor, 60 Gy to high-risk involved tissues and suspicious node drainage area, and 54 Gy to the low-risk regions and group II, III, and Va nodes in 30 daily fractions by the simultaneous modulated accelerated radiotherapy (SMART) technique.[6] For patients in the WNI group, a lower neck anterior field delivered by a 6-MV photon with a 3-cm central block was used to give a total dose of 50 Gy, regardless of the radiotherapy technique used. The junction point to lower neck was set at the lower margin of the cricoid bone, whether the patients were treated by IMRT or conventional radiotherapy.


Treatment results were assessed by complete physical examination, fiberoptic nasopharyngoscopy, MRI (or CT if MRI was contraindicated) scans of the head and neck, chest x-ray, and ultrasound of the liver 3 months after radiotherapy. The patients were then followed up every 3 months for the first 3 years, every 6 months for the next 2 years, and yearly thereafter. At each follow-up visit, patients would undergo a complete physical examination and nasopharyngoscopy. Repeated imaging studies, including MRI scans of the head and neck, chest x-ray or CT scan, liver sonography or CT scan, and bone scans, were performed whenever a clinical suspicion existed, and routinely every 6 months for the first 3 years and yearly thereafter.

Local relapse was confirmed by pathological study of any mass on nasopharyngoscopy or imaging studies that showed progressive bone erosion or continuous enlargement of the parapharyngeal mass. Regional relapse was diagnosed by fine-needle aspiration of the suspicious nodes. Distant metastases were diagnosed based on symptoms, physical examination, and imaging methods as described above. Adverse events were evaluated according to the Common Terminology Criteria for Adverse Events (CTCAE version 3.0).[20]

Statistical Analysis

The primary endpoint of the study was to compare the lower neck control rate of node-negative NPC treated with primary and upper neck irradiation versus primary and whole-neck irradiation. The second endpoint was to compare the overall survival (OS), metastasis-free survival (MFS), and the relapse-free survival (RFS) in the 2 arms. A noninferiority design was used to test the hypothesis that for patients with node-negative NPC, prophylactic irradiation of the upper neck is not inferior to the irradiation of the whole-neck in terms of lower neck disease control.

To determine the sample size when the study was designed, we used data from the retrospective study by Lee et al,[21] in which 3 of 189 patients developed a lower neck relapse when the elective node irradiation was omitted in stage I NPC (using Ho's classification). We assumed that the node control rate in the lower neck after prophylactic WNI or UNI would be 98%, and that a difference of no more than 5% in the rate of lower neck node control would be clinically relevant and of clinical interest. If there is truly no difference between UNI and WNI, then 290 patients (with a 15% rate of dropouts and early loss to follow-up) would be required to be 80% certain that the upper limit of a 2-sided 95% confidence interval (CI) will exclude a difference in favor of WNI of more than 5%.

Analyses were performed on an intention-to-treat (ITT) basis. OS was measured from the date of random assignment to the date of death from any cause. Time to relapse and distant metastases was measured from the date of randomization to the date of first recurrence or first metastasis site. The OS, RFS, and MFS were calculated with the Kaplan-Meier method, and differences between groups were compared by means of the log-rank test. The Cox proportional hazard model was performed to evaluate the hazard ratios (HRs) between the treatment groups. Toxicity was compared using the chi-square test. Fisher's exact test was used whenever a small sample size existed. SPSS version 17.0 software (SPSS Inc., Chicago, Ill) was used for the statistical analysis. A 2-sided P value < .05 was considered to be statistically significant.


Patient Characteristics and Treatment Compliance

From November 2005 to June 2012, 301 eligible patients were recruited to the study (153 in the UNI group and 148 in the WNI group). Patient characteristics were well balanced with respect to sex, age, staging, suspicious nodes, and radiotherapy modality (Table 1). Compliance with the protocol was excellent. Overall, 147 of the 153 in the UNI group and 145 of 148 patients in the WNI group received the full dose of radiotherapy as per the protocol (Fig. 1).

Figure 1.

CONSORT diagram shows recruitment of the patients in this study. Abbreviations: ITT, intention-to-treat; UNI, upper neck irradiation; WNI, whole-neck irradiation.

Table 1. Clinical Characteristics of Patients From the Intent-to-Treat Population
CharacteristicWNI (n = 148) No. of Patients (%)UNI (n = 153) No. of Patients (%)
  1. Abbreviations: CRT, conventional 2-dimensional radiotherapy; IMRT, intensity-modulated radiotherapy; RT, radiotherapy; UNI, upper neck irradiation; WNI, whole-neck irradiation.

  2. a

    Based on International Union Against Cancer (UICC) TNM staging, 6th edition.

  3. b

    Any node that does not fulfill the diagnostic criteria but has a shortest axial diameter of 8 mm, or grouped nodes in which at least 1 has a minimal dimension of 5 mm.

  4. c

    RT modality was unknown in 1 patient in the WNI group.

Male107 (72)115 (75)
Female41 (28)38 (25)
Age (median, years)4846
T stagea
T1/242 (28)47 (31)
T3/4106 (72)106 (69)
Suspicious nodeb12 (8)15 (10)
RT modalityc
CRT98 (66)97 (63)
IMRT49 (33)56 (37)

Nodal Control

The median follow-up period was 39 months (range, 6-84 months). There was no confirmed case of cervical node relapse including the lower neck relapse in either of the 2 groups. However, 1 patient in the WNI arm had a 5-mm node in the level II region 1 year after treatment. The node progressed to 7 mm after observation for another year. The patient refused a fine-needle aspiration, and remains alive with the node. Another patient in the UNI group had a small node in the left supraclavicular fossa identified on MRI immediately after radiotherapy with a diameter of 4 mm. The patient has been followed for 2 years to date, and the node remains stable in size. Unfortunately, the patient developed lung metastases 14 months after treatment.


During the follow-up period, there were 42 deaths, 28 relapses, and 29 distant metastases in the 2 groups. The Kaplan-Meier curves for OS are shown in Figure 2. The estimated 3-year OS rates were 89.5% (95% CI = 84.1%-95.0%) in the UNI group and 87.4% (95% CI = 81.4%-93.5%) in the WNI group (HR = 0.866; 95% CI = 0.41-1.82; log-rank P = .70). The 3-year RFS rates were 89.8% (95% CI = 84.2%-95.3%) and 89.3% (95% CI = 83.7%-94.8%) in the UNI and WNI arms, respectively (HR = 0.914, 95% CI = 0.42-2.00; log-rank P = .82; Fig. 3). The 3-year cumulative MFS rates were 91.7% (95% CI = 87.0%-96.5%) and 90.9% (95% CI = 85.7%-96.1%) in the UNI and WNI arms, respectively (HR = 1.007, 95% CI = 0.44-2.32, log-rank P = .99; Fig. 4).

Figure 2.

Cumulative overall survival rates are shown between the 2 groups.

Figure 3.

Cumulative relapse-free survival rates are shown between the 2 groups.

Figure 4.

Cumulative metastasis-free survival rates are shown between the 2 groups.

Adverse Events

Adverse events related to the irradiation are listed in Table 2. There were no significant differences in the types and degrees of the acute and late toxicities, except that 68.2% of patients in the WNI group had lower neck grade I acute dermatitis, and 13.5% and 4.7% patients in the WNI group had grade I skin atrophy in the lower neck regions and fibrosis of the lung apex, respectively.

Table 2. Radiation-Related Adverse Events Between Treatment Groups
EventWNI (n = 148), GradeUNI (n = 153), GradeP
  1. Abbreviations: CNS, central nervous system; UNI, upper neck irradiation; WNI, whole-neck irradiation.

Acute complications (%)
Dry mouth8.
Lower neck dermatitis31.868.20100.000.000
Late complications (%)
Skin atrophy62.237.8064.735.30.647
Dry mouth14.183.22.717.
Hearing loss20.275.14.713.878.47.8.205
Cranial neuropathy92.67.4096.13.90.187
CNS necrosis93.25.41.495.44.60.331
CNS hemorrhage99.300.7100.000.492
Secondary malignancy100.00099.300.71.000
Lower neck skin atrophy86.513.50100.000.000
Pulmonary apex fibrosis95.34.70100.000.019


Our results confirmed the hypothesis that the probability of lower neck occult metastases by subclinical nodes was extremely low in patients who had node-negative NPC, and the omission of the lower neck during prophylactic node irradiation did not decrease the rate of lower neck control. At a median follow-up period of 39 months, there was no case of confirmed cervical nodal relapse. This exceptionally low neck relapse rate might be partly due to the 10-Gy boost to the suspicious nodes, which could have improved the treatment of subclinical disease. Our result was similar to those of the several recently reported trials in which node metastases were diagnosed on the basis of imaging studies. Chen et al[15] demonstrated that the regional control of 128 patients with node-negative (N0) NPC was 98.2% with no relapse in the IV and Vb regions, which were omitted during elective node irradiation. In a study by Gao et al,[8] 410 patients with N0 NPC were given prophylactic irradiation to the upper neck region only. Four of these 410 patients developed neck recurrences, but only 1 had a lower neck relapse outside the irradiation field. Another study by Tang et al[9] presented a series of 138 patients who underwent selective irradiation to the upper neck alone (37 patients) or to the entire neck (101 patients), with no neck relapse in either group.

Although the prophylactic irradiation of the whole neck has been recommended in patients with NPC, irrespective of nodal status, this was based on clinical observations and a small number of retrospective studies. The nodal volume needed for prophylactic irradiation in patients with N0 NPC remains controversial.[7] The only randomized clinical trial concerning this question was reported by Ho et al in 1978.[22] In this study, patients with N0 NPC were randomized to receive either upper neck prophylactic irradiation to both sides or no irradiation if the node was not palpable. The results of the patient with stage I disease as described by Ho et al demonstrated that although 5 of the 32 patients in the nonirradiation group developed palpable nodes during the follow-up period, all of them were successfully cured by salvage radiotherapy, and the survival and tumor control rates did not significantly differ between the 2 arms. The authors of that study recommended that irradiation of the neck should be withheld until clinical metastases become evident. However, 2 later retrospective studies from the same center demonstrated that the rate of node recurrence was unacceptably high in patients who did not receive prophylactic node irradiation (30%-40%), and although most of the regional relapses could be successfully rescued, distant metastases were higher in those patients with regional recurrence.[3, 21]

Our randomized trial differed from the former cited research in 2 main respects: first, nodal status was evaluated on the basis of modern imaging techniques (MRI or CT) in our study, whereas the diagnoses of nodal spread in previous studies were mainly based on clinical palpation. CT or MRI scans are more sensitive in detecting nodal metastases in patients with NPC, and some of the node-negative patients in older studies would already have had subclinical metastases, resulting in the overdiagnosis of “node-negative” patients by palpation. These would have eventually manifested as nodal recurrence when the nodal drainage areas were left untreated. The second and probably most important difference was that the entire neck of patients in former studies was not irradiated, whereas only the lower neck in our UNI arm was omitted from the prophylactic irradiation protocol. Because nodal spread occurs in an orderly pattern, the upper neck would be at a relatively high risk of subclinical node metastases, even though no nodal metastases were considered to be present after a complete clinical and radiographic study. Because the aforementioned study showed that more than 95% of recurrent node disease occurred in the upper neck region, the omission of UNI would certainly result in a higher rate of regional relapse.

Although lower neck irradiation is relatively safe and well-tolerated, it is not without its adverse effects. The soft tissues, part of the thyroid gland, carotid, and apex of the lungs were inevitably irradiated when the lower neck region was treated, which could damage these tissues.[23] Although no significant differences existed in the overall toxicities between the 2 groups, more than two-thirds of the patients in the WNI group had faint erythema during radiotherapy in the lower neck regions, and 13.5% and 4.7% of patients had grade I skin atrophy in the lower neck regions and fibrosis of the lung apex, respectively. Whereas the long-term effect of these mild complications on the quality of life of patients remained unclear, based on our relatively short follow-up period, reducing the irradiated volume would surely be beneficial provided that tumor control and survival were not negatively affected.

The OS, RFS, and MFS rates did not significantly differ between groups. The results were comparable to those found in recent studies, despite a relatively higher rate of advanced local disease in our study. Xie et al[24] reported a 3-year OS of 94.2% for patients treated with UNI and 91.9% for those treated with WNI. Results of 2 studies with longer follow-up periods were also reported recently.[15, 25] The 5-year OS rates were 78.7% and 89.8%, 5-year RFS rates were 81.2% and 88.6%, and 5-year MFS rates were 88.5% and 90.6%, for the studies by Chen et al[15] and Sun et al,[25] respectively, when the prophylactic irradiation volume included only the upper neck region. These results suggest that tumor control and survival are not adversely affected when the lower neck is omitted during WNI in the treatment of patients with N0 NPC.

A major concern about our study was the relatively slow enrollment of the patients, because the enrollment period lasted for more than 6 years. Most patients were treated by conventional radiotherapy before 2009 and by IMRT thereafter, which could bring a potential bias into interpreting the results of treatment. However, the comparability of the treatment groups because of the randomized nature of our study, the identical lower margins (which was at the inferior margin of the cricoid bone) whether by conventional radiotherapy or IMRT during the upper neck irradiation, the same imaging techniques used in detecting nodal metastases, and the lack of cervical nodal relapse in the 2 groups make us confident that omitting the lower neck would not increase the regional relapse.

In conclusion, UNI is sufficient as prophylactic node irradiation for patients with N0 NPC.


Supported by Scientific Research Fund of The Health Administration of Jiangxi Province (No. 7111).


The authors made no disclosure.