Original Research
Neuroradiology/Head and Neck Imaging
September 27, 2023

Optimal Size Threshold for MRI-Detected Retropharyngeal Lymph Nodes to Predict Outcomes in Nasopharyngeal Carcinoma: A Two-Center Study

Please see the Editorial Comment by Paul M. Bunch discussing this article.
Chinese (audio/PDF) and Spanish (audio/PDF) translations are available for this article's abstract.
To listen to the podcast associated with this article, please select one of the following:
iTunes, Spotify, Google, or direct download.

Abstract

BACKGROUND. Retropharyngeal lymph node (RLN) metastases have profound prognostic implications in patients with nasopharyngeal carcinoma (NPC). However, the AJCC staging system does not specify a size threshold for determining RLN involvement, resulting in inconsistent thresholds in practice.
OBJECTIVE. The purpose of this article was to determine the optimal size threshold for determining the presence of metastatic RLNs on MRI in patients with NPC, in terms of outcome predictions.
METHODS. This retrospective study included 1752 patients (median age, 46 years; 1297 men, 455 women) with NPC treated by intensity-modulated radiotherapy (RT) from January 2010 to March 2014 from two hospitals; 438 patients underwent MRI 3–4 months after treatment. Two radiologists measured the minimal axial diameter (MAD) of the largest RLN for each patient using a consensus process. A third radiologist measured MAD in 260 randomly selected patients to assess interobserver agreement. Initial ROC and restricted cubic spline (RCS) analyses were used to derive an optimal MAD threshold for predicting progression-free survival (PFS). The threshold's predictive utility was assessed in multivariable Cox regression analyses, controlling for standard clinical predictors. The threshold's utility for predicting PFS and overall survival (OS) was compared with a 5-mm threshold using Kaplan-Meier curves and log-rank tests.
RESULTS. The intraclass correlation coefficient for MAD was 0.943. ROC and RCS analyses yielded an optimal threshold of 6 mm. In multivariable analyses, MAD of 6 mm and greater independently predicted PFS in all patients (HR = 1.35, p = .02), patients with N0 or N1 disease (HR = 1.80, p = .008), and patients who underwent posttreatment MRI (HR = 1.68, p = .04). In patients with N1 disease without cervical lymph node involvement, 5-year PFS was worse for MAD greater than or equal to 6 mm than for MAD that was greater than or equal to 5 mm but less than 6 mm (77.2% vs 89.7%, p = .03). OS was significantly different in patients with stage I and stage II disease defined using a 6-mm threshold (p = .04), but not using a 5-mm threshold (p = .09). The 5-year PFS rate was associated with a post-RT MAD of 6 mm and greater (HR = 1.68, p = .04) but not a post-RT MAD greater than or equal to 5 mm (HR = 1.09, p = .71).
CONCLUSION. The findings support a threshold MAD of 6 mm for determining RLN involvement in patients with NPC.
CLINICAL IMPACT. Future AJCC staging updates should consider incorporation of the 6-mm threshold for N-category and tumor-stage determinations.

Highlights

Key Finding
OS for patients with stage I and II NPC was significantly different using a 6-mm (p = .04) but not a 5-mm (p = .09) threshold. Five-year PFS was associated with post-RT MAD 6 mm and greater (HR = 1.68, p = .04) but not 5 mm and greater (HR = 1.09, p = .71).
Importance
Use of a 6-mm threshold, rather than a 5-mm threshold, can facilitate better risk stratification and treatment decision-making in patients with NPC.
Nasopharyngeal carcinoma (NPC) is a significant global health concern, with 133,354 new diagnoses and 80,008 deaths worldwide in 2020 [1]. China is disproportionately impacted by NPC, having higher morbidity and mortality rates than the global average and accounting for approximately 45% of global NPC-related deaths [2].
Retropharyngeal lymph nodes (RLNs) represent the first-echelon lymph nodes in NPC [3], playing a crucial role in disease spread. RLN involvement is common [4, 5] and an independent predictor of survival [6, 7], thus having profound prognostic and therapeutic implications [8, 9]. The current AJCC cancer staging system (8th edition) [10] categorizes metastatic RLNs measuring up to 6 cm as representing N1 disease, but does not provide a minimum size threshold for determining whether RLNs are meta-static. According to the current NCCN guidelines, the presence of metastatic RLNs impacts treatment selection in patients with NPC and T1–T3 disease, warranting definitive radiotherapy (RT) with possible concurrent systemic therapy and induction chemotherapy [11]. Therefore, establishing optimal criteria for defining RLN metastases is critical. Pathologic evaluation would represent the ideal method to determine the presence versus absence of nodal metastasis. However, the RLNs are rarely surgically resected during neck dissection, such that a pathologic reference standard is generally unavailable, and clinical and imaging follow-up are instead used to provide a surrogate reference standard.
MRI can clearly define the distribution and extent of RLNs [4]. Previous studies focused predominantly on size criteria for defining RLN metastases based on the lymph node's minimal axial diameter (MAD), although consensus for an optimal MAD is lacking [1218]. For example, some studies suggested an optimal MAD threshold of 5 mm for diagnosing metastatic RLNs [1214]. Other studies proposed an optimal MAD threshold of 6 mm [1518]. However, these were all single-center studies, and the largest study, by Li et al. [18], had a sample size of 817 patients. In addition, these earlier studies did not comprehensively compare the use of 5-mm and 6-mm thresholds for differentiating N0 from N1 disease or assess the impact of the different thresholds on disease staging using the AJCC 8th edition [1218]. Additionally, earlier studies' inclusion of patients with cervical lymph node (CLN) metastases among patients with N1 disease may have introduced statistical bias. Moreover, a paucity of literature has explored the role of RLN size on post-RT MRI in predicting outcomes.
This study aimed to determine the optimal size threshold for determining the presence of metastatic RLNs on MRI in patients with NPC, in terms of prediction of clinical outcomes.

Methods

Patients

This retrospective study was conducted at two centers, Sun Yat-sen University Cancer Center (hereafter, Hospital A) and the First People's Hospital of Foshan (hereafter, Hospital B). The ethics committees of both hospitals approved the study and waived the requirement for informed consent.
All patients in the present sample were also included in at least one prior study [1922] that evaluated factors predicting outcomes in patients with NPC, but none of which compared different minimum size thresholds for defining metastatic RLNs. One of those studies [20] used the same sample of 1752 patients as in the current study. As part of the completion of those earlier studies, the study investigators assembled a research database comprising comprehensive information for patients with NPC, including clinical information, MRI findings, and follow-up events. All data for the current study were extracted from that database.
To initially establish the research database, the EMR was searched from January 2010 to January 2013 at Hospital A and from April 2010 to March 2014 at Hospital B for patients with a new pathologically confirmed diagnosis of NPC who completed treatment by intensity-modulated radiotherapy (IMRT), yielding 4561 and 894 patients at Hospitals A and B, respectively. Patients were then excluded for the following reasons: incomplete medical records (3142 and 410 patients from Hospitals A and B, respectively), distant metastatic disease at the time of NPC diagnosis (49 and 25 patients, respectively), synchronous primary cancer at the time of NPC diagnosis (40 and seven patients, respectively), and did not undergo complete staging MRI at the time of diagnosis (10 and 20 patients, respectively). The final study sample thus included 1320 patients from Hospital A and 432 patients from Hospital B, for a total of 1752 patients (median age, 46 years; 1297 men, 455 women). Figure 1 shows the flow of patient selection.
Fig. 1 —Flow diagram shows patient selection at Sun Yat-sen University Cancer Center (Hospital A) and First People's Hospital of Foshan (Hospital B). NPC = nasopharyngeal carcinoma, IMRT = intensity-modulated radiotherapy.

Treatments

In all patients, treatments were compliant with the hospital's standardized treatment protocols for NPC and followed the NCCN guidelines in place from 2010 to 2014. All patients received IMRT, per the selection criteria. Patients with stage II–IV disease were recommended to receive concurrent chemotherapy, with further optional induction chemotherapy based on the treating physician's protocol. Induction chemotherapy was always delivered before IMRT. Patients with persistent disease or relapse after initial IMRT received repeat radiation, surgery, or chemotherapy. The Supplemental Methods provides additional information regarding IMRT and chemotherapy regimens.

Follow-Up

At both hospitals, patients were followed by a clinic visit with nasopharyngeal evaluation every 3 months for the first 2 years after IMRT, and then biannually. Patients who were unable to attend follow-up visits in person were instead interviewed by telephone. Follow-up records were reviewed to identify documentation of disease progression, locoregional recurrence, distant metastasis, or death. The study's endpoints included progression-free survival (PFS), calculated from the date of treatment initiation to the date of progression, recurrence, metastasis, or death from any cause, and overall survival (OS), calculated from the date of treatment initiation until the date of death from any cause. Both PFS and OS were determined at the date of last follow-up. A subset of patients from Hospital A underwent additional testing 3–4 months after completion of IMRT, hereafter described as the post-RT time point. These various follow-up endpoints were captured in the research database.

Clinical Variables

The EMR was reviewed for each patient to record age, sex, histologic type of NPC (classified as WHO type 1 or 2 vs WHO type 3), and treatment approach (classified as IMRT, IMRT with concurrent chemotherapy, or IMRT with concurrent chemotherapy and induction chemotherapy). The EMR was additionally reviewed to record plasma Epstein-Barr virus (EBV) DNA levels, determined using real-time fluorescence quantitative PCR, before treatment and, when available, at the post-RT time point. Pretreatment plasma EBV DNA levels were categorized as less than 1000, 1000 to less than 10,000, or 10,000 and greater copies/mL, on the basis of thresholds from a previous study [23]; post-RT plasma EBV DNA levels were classified as 0, 1 to 1000, or more than 1000 copies/mL, determined empirically in the absence of previously established thresholds. These variables extracted from the EMR were captured in the research database. In addition, at the time of research database construction, each patient's T category, N category, and tumor stage at diagnosis were classified using the AJCC 8th edition, on the basis of the available information within the EMR. Because the AJCC 8th edition does not specify a size threshold for determining the presence of metastatic RLNs, this determination was performed using the common threshold of 5 mm.

MRI Acquisition and Interpretation

MRI examinations of the neck were performed before treatment and in a subset of patients at the post-RT evaluation. Examinations were performed using 1.5-T or 3-T systems equipped with 20-channel combined head-neck coils and covered from the suprasellar cistern to the clavicle's inferior margin. The examinations included axial, sagittal, and coronal unenhanced T1-weighted images using a TR/TE of 540/11.8 and axial T2-weighted images using a TR/TE of 4000/99. After IV administration of contrast agent (gadopentetate dimeglumine, dose: 0.1 mmol/kg body weight), axial and sagittal contrast-enhanced T1-weighted images and coronal contrast-enhanced fat-suppressed T1-weighted images were acquired. The axial, sagittal, and coronal contrast-enhanced acquisitions used a section thickness of 5, 3, and 2 mm, respectively, and an interslice gap of 1, 0.5, and 0.5 mm, respectively. Table S1 summarizes MRI acquisition parameters.
As part of the development of the research database, MRI examinations were evaluated by two radiologists (L.L. and a non-author, with 18 and 10 years of posttraining experience, respectively), using a consensus process, blinded to clinical outcomes and to whether an examination was a pretreatment or post-RT examination. For each patient, the radiologists identified the largest visible RLN on axial T2-weighted image and recorded its MAD to the nearest millimeter (Fig. 2). The radiologists also assessed the largest visible RLN in terms of the presence versus absence of central nodal necrosis (CNN) and extracapsular nodal spread. CNN was defined as the presence of focal areas of high and low signal intensity on T2- and T1-weighted images, respectively, with or without a surrounding rim of enhancement [24]. Extracapsular nodal spread was defined as the presence of indistinct nodal margins on any sequence [25]. A prior study evaluating interobserver agreement for characteristics of RLNs on MRI reported a kappa coefficient of 0.877 for CNN and of 0.824 for extracapsular nodal spread [19]. The radiologists also characterized each examination in terms of the presence versus absence of metastatic CLN (hereafter, CLN involvement), using previously published criteria [25]. The results of these MRI interpretations were captured in the research database.
Fig. 2A —49-year-old woman with nasopharyngeal carcinoma.
A, Axial T2-weighted image from MRI examination performed before treatment shows largest retropharyngeal lymph node (RLN) is on right side and has minimal axial diameter (MAD) (red line) measuring 5 mm. Patient was treated by intensity-modulated radiotherapy (IMRT).
Fig. 2B —49-year-old woman with nasopharyngeal carcinoma.
B, Axial T2-weighted image obtained 3–4 months after completion of IMRT shows decrease in size of right RLN, now with MAD (red line) measuring 4 mm. Patient would be considered to have RLN involvement at baseline based on 5-mm threshold commonly used in practice but not by 6-mm threshold supported by current study. Patient would not be considered to have RLN involvement using either threshold on examination after IMRT. No disease progression, locoregional recurrence, distant metastasis, or death occurred during follow-up period of 60.7 months.
For purposes of the present investigation, an additional radiologist (H.L., with 10 years of posttraining experience) independently reviewed 260 randomly selected pretreatment MRI examinations to measure the MAD of the largest visible RLN. This reader's measurements were used solely for assessing interob-server agreement of RLN MAD with respect to the measurements within the database.

Statistical Analyses

Patient characteristics were summarized descriptively and compared between patients from Hospital A and Hospital B using Fisher exact or chi-square tests for categoric variables or t test for continuous variables. Interobserver agreement for RLN MAD measurements was assessed using the intraclass correlation coefficient (ICC). Prognostic analyses were performed in three patient groups: all patients, patients with N0 or N1 disease based on the AJCC 8th edition, and patients who underwent evaluation after RT (including both MRI and plasma EBV DNA measurement). Analyses in patients who underwent evaluation after RT used assessments of RLN MAD, RLN CNN, RLN extracapsular nodal spread, and CLN involvement from the post-RT MRI, as well as both the pretreatment and post-RT plasma EBV DNA measurements, and otherwise used pretreatment data.
Histograms were constructed of the distribution of RLN MAD measurements, and the RLN MAD measurements corresponding the histogram peaks were identified. ROC curves were constructed for prediction of 5-year PFS based on RLN MAD, and the AUC was determined. The optimal threshold was determined using the Youden index, for which the corresponding sensitivity and specificity were derived. Restricted cubic spline (RCS) curves were also constructed, which were used to identify an optimal threshold on the basis of the point at which the lower confidence bound of the curve crossed the reference line (i.e., log relative hazard of zero). Based on these various tests, a single optimal threshold, rounded to the nearest millimeter, was selected for further testing.
Univariable analyses for predicting PFS were performed in all patients and in patients who underwent evaluation after RT using the log-rank tests. Multivariable analyses for predicting PFS were performed in all patients, patients with N0 or N1 disease, and patients who underwent evaluation after RT using Cox proportional hazards models, incorporating variables that were statistically significant in the univariable analyses. If multiple variables related to RLN MAD were significant at univariable analysis, then the multivariable analysis used the variable relating to the derived optimal threshold to mitigate multicollinearity. The multivariable analysis in patients with N0 or N1 disease used the features identified in all patients in the univariable analyses. However, given the potential confounding effect of CLN involvement on the prognostic impact of RLN involvement, analyses in patients with N0 or N1 disease stratified patients into four groups for purposes of assessing RLN MAD: those with N0 disease; those with N1 disease, RLN MAD less than the derived optimal threshold, and no CLN involvement; those with N1 disease, RLN MAD equal to or greater than the derived optimal threshold, and no CLN involvement; and those with N1 disease and CLN involvement. Otherwise, the multivariable models in all patients and in patients with N0 or N1 disease did not include N category or tumor stage, which had been defined using an RLN MAD threshold of 5 mm, even if these variables were statistically significant in univariable analyses, given the investigation's focus on the optimal RLN MAD threshold.
Kaplan-Meier curves were constructed for predicting PFS in patients with N0 or N1disease, stratifying by the previously noted four subgroups. All possible pairwise comparisons for predicting 5-year PFS among the subgroups were assessed using log-rank tests. These analyses were performed using both the derived optimal threshold and a threshold 1 mm greater than the derived optimal threshold.
The impact of the derived optimal threshold for determining N category and tumor stage was assessed in terms of shifts in these classifications in the absence of RLN CNN, RLN extracapsular nodal spread, and CLN involvement. Kaplan-Meier curves and log-rank tests were used to compare PFS and OS among patients stratified by N category and tumor stage on the basis of the original assignments recorded in the research database and from the new assignments incorporating the derived optimal threshold. The impact of the derived optimal threshold for determining post-RT RLN involvement was assessed in a similar manner, incorporating Kaplan-Meier PFS curves and determination of 5-year PFS rates based on a 5-mm threshold and the derived optimal threshold.
Two-sided p values less than .05 were considered statistically significant. All statistical analyses were conducted in R (version 4.0.1).

Results

Patients

The clinical characteristics of the 1752 patients are summarized in Table 1. Table 1 also provides comparisons of clinical characteristics between patients from Hospitals A and B. A total of 1345 of 1752 (76.8%) patients had N0 or N1 disease, based on AJCC 8th edition criteria. A total of 438 of 1320 (33.2%) patients, all from Hospital A, underwent evaluation after RT. During a median follow-up of 61.5 months, 387 (22.1%) patients developed progressive disease, 159 (9.1%) patients developed recurrent disease after treatment, 225 (12.8%) patients developed distant metastatic disease, and 244 (13.9%) patients died.
TABLE 1: Study Variables in All Patients and in Patients Who Underwent Post-RT Evaluation
VariableAll PatientsHospital A Patients Who Underwent Post-RT Evaluation (n = 438)
Total (n = 1752)Hospital A (n = 1320)Hospital B (n = 432)pa
Age (y), median (IQR)46 (39–55)45 (38–54)47 (40–58)< .00146 (38–54)
Sex   .31 
Male1297 (74.0)969 (73.4)328 (75.9) 317 (72.4)
Female455 (26.0)351 (26.6)104 (24.1) 121 (27.6)
Histologic typeb   < .001 
WHO type 1 or 263 (3.6)63 (4.8)0 (0) 16 (3.7)
WHO type 31689 (96.4)1257 (95.2)432 (100) 422 (96.3)
Pretreatment plasma EBV DNA (copies/mL)   < .001 
< 1000852 (48.6)583 (44.2)269 (62.3) 86 (19.6)
1000 to < 10,000493 (28.1)339 (25.7)154 (35.6) 165 (37.7)
≥ 10,000407 (23.2)398 (30.2)9 (2.1) 187 (42.7)
T categoryc   .10 
T1451 (25.7)331 (25.1)120 (27.8) 73 (16.7)
T2213 (12.2)153 (11.6)60 (13.9) 54 (12.3)
T3678 (38.7)532 (40.3)146 (33.8) 187 (42.7)
T4410 (23.4)304 (23.0)106 (24.5) 124 (28.3)
N categoryc   .10 
N0352 (20.1)278 (21.1)74 (17.1) 51 (11.6)
N1993 (56.7)752 (57.0)241 (55.8) 263 (60.0)
N2289 (16.5)205 (15.5)84 (19.4) 89 (20.3)
N3118 (6.7)85 (6.4)33 (7.6) 35 (8.0)
Tumor stagec   .41 
I151 (8.6)115 (8.7)36 (8.3) 7 (1.6)
II388 (22.1)288 (21.8)100 (23.1) 89 (20.3)
III708 (40.4)547 (41.4)161 (37.3) 194 (44.3)
IVa505 (28.8)370 (28.0)135 (31.2) 148 (33.8)
Treatment   .004 
IMRT225 (12.8)156 (11.8)69 (16.0) 21 (4.8)
CCRT635 (36.2)504 (38.2)131 (30.3) 153 (34.9)
CCRT and IC892 (50.9)660 (50.0)232 (53.7) 264 (60.3)
RLNd     
MAD (mm), median (IQR)7 (2–12)7 (0–11)7 (5–12)< .0010 (0–5)
5-mm threshold   .59 
MAD < 5 mm564 (31.2)420 (31.8)144 (33.3) 319 (72.8)
MAD ≥ 5 mm1188 (67.8)900 (68.2)288 (66.7) 119 (27.2)
6-mm threshold   .49 
MAD < 6 mm694 (39.6)529 (40.1)165 (38.2) 382 (87.2)
MAD ≥ 6 mm1058 (60.4)791 (59.9)267 (61.8) 56 (12.8)
CNN   .20 
Absent1348 (76.9)1006 (76.2)342 (79.2) 346 (79.0)
Present404 (23.1)314 (23.8)90 (20.8) 92 (21.0)
ENS   .49 
Absent1385 (79.1)1038 (78.6)347 (80.3) 354 (80.8)
Present 367 (20.9)282 (21.4)85 (19.7) 84 (19.2)
CLN involvementd   .008 
Absent687 (39.2)541 (41.0)146 (33.8) 380 (86.8)
Present1065 (60.8)779 (59.0)286 (66.2) 58 (13.2)
Post-RT plasma EBV DNA (copies/mL)    
0 377 (86.1)
1 to < 1000 28 (6.4)
≥ 1000 33 (7.5)

Note—All findings assessed pretreatment unless otherwise indicated. Dashes indicate not applicable. RT = radiotherapy, Hospital A = Sun Yat-sen University Cancer Center, Hospital B = First People's Hospital of Foshan, EBV = Epstein-Barr virus, IMRT = intensity-modulated radiotherapy, CCRT = concurrent chemoradiotherapy (i.e., IMRT plus concurrent chemotherapy), IC = induction chemotherapy, RLN = retropharyngeal lymph node, MAD = minimal axial diameter, CNN = central nodal necrosis, ENS = extracapsular nodal spread, CLN = cervical lymph node.

a
Comparison of Hospitals A and B using Fisher exact test or chi-square test for categoric variables or t test for continuous variables.
b
According to the 5th edition of the WHO classification of tumors.
c
According to the 8th edition of the AJCC staging system, with N category incorporating the commonly used MAD threshold of 5 mm to determine RLN involvement.
d
Determined before treatment or post-RT MRI, depending on patient group.

Derivation of Optimal Retropharyngeal Lymph Node Minimum Axial Diameter Threshold

The ICC for RLN MAD was 0.943 (95% CI, 0.911–0.961). Figure 3A shows the histogram of RLN MAD measurements in the entire sample. The histogram peak was at an RLN MAD of 6.3 mm. Figure 3B shows an ROC curve for prediction of 5-year PFS by RLN MAD. The AUC was 0.605. Based on the Youden index, the optimal threshold was 6.1 mm, which had a sensitivity of 68.7% (1203/1752) and specificity of 49.2% (862/1752) for 5-year PFS. Figure 3C shows the histogram of RLN MAD measurements in patients with N0 or N1 disease. The histogram peak was at an RLN MAD of 6.1 mm. Figure 3D shows the RCS curve for the association between RLN MAD and PFS in patients with N0 or N1 disease. The log relative hazard of the lower confidence bound equaled 0 at a MAD of 6.2 mm. Figure 3E shows the ROC curve for prediction of 5-year PFS by RLN MAD in patients who underwent evaluation after RT. The AUC was 0.560. Based on the Youden index, the optimal threshold was 5.5 mm, which had a sensitivity of 22.8% (100/438) and specificity of 90.4% (396/438) for 5-year PFS. Figure 3F shows the RCS curve for the association between RLN MAD and PFS in patients who underwent evaluation after RT. The log relative hazard of the lower confidence bound equaled 0 at a MAD of 6.2 mm. Based on these tests, an optimal threshold (to the nearest millimeter) of 6 mm was selected.
Fig. 3A —Derivation of optimal retropharyngeal lymph node (RLN) minimal axial diameter (MAD) threshold.
A, Graphs show RLN MAD measurements before treatment in all patients. Histogram (A) shows distribution of RLN MAD measurements. Peak amplitude is at MAD of 6.3 mm. ROC curve (B) shows prediction of progression-free survival (PFS) by RLN MAD. AUC is 0.605. Optimal threshold based on Youden index is 6.1 mm, which had sensitivity of 68.7% and specificity of 49.2%.
Fig. 3B —Derivation of optimal retropharyngeal lymph node (RLN) minimal axial diameter (MAD) threshold.
B, Graphs show RLN MAD measurements before treatment in all patients. Histogram (A) shows distribution of RLN MAD measurements. Peak amplitude is at MAD of 6.3 mm. ROC curve (B) shows prediction of progression-free survival (PFS) by RLN MAD. AUC is 0.605. Optimal threshold based on Youden index is 6.1 mm, which had sensitivity of 68.7% and specificity of 49.2%.
Fig. 3C —Derivation of optimal retropharyngeal lymph node (RLN) minimal axial diameter (MAD) threshold.
C, Graphs show RLN MAD measurements before treatment in patients with N0 and N1 disease. Histogram (C) shows distribution of RLN MAD measurements. Peak amplitude is at MAD of 6.1 mm. Restricted cubic spline (RCS) analysis (D) shows prediction of PFS by RLN MAD. Lower bound of confident interval crosses zero at MAD of 6.2 mm. Log = logarithm.
Fig. 3D —Derivation of optimal retropharyngeal lymph node (RLN) minimal axial diameter (MAD) threshold.
D, Graphs show RLN MAD measurements before treatment in patients with N0 and N1 disease. Histogram (C) shows distribution of RLN MAD measurements. Peak amplitude is at MAD of 6.1 mm. Restricted cubic spline (RCS) analysis (D) shows prediction of PFS by RLN MAD. Lower bound of confident interval crosses zero at MAD of 6.2 mm. Log = logarithm.
Fig. 3E —Derivation of optimal retropharyngeal lymph node (RLN) minimal axial diameter (MAD) threshold.
E, Graphs show RLN MAD measurements in patients who underwent evaluation after radiotherapy (post-RT). ROC curve (E) shows prediction of PFS by post-RT RLN MAD measurements. AUC is 0.560. Optimal threshold based on Youden index is 5.5 mm, which had sensitivity of 22.9% and specificity of 90.4%. RCS analysis (F) shows prediction of PFS by post-RT RLN MAD. Lower bound of confident interval crosses zero at MAD of 6.2 mm.
Fig. 3F —Derivation of optimal retropharyngeal lymph node (RLN) minimal axial diameter (MAD) threshold.
F, Graphs show RLN MAD measurements in patients who underwent evaluation after radiotherapy (post-RT). ROC curve (E) shows prediction of PFS by post-RT RLN MAD measurements. AUC is 0.560. Optimal threshold based on Youden index is 5.5 mm, which had sensitivity of 22.9% and specificity of 90.4%. RCS analysis (F) shows prediction of PFS by post-RT RLN MAD. Lower bound of confident interval crosses zero at MAD of 6.2 mm.

Univariable and Multivariable Analyses for Predicting 5-Year Progression-Free Survival

Table 2 summarizes the results of the univariable analyses, including 5-year PFS percentages and the p values for comparison based on log-rank tests. In the entire study sample, significant predictors of PFS included age, pretreatment plasma EBV DNA, T category, N category, tumor stage, treatment modality, RLN MAD, RLN MAD 5 mm and greater, RLN MAD 6 mm and greater, RLN CNN, RLN extracapsular nodal spread, and CLN involvement (all p < .05). In patients who underwent evaluation after RT, significant predictors of PFS included pretreatment plasma EBV DNA, N category, post-RT RLN MAD, post-RT RLN MAD 5 mm and greater, post-RT RLN MAD 6 mm and greater, post-RT RLN CNN, post-RT RLN extracapsular nodal spread, and post-RT plasma EBV DNA (all p < .05).
TABLE 2: Univariable Analysis to Identify Predictors of PFS
VariableAll Patients (n = 1752)Patients Who Underwent Post-RT Evaluation (n = 438)
5-Year PFSapb5-Year PFSapb
Age (y).005.73
Sex .38 .08
Male76.2 (1004/1297) 77.0 (248/317) 
Female78.3 (361/455) 68.8 (85/121) 
Histologic typec .13 .26
WHO type 1 or 268.6 (44/63) 61.9 (10/16) 
WHO type 377.0 (1321/1689) 75.3 (323/422) 
Pretreatment plasma EBV DNA (copies/mL) < .001 .01
< 100084.1 (718/852) 87.8 (75/86) 
1000 to < 10,00070.0 (356/493) 73.8 (125/165) 
≥ 10,00068.7 (291/407) 69.6 (133/187) 
T categoryd < .001 .38
T188.0 (397/451) 80.4 (59/73) 
T278.6 (170/213) 78.0 (43/54) 
T375.4 (522/678) 74.7 (142/187) 
T465.2 (276/410) 69.9 (89/124) 
N categoryd < .001 .02
N089.8 (315/352) 88.1 (45/51) 
N177.5 (783/993) 75.7 (203/263) 
N266.7 (198/289) 69.8 (63/89) 
N353.8 (69/118) 60.9 (22/35) 
Tumor staged < .001 .29
I94.1 (143/151) 100.0 (7/7) 
II87.0 (337/388) 76.5 (69/89) 
III76.2 (550/708) 76.3 (150/194) 
Iva63.8 (335/505) 70.5 (107/148) 
Treatment < .001 .52
IMRT85.3 (193/225) 84.0 (18/21) 
CCRT78.9 (506/635) 75.6 (117/153) 
CCRT and IC72.9 (665/892) 73.6 (198/264) 
RLNe    
MAD (mm)< .001.004
5-mm threshold < .001 .04
MAD < 5 mm87.3 (494/564) 77.6 (251/319) 
MAD ≥ 5 mm71.6 (871/1188) 67.7 (82/119) 
6-mm threshold < .001 < .001
MAD < 6 mm85.1 (593/694) 77.7 (301/382) 
MAD ≥ 6 mm71.2 (772/1058) 55.2 (32/56) 
CNN < .001 .003
Absent80.2 (1096/1348) 78.1 (273/346) 
Present65.0 (269/404) 61.6 (60/92) 
ENS < .001 < .001
Absent80.1 (1122/1385) 79.1 (283/354) 
Present63.1 (243/367) 55.9 (50/84) 
CLN involvemente < .001 .22
Absent85.1 (586/687) 75.7 (292/380) 
Present71.2 (779/1065) 68.5 (41/58) 
Post-RT plasma EBV DNA (copies/mL)  < .001
0 80.0 (305/377) 
1 to < 1000 61.2 (18/28) 
≥ 1000 24.0 (10/33) 

Note—Unless otherwise indicated, values are expressed as percentages with raw data in parentheses. Percentage represents estimated survival probability by Kaplan-Meier method while accounting for censored data. Dashes indicate not applicable. PFS = progression-free survival, RT = radiotherapy, EBV = Epstein-Barr virus, IMRT = intensity-modulated radiotherapy, CCRT = concurrent chemoradiotherapy (i.e., IMRT plus concurrent chemotherapy), IC = induction chemotherapy, RLN = retropharyngeal lymph node, MAD = minimal axial diameter, CNN = central nodal necrosis, ENS = extracapsular nodal spread, CLN = cervical lymph node.

a
Expressed as percentage with value in parentheses.
b
For prediction of PFS, calculated using the log-rank test.
c
According to the 5th edition of the WHO classification of tumors.
d
According to the 8th edition of the AJCC staging system, with N category incorporating the commonly used MAD threshold of 5 mm to determine RLN involvement.
e
Determined before treatment or post-RT MRI, depending on patient group.
Tables 3 and 4 show the results of the multivariable analyses, which assessed RLN MAD at a threshold of 6 mm. In all patients, RLN MAD 6 mm and greater was a significant independent predictor of PFS relative to RLN MAD less than 6 mm (HR = 1.35; 95% CI, 1.05–1.73; p = .02). In patients with N0 or N1 disease, the presence of N1 disease, RLN MAD 6 mm and greater, and absence of CLN involvement was a significant independent predictor of PFS relative to N0 (HR = 1.80; 95% CI, 1.17–2.78; p = .008), and relative to the presence of N1 disease, RLN MAD less than 6 mm, and absence of CLN involvement (i.e., RLN MAD ranging from ≥ 5 mm to < 6 mm) (HR = 1.72; 95% CI, 0.82–3.64; p = .15). In patients who underwent evaluation after RT, RLN MAD 6 mm and greater on post-treatment MRI was a significant independent predictor of PFS relative to RLN MAD less than 6 mm on posttreatment MRI (HR = 1.68; 95% CI, 1.03–2.76; p = .04).
TABLE 3: Multivariable Analysis to Identify Independent Predictors of Progression-Free Survival (PFS)
VariableAll Patients (n = 1752)Patients With N0–N1 Disease (n = 1345)
No. of PatientsHR (95% CI)pNo. of PatientsHR (95% CI)p
Age (y)1.01 (1.00–1.02).0041.01 (1.00–1.02).12
Pretreatment plasma EBV DNA (copies/mL)      
< 10008521 (Reference) 7681 (Reference) 
1000 to < 10,0004931.35 (1.05–1.74).023431.32 (0.98–1.79).07
≥ 10,0004071.17 (0.88–1.55).282341.28 (0.91–1.82).16
T category      
T14511 (Reference) 3851 (Reference) 
T22131.29 (0.85–1.95).231541.05 (0.61–1.82).86
T36781.67 (1.21–2.30).0025041.91 (1.30–2.81).001
T44102.42 (1.73–3.40)< .0013023.02 (1.99–4.57)< .001
Treatment      
IMRT2251 (Reference) 2111 (Reference) 
CCRT6351.07 (0.71–1.62).745320.80 (0.51–1.25).33
CCRT and IC8920.96 (0.63–1.45).836020.72 (0.45–1.16).18
RLN MAD (mm)      
< 66941 (Reference) 
≥ 610581.35 (1.05–1.73).02
N category and RLN MAD      
N03521 (Reference) 
N1, MAD < 6 mm, no CLN771.05 (0.49–2.26).91
N1, MAD ≥ 6 mm, no CLN2561.80 (1.17–2.78).008
N1, CLN6601.89 (1.26–2.81).002
RLN CNN      
Absent13481 (Reference) 11271 (Reference) 
Present4041.35 (1.07–1.71).012181.20 (0.86–1.67).28
RLN ENS      
Absent13851 (Reference) 11671 (Reference) 
Present3671.40 (1.10–1.78).0061781.26 (0.90–1.78).18

Note—All variables were determined before treatment. HR and p values were calculated using multivariable Cox regression analysis. Survival curves are shown in Figure 4. Dashes indicate not applicable. EBV = Epstein-Barr virus, IMRT = intensity-modulated radiotherapy, CCRT = concurrent chemoradiotherapy (i.e., IMRT plus concurrent chemotherapy), IC = induction chemotherapy, RLN = retropharyngeal lymph node, MAD = minimal axial diameter, CLN = cervical lymph node, CNN = central nodal necrosis, ENS = extracapsular nodal spread.

TABLE 4: Multivariable Analysis to Identify Independent Predictors of PFS in 438 Patients Who Underwent Post-RT Evaluation
VariableNo. of PatientsHR (95% CI)p
Pretreatment plasma EBV DNA (copies/mL)   
< 1000861 (Reference) 
1000 to < 10,0001651.99 (1.01–3.92).047
≥ 10,0001871.75 (0.89–3.44).10
N categorya   
N0511 (Reference) 
N12631.50 (0.64–3.51).34
N2891.70 (0.67–4.30).27
N3351.87 (0.67–5.21).23
RLN MAD (mm)b   
< 63821 (Reference) 
≥ 6561.68 (1.03–2.76).04
RLN CNNb   
Absent3461 (Reference) 
Present921.11 (0.64–1.91).72
RLN ENSb   
Absent3541 (Reference) 
Present841.59 (0.92–2.73).10
Post-RT plasma EBV DNA (copies/mL)   
03771 (Reference) 
1 to < 1000282.20 (1.13–4.30).02
≥ 1000336.77 (4.13–11.08)< .001

Note—HR and p values were calculated using multivariable Cox regression analysis. Factors were selected for model reflected significant variables (p < .05) in univariable analyses (Table 2). Survival curves are shown in Figure S2. PFS = progression-free survival, RT = radiotherapy, EBV = Epstein-Barr virus, RLN = retropharyngeal lymph node, MAD = minimal axial diameter, CNN = nodal central necrosis, ENS = extracapsular nodal spread.

a
Assessed before treatment according to the 8th edition of the AJCC staging system, incorporating the commonly used MAD threshold of 5 mm to determine RLN involvement.
b
Assessed on posttreatment MRI.

Kaplan-Meier Analyses in Patients With N0 or N1 Disease

Figures 4A and 4B show the Kaplan-Meier PFS curves using 6-mm and 7-mm thresholds, respectively, in patients with N0 disease; patients with N1 disease, RLN MAD below threshold, and no CLN involvement; patients with N1 disease, RLN MAD above the threshold, and no CLN involvement; and patients with N1 disease and CLN involvement. The 5-year PFS was not significantly different between patients with N0 disease and patients with N1 disease, RLN MAD less than 6 mm, and no CLN involvement (89.8% vs 89.7%, p = .97) or between patients with N1 disease, RLN MAD 6 mm and greater, and no CLN involvement and patients with N1 disease with CLN involvement (77.2% vs 76.3%, p = .96). However, the 5-year PFS was significantly lower for patients with N1 disease, RLN MAD 6 mm and greater, and no CLN involvement (77.2%) and for patients with N1 disease and CLN involvement (76.3%) than for patients with N1 disease, RLN MAD less than 6 mm, and no CLN involvement (89.7%, both p = .03).
Fig. 4A —Kaplan-Meier progression-free survival (PFS) curves in patients with N0 or N1 disease, stratified into four groups: N0 disease (blue line); N1 disease, retropharyngeal lymph node (RLN) minimal axial diameter (MAD) below threshold, and no cervical lymph node (CLN) involvement (yellow line); N1 disease, RLN MAD above threshold, and no CLN involvement (gray line); and N1 disease and CLN involvement (red line).
A, Graph shows PFS curve for 6-mm threshold.
Fig. 4B —Kaplan-Meier progression-free survival (PFS) curves in patients with N0 or N1 disease, stratified into four groups: N0 disease (blue line); N1 disease, retropharyngeal lymph node (RLN) minimal axial diameter (MAD) below threshold, and no cervical lymph node (CLN) involvement (yellow line); N1 disease, RLN MAD above threshold, and no CLN involvement (gray line); and N1 disease and CLN involvement (red line).
B, Graph shows PFS curve for 7-mm threshold.
The 5-year PFS was not significantly different between patients with N0 disease and patients with N1 disease, RLN MAD less than 7 mm, and no CLN involvement (89.8% vs 85.6%, p = .21). In addition, the 5-year PFS was not significantly different between patients with N1 disease, RLN MAD less than 7 mm, and no CLN involvement (85.6%) and either patients with N1 disease, RLN MAD 7 mm and greater, and no CLN involvement (76.1%, p = .06) or patients with N1 disease and CLN involvement (76.3%, p = .05).

Prognostic Stratification of N Category and Tumor Stage When Applying Optimal Threshold

Using the AJCC 8th edition criteria and a 5-mm threshold, 352 patients had N0 disease, and 993 patients had N1 disease. If applying a RLN MAD threshold of 6 mm rather than 5 mm for determining the presence of RLN involvement in the absence of RLN CNN, RLN extracapsular nodal spread, and CLN involvement, then 77/993 (7.8%) patients with RLN MAD less than 6 mm would be downgraded from having N1 disease to N0 disease, resulting in 429 patients with N0 disease and 916 patients with N1 disease. In addition, using the AJCC 8th edition criteria and a 5-mm threshold, 151 patients had stage I disease, and 388 patients had stage II disease. If applying a RLN MAD threshold of 6 mm rather than 5 mm for determining the presence of RLN involvement in the absence of RLN CNN, RLN extracapsular nodal spread, and CLN involvement, then 23/388 (5.9%) patients with RLN MAD less than 6 mm would be downgraded from having stage II disease to stage I disease, resulting in 174 patients with stage I disease and 365 patients with stage II disease.
Figures 5 and S1 show the Kaplan-Meier OS and PFS curves, respectively, stratifying patients on the basis of N category and tumor stage using the 5-mm and 6-mm thresholds. PFS was significantly different between the 6-mm–based N0 and N1 categories (p < .001) and between the 6-mm–based stage I and stage II groups (p = .004). PFS was also significantly different between the 5-mm–based N0 and N1 categories (p < .001) and between the 5-mm–based stage I and stage II groups (p = .01). OS was significantly different between the 6-mm–based N0 and N1 categories (p < .001) and between the 6-mm–based stage I and stage II groups (p = .04). However, OS was significantly different between the 5-mm–based N0 and N1 categories (p = .006) but not between the 5-mm–based stage I and stage II groups (p = .09).
Fig. 5A —Kaplan-Meier overall survival (OS) curves in all patients, showing relationship with retropharyngeal lymph node minimal axial diameter threshold.
A, Graphs show curves stratified by N category, using 5-mm (A) and 6-mm (B) thresholds. Orange shading shows variation in minimal axial diameter threshold only affects N0 and N1 categories. N categories shown are N0 (blue line), N1 (yellow line), N2 (gray line), and N3 (red line).
Fig. 5B —Kaplan-Meier overall survival (OS) curves in all patients, showing relationship with retropharyngeal lymph node minimal axial diameter threshold.
B, Graphs show curves stratified by N category, using 5-mm (A) and 6-mm (B) thresholds. Orange shading shows variation in minimal axial diameter threshold only affects N0 and N1 categories. N categories shown are N0 (blue line), N1 (yellow line), N2 (gray line), and N3 (red line).
Fig. 5C —Kaplan-Meier overall survival (OS) curves in all patients, showing relationship with retropharyngeal lymph node minimal axial diameter threshold.
C, Graphs show curves stratified by tumor stage, using 5-mm (C) and 6-mm (D) thresholds. Stages shown are I (blue line), II (yellow line), III (gray line), and IVa (red line).
Fig. 5D —Kaplan-Meier overall survival (OS) curves in all patients, showing relationship with retropharyngeal lymph node minimal axial diameter threshold.
D, Graphs show curves stratified by tumor stage, using 5-mm (C) and 6-mm (D) thresholds. Stages shown are I (blue line), II (yellow line), III (gray line), and IVa (red line).
A total of 119 patients who underwent evaluation after RT had RLN involvement on post-RT MRI based on RLN MAD 5 mm and greater. If applying a post-RT RLN MAD threshold of 6 mm rather than 5 mm for determining the presence of RLN involvement in the absence of RLN CNN, RLN extracapsular nodal spread, and CLN involvement, then 63 of 119 (52.9%) patients with post-RT RLN MAD less than 6 mm would be downgraded to not having RLN involvement on post-RT MRI. Figure S2 shows the Kaplan-Meier PFS curves in patients who underwent evaluation after RT, stratifying patients on the basis of 5-mm and 6-mm thresholds on post-RT MRI. The 5-year PFS rate was significantly associated with presence of a post-RT RLN MAD 6 mm or greater (HR = 1.68; 95% CI, 1.03–2.76; p = .04) but not with presence of post-RT RLN MAD 5 mm or greater (HR = 1.09; 95% CI, 0.69–1.71; p = .71). In addition, in patients who underwent evaluation after RT, 5-year PFS was 77.7% for post-RT RLN MAD less than 6 mm versus 55.2% for post-RT MAD 6 mm and greater (p < .001), but 77.5% for post-RT RLN MAD less than 5 mm versus 67.7% for post-RT RLN MAD 5 mm and greater (p = .04).

Discussion

This study of 1752 patients with NPC from two hospitals supports the use of a threshold MAD of 6 mm rather than 5 mm for determining the presence of RLN involvement on MRI. Initial ROC and RCS analyses in various patient groups yielded an optimal threshold of 6 mm for further evaluation. RLN MAD 6 mm and greater was subsequently found to be an independent negative predictor of PFS in all patients, patients with N0 or N1 disease, and patients who underwent post-RT evaluation. In patients with N1 disease based on the AJCC 8th edition using a 5-mm threshold but without CLN involvement, 5-year PFS was significantly worse for those with RLN MAD 6 mm and greater than for those with RLN MAD less than 6 mm (i.e., RLN MAD ranging from ≥ 5 mm to < 6 mm). Tumor stage significantly predicted OS when differentiating stage I and stage II using a 6-mm, but not a 5-mm, threshold. Finally, in patients who underwent evaluation after RT, 5-year PFS was significantly associated with post-RT RLN involvement using a 5-mm, but not a 6-mm, threshold.
RLN metastases represent an important prognostic factor in patients with NPC [8, 9]. RLNs are in a deep anatomic location with difficult surgical access [26]. MRI has high soft-tissue contrast and has become established as a routine imaging modality for staging patients with NPC [27]. Morphologic imaging criteria for determining RLN involvement include increased lymph node size, CNN, and extracapsular nodal spread [18, 28]. However, consensus is lacking regarding the minimum size to consider RLNs to be pathologically enlarged.
A MAD of 5 mm and greater was initially proposed as the size criterion for diagnosing RLN involvement in patients with NPC [1214] and was widely adopted into clinical practice in the absence of a recommended threshold within the AJCC 8th edition. Yet, subsequent studies suggested that MAD 6 mm and greater was a better cutoff [1518]. Zhang et al. [15] used follow-up MRI as a reference standard for differentiating benign and metastatic RLNs, finding that a threshold of 6 mm and greater provided an accuracy of 87.5% with a high predictive value; however, most patients in their study were treated with conventional 2D RT instead of the widely used IMRT. To better align with current treatment paradigms, the current study included only patients who underwent IMRT. Earlier studies also had single-center designs and smaller sample sizes than the current study [1218]. Moreover, the presence of CLN metastases may have introduced a bias in multivariable analyses in patients with N0 or N1 disease in prior work, given that patients with RLN and CLN involvement have a significantly worse prognosis than patients with only RLN involvement [13]. Therefore, for purposes of outcome prediction, we stratified patients with N1 disease into separate groups on the basis of the presence or absence of CLN involvement. Finally, the current study incorporated a group of patients who underwent evaluation 3–4 months after RT. All of these various analyses consistently supported RLN MAD 6 mm and greater as a significant predictor of worse clinical outcomes.
Adoption of the proposed threshold for determination of N category and tumor stage could have important clinical implications, as an optimal staging system helps patients receive timely and effective therapy while avoiding unnecessary treatment. Indeed, according to the current NCCN guidelines [11], the differentiation between N0 and N1 disease and between stage I and stage II disease may lead to changes in treatment strategy, such as determination of the gross target volume for RT and the decision to administer concurrent chemotherapy and/or induction chemotherapy. The use of a threshold of 6 mm rather than 5 mm would result in the downgrading of patients with RLN MAD measuring between 5 and 6 mm from N1 disease to N0 disease, in turn also downgrading patients from stage II to stage I disease. These patients with a downgrading of disease status before treatment could receive less intensive therapeutic interventions. Likewise, residual RLN involvement after RT determines whether patients require salvage therapies. The current study found a substantial decrease in the fraction of patients classified as having post-RT RLN involvement using the 6-mm threshold. These patients with downgrading of post-RT disease status could potentially avoid aggressive additional treatments that have associated toxicities. The downgrading of disease before treatment and after RT using the 6-mm threshold was supported by improved prognostication.
This study had limitations. First, it was performed retrospectively and had likely selection bias. In addition, patients who were downstaged for purposes of this investigation on the basis of a 6-mm threshold were still treated as having stage II disease. Thus, whether a prognostic difference would persist in the clinical setting after treatment de-escalation is unknown. Also, the analysis lacked histologic correlation from the evaluated RLNs. Finally, although data were gathered from two hospitals, the derived optimal threshold of 6 mm did not undergo further validation at additional centers.
In conclusion, this study supports a MAD of 6 mm as the optimal threshold for determining the presence of RLN involvement in patients with NPC. In some of the analyses, this threshold provided better prognostic performance than the 5-mm threshold that is commonly used in clinical practice. Given the absence of a defined size threshold in the AJCC 8th edition staging manual, we propose that future updates to the manual incorporate this threshold for N category and tumor stage determinations.

Footnotes

Based on a presentation at the European Congress of Radiology 2023 annual meeting, Vienna, Austria.
Supported by the National Natural Science Foundation of China (no. 82171906), the National Natural Science Foundation of China–Regional Science Foundation Project (no. 82260358), and China Zhongshan City Foundation for Social Welfare and Basic Research Projects (no. 2021B1029).
Provenance and review: Not solicited; externally peer reviewed.
Peer reviewers: Guarang V. Shah, Michigan Medicine; additional individual(s) who chose not to disclose their identity.

Supplemental Content

File (23_29984_suppl.pdf)

References

1.
Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2021; 71:209–249
2.
Xia C, Dong X, Li H, et al. Cancer statistics in China and United States, 2022: profiles, trends, and determinants. Chin Med J (Engl) 2022; 135:584–590
3.
Liu LZ, Zhang GY, Xie CM, Liu XW, Cui CY, Li L. Magnetic resonance imaging of retropharyngeal lymph node metastasis in nasopharyngeal carcinoma: patterns of spread. Int J Radiat Oncol Biol Phys 2006; 66:721–730
4.
King AD, Ahuja AT, Leung SF, et al. Neck node metastases from nasopharyngeal carcinoma: MR imaging of patterns of disease. Head Neck 2000; 22:275–281
5.
Ho FC, Tham IW, Earnest A, Lee KM, Lu JJ. Patterns of regional lymph node metastasis of nasopharyngeal carcinoma: a meta-analysis of clinical evidence. BMC Cancer 2012; 12:98
6.
Ma J, Liu L, Tang L, et al. Retropharyngeal lymph node metastasis in nasopharyngeal carcinoma: prognostic value and staging categories. Clin Cancer Res 2007; 13:1445–1452
7.
Tham IW, Hee SW, Yap SP, Tuan JK, Wee J. Retropharyngeal nodal metastasis related to higher rate of distant metastasis in patients with N0 and N1 nasopharyngeal cancer. Head Neck 2009; 31:468–474
8.
Huang L, Zhang Y, Liu Y, et al. Prognostic value of retropharyngeal lymph node metastasis laterality in nasopharyngeal carcinoma and a proposed modification to the UICC/AJCC N staging system. Radiother Oncol 2019; 140:90–97
9.
Ng WT, Tsang RKY, Beitler JJ, et al. Contemporary management of the neck in nasopharyngeal carcinoma. Head Neck 2021; 43:1949–1963
10.
He T, Yan RN, Chen HY, et al. Comparing the 7th and 8th editions of UICC/AJCC staging system for nasopharyngeal carcinoma in the IMRT era. BMC Cancer 2021; 21:327
11.
[No authors listed]. NCCN clinical practice guidelines in oncology: head and neck cancers, version 2.2023. NCCN, 2023
12.
Tang LL, Guo R, Zhou G, et al. Prognostic value and staging classification of retropharyngeal lymph node metastasis in nasopharyngeal carcinoma patients treated with intensity-modulated radiotherapy. PLoS One 2014; 9:e108375
13.
Shi Q, Shen C, Kong L, et al. Involvement of both cervical lymph nodes and retropharyngeal lymph nodes has prognostic value for N1 patients with nasopharyngeal carcinoma. Radiat Oncol 2014; 9:7
14.
Wu IS, Hung GU, Chang BL, et al. Is unenhanced 18F-FDG-PET/CT better than enhanced CT in the detection of retropharyngeal lymph node metastasis in nasopharyngeal carcinoma? Ear Nose Throat J 2016; 95:178–184
15.
Zhang GY, Liu LZ, Wei WH, Deng YM, Li YZ, Liu XW. Radiologic criteria of retropharyngeal lymph node metastasis in nasopharyngeal carcinoma treated with radiation therapy. Radiology 2010; 255:605–612
16.
Chen J, Luo J, He X, Zhu C. Evaluation of contrast-enhanced computed tomography (CT) and magnetic resonance imaging (MRI) in the detection of retropharyngeal lymph node metastases in nasopharyngeal carcinoma patients. Cancer Manag Res 2020; 12:1733–1739
17.
Tu DG, Chen HY, Yao WJ, et al. Verification of the efficacy of new diagnostic criteria for retropharyngeal nodes in a cohort of nasopharyngeal carcinoma patients. Int J Med Sci 2021; 18:3463–3469
18.
Li YZ, Xie CM, Wu YP, et al. Nasopharyngeal carcinoma patients with retropharyngeal lymph node metastases: a minimum axial diameter of 6 mm is a more accurate prognostic predictor than 5 mm. AJR 2015; 204:20–23
19.
Zhao Q, Dong A, Cui C, et al. MRI-based metastatic nodal number and associated nomogram improve stratification of nasopharyngeal carcinoma patients: potential indications for individual induction chemotherapy. J Magn Reson Imaging 2023; 57:1790–1802
20.
Li S, Luo C, Huang W, et al. Value of skull base invasion subclassification in nasopharyngeal carcinoma: implication for prognostic stratification and use of induction chemotherapy. Eur Radiol 2022; 32:7767–7777
21.
Ma H, Liang S, Cui C, et al. Prognostic significance of quantitative metastatic lymph node burden on magnetic resonance imaging in nasopharyngeal carcinoma: a retrospective study of 1224 patients from two centers. Radio-ther Oncol 2020; 151:40–46
22.
Wan Y, Tian L, Zhang G, et al. The value of detailed MR imaging report of primary tumor and lymph nodes on prognostic nomograms for nasopharyngeal carcinoma after intensity-modulated radiotherapy. Radiother Oncol 2019; 131:35–44
23.
Tang LQ, Li CF, Li J, et al. Establishment and validation of prognostic nomograms for endemic nasopharyngeal carcinoma. J Natl Cancer Inst 2015; 108:djv291
24.
Fu YC, Liang SB, Huang WJ, et al. Prognostic value of lymph node necrosis at different N stages in patients with nasopharyngeal carcinoma. J Cancer 2023; 14:2085–2092
25.
Mao YP, Liang SB, Liu LZ, et al. The N staging system in nasopharyngeal carcinoma with radiation therapy oncology group guidelines for lymph node levels based on magnetic resonance imaging. Clin Cancer Res 2008; 14:7497–7503
26.
Coskun HH, Ferlito A, Medina JE, et al. Retropharyngeal lymph node metastases in head and neck malignancies. Head Neck 2011; 33:1520–1529
27.
Dong D, Zhang F, Zhong LZ, et al. Development and validation of a novel MR imaging predictor of response to induction chemotherapy in locoregionally advanced nasopharyngeal cancer: a randomized controlled trial substudy (NCT01245959). BMC Med 2019; 17:190
28.
Sharma M, Bartlett E, Yu E. Metastatic retropharyngeal lymph nodes in nasopharyngeal carcinoma: imaging criteria. Expert Rev Anticancer Ther 2010; 10:1703–1706

Information & Authors

Information

Published In

American Journal of Roentgenology
PubMed: 37753859

History

Submitted: July 20, 2023
Revision requested: August 2, 2023
Revision received: September 1, 2023
Accepted: September 19, 2023
Version of record online: September 27, 2023

Keywords

  1. minimal axial diameter
  2. nasopharyngeal carcinoma
  3. prognosis
  4. progression-free survival
  5. retropharyngeal lymph node

Authors

Affiliations

Yuliang Zhu, PhD
Nasopharyngeal Head and Neck Tumor Radiotherapy Department, Zhongshan City People's Hospital, Zhongshan, P. R. China.
Chao Luo, MD
State Key Laboratory of Oncology in South China, Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-sen University Cancer Center, Guangdong 510060, P. R. China.
Shumin Zhou, BS
State Key Laboratory of Oncology in South China, Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-sen University Cancer Center, Guangdong 510060, P. R. China.
Haojiang Li, PhD
State Key Laboratory of Oncology in South China, Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-sen University Cancer Center, Guangdong 510060, P. R. China.
Lizhi Liu, PhD
State Key Laboratory of Oncology in South China, Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-sen University Cancer Center, Guangdong 510060, P. R. China.
Kit Ian Kou, PhD
Department of Mathematics, Faculty of Science and Technology, University of Macau, Macao Special Administrative Region, P. R. China.
Feng Lei, PhD
Nasopharyngeal Head and Neck Tumor Radiotherapy Department, Zhongshan City People's Hospital, Zhongshan, P. R. China.
Guoyi Zhang, MD
Department of Radiation Oncology, Cancer Center, the First People's Hospital of Foshan, Foshan, P. R. China.
Di Cao, PhD
State Key Laboratory of Oncology in South China, Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-sen University Cancer Center, Guangdong 510060, P. R. China.
Zhiying Liang, MD [email protected]
State Key Laboratory of Oncology in South China, Guangdong Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy, Guangdong Provincial Clinical Research Center for Cancer, Sun Yat-sen University Cancer Center, Guangdong 510060, P. R. China.

Notes

Address correspondence to Z. Liang ([email protected]).
First published online: Sep 27, 2023
Version of record: Dec 20, 2023
Y. Zhu, C. Luo, and S. Zhou contributed equally to this work.
The authors declare that there are no disclosures relevant to the subject matter of this article.

Metrics & Citations

Metrics

Citations

Export Citations

To download the citation to this article, select your reference manager software.

Articles citing this article

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share on social media