Magnetic Resonance Imaging of Retropharyngeal Lymphadenopathy in Nasopharyngeal Carcinoma

DISCUSSION

The incidence of metastatic lymph nodes in NPC is higher than in other head and neck cancers, with a RLN involvement of 29.1%~89% based on previous literature^[4-7].The difference in the incidence of RLNs which has been higher in various reports may result from different radiological size criteria or different methods of examination. Lam et al.^[5] and King et al. ^[4] found that normal RLNs were smaller than 4 mm or 4.5 mm in their maximum diameter of the shortest axis according to observation respectively in normal groups of 20 and 100 people. Therefore, our study used a maximal axial diameter of ≥5 mm as the size criteria for metastatic RLNs based on the recommendation of published reports. The vast majority of RLN involvement in NPC patients was among the lateral group. Since the medial retropharyngeal nodes are usually not visible, any MRI finding of a medial retropharyngeal node is regarded as abnormal.

The incidence of involvement of RLNs was 64.6% in our study, whereas in a study by Chua et al. ^[7], metastatic rates were 29.1% in NPC patients with metastatic RLNs. However, that study was based on CT imaging, and a dimension of ≥10 mm was used to assign RLN involvement, possibly leading to a higher difference in the incidence of RLN involvement. It is difficult to discriminate RLNs from front muscle of the vertebra from the primary tumor on CT imaging, for they are very similar in density and location. Compared to CT, MRI provides a better option for identification of RLNs and the primary extent into the parapharyngeal space and surrounding normal tissue identification, because MRI shows a higher resolution than CT of soft tissue^[8]. Therefore, MR imaging is superior to CT imaging for delineating the target volume for three dimensional conformal radiotherapy and intensity modulated radiation therapy in NPC cases.

Our study showed that the majority of metastatic RLNs were located superior to the oropharynx level (from the base of skull to the C1/2 disk level), and the multiple metastatic RLNs were common in NPC. Metastatic RLNs at the oropharyngeal level might require special consideration in radiation planning. The definition of RLNs proposed by Som et al.^[2] states that RLNs must lie medial to the internal carotid arteries within 2 cm of the skull base. This definition may be not suitable for NPC, because there are so many RLNs located beyond that region. Based on our observations, from the skull base to tongue bone level may be suitable for NPC.

The lymphatic drainage of carcinoma of the head and neck was associated with the primary tumor location and adjacent subsites to which the tumor spread ^[9]. Drainage regions of the retropharyngeal nodes are the nasal cavities, eustachian tubes, nasopharynx, oropharynx, hypopharynx, pharyngeal wall, and palate^[10]. Tumor invasion into the levator muscle of the velum palatini was associated with a higher risk of RLN metastasis, possibly due to submucosal invasion. The risk of RLN metastasis markedly increased when the primary tumor invaded beyond the confinement of the pharyngobasilar fascia^[7]. Furthermore, a higher incidence of metastatic RLNs was associated with primary tumor invasions into the oropharynx, nasal cavity, pre-styloid space, post-styloid space and tensor muscle of the velum palatini. Drainage regions of the retropharyngeal nodes are located mostly of these subsites, and involvement of these sites is classified as T2 according to the 2002 UICC/AJCC Staging System. Also there was no apparent relationship between RLN metastasis and the laryngopharynx, pterygopalatine fossa, infratemporal fossa, skull base, paranasal sinus or intracranial extension. The structures are not drainage regions of the retropharyngeal nodes except for the laryngopharynx, and involvement of these sites has not markedly increased in the incidence of RLNs spread. The findings of our study suggest that metastatic RLNs were not associated with involvement of the laryngopharynx in this series. The lymphatic drainage of laryngopharynx may primarily drain to the retropharyngeal lymph nodes in NPC cases. The incidence of laryngopharyngeal invasion was very low, so we could not evaluate the association between the laryngopharynx and RLNs.

The regional lymph node spread in NPC includes cervical lymph nodes and retropharyngeal lymph nodes, which have an apparent relationship. A higher incidence of metastatic RLNs was found in patients with cervical lymph node metastases. This finding is in agreement with the report of Chua et al.^[7] We also found that the incidence of metastatic RLNs in patients with bilateral cervical lymph node metastases was lower than in patients with unilateral cervical lymph node metastases. Therefore, the incidence of metastatic RLNs in N2 patients is higher than in N1 patients. A statistically significant association was found between the involvement of RLNs and level II, level III and level V nodes. This was in agreement with the observation that the efferent vessels of RLNs drain to the upper jugular chain and to the posterior triangle ^[10]. This is the reason for the relationship of retropharyngeal lymph nodes and cervical lymph nodes.

The lymphatic vessels drain in two general directions, lateral and medial. The primary drainage is the lateral pathway, which drains the lateral nasopharynx (pharyngeal recess), and flows into the lateral RLNs or into the lateral half of the upper internal jugular chain (group II lymph nodes) by them or directly into group II lymph nodes. The medial drainage pathway drains the roof and posterior wall of the nasopharynx, and flows into the median RLNs penetrating the pharyngobasilar fascia^[9]. Our study revealed that the incidence of metastatic RLNs was almost equal to the incidence of cervical lymph nodes metastases, and was also nearly equivalent to the incidence of cervical level II node metastases. In 41 (17.7%) of the 231 patients with nodal metastases, the metastasis bypassed the RLNs and spread directly to the level II nodes. Of 132 patients with unilateral RLNs involvement, 23.5% had bilateral cervical lymphadenopathy and 8.3% had contralateral cervical lymphadenopathy. These results support the view that another lymphatic pathway exists in NPC, which allows direct lymphatic spread to level II nodes without RLN involvement ^[6,11]. Consequently, these results indicate that both RLNs and cervical level II nodes might be the first-echelon nodes in nasopharyngeal carcinoma.

In conclusion, a high incidence of RLN involvement is present in patients with nasopharyngeal carcinoma. The majority of metastatic RLNs were located superior to the oropharynx level, and the incidence of RLNs showed an orderly decrease from the level of C1 to C3. Both RLNs and cervical level II nodes appear to be the first-echelon nodes in NPC. The patterns of RLN spread, which have been demonstrated in our research, are helpful for the delineation the target area in radiotherapy in three dimensional conformal radiotherapy and intensity modulated radiation therapy. A higher incidence of metastatic RLNs was associated with early stage primary involvement beyond the pharyngobasilar fascia and the upper cervical lymph node metastasis. The implication is that RLN involvement might affect treatment outcome in NPC. The prognostic significance of RLNs in NPC must be reevaluated with a larger follow-up study.