Radiologic criteria of retropharyngeal lymph node metastasis in maxillary sinus cancer

Maxillary sinus cancer (MSC) is a relatively rare disease, with an incidence of 1 per 100 000 person-years, and accounts for 3% of all head and neck cancers. Most MSC cases are squamous cell carcinoma (SCC), representing 60% to 90% of total cases [1]. Tobacco use is a significant risk factor. The prognosis of MSC with lymph node (LN) metastases is significantly poor. Five-year overall survival (OS) of patients with SCC of the paranasal sinuses is approximately 50%, and 30% among those with localized and regional disease [2].

Surgery and (chemo)radiation therapy are recommended as the standard treatments for MCS [2]. The combination of radiation therapy and intra-arterial (IA) chemotherapy is a promising treatment for patients who had unresectableMSC or refused surgery because of its high control rate and high OS [3, 4]. This is a popular treatment in Japan, and although it is not a standard method, we have observed good results in patients treated with locally advanced MSC using this combination [5]. In this setting, we have performed elective nodal irradiation (ENI) only for patients with regional LN metastasis (cN +). From the results of a meta-analysis, the guidelines of the National Comprehensive Cancer Network (NCCN) recommended ENI for patients with advanced MSC [6], but ENI is controversial because of reports of increased late adverse events comprising dysphagia and trismus [7]. We reported that ENI increased the radiation dose in the pharyngeal contractile muscles and increased late aspiration pneumonia [8]. The irradiation field setting of the retropharyngeal lymph node (RLN) region is important because it affects the radiation dose of the pharyngeal contractile muscle and may cause severe acute mucositis and late dysphagia in ENI [9‐12]. Recently, intensity-modulated radiation therapy (IMRT) for MSC has been reported to be useful in reducing adverse events [2]. Similarly, IMRT highlights the risk of marginal recurrence, and a more accurate irradiation field setting is required [13]. Because of the high frequency of metastatic RLNs detected by magnetic resonance imaging (MRI) in recent years [14], the European Society for Radiotherapy & Oncology (ESTRO) guidelines recommended including the RLN area as a prophylactic region for radiation therapy in MSC [15]. However, radiologic criteria for metastatic RLNs in MSC have not been established.

Therefore, in this study, we determined metastatic RLNs on the basis of the therapeutic response of RLNs using MRI images taken before and after treatment and investigated the appropriate radiologic criteria for metastatic RLNs in MSC.

There were few reports of RLNs in MSC before 2000 [1], but in recent years there have been reports of a high frequency of RLNs detected by MRI [14]. The inconsistent frequency of metastatic RLNs may be due to changes in diagnostic modalities and the diagnostic criteria [8]. In a previous study, a maximal axial diameter of 10 mm or longer was used as the diagnostic criterion for metastatic RLNs of head and neck cancer [18]. Recently, from the examination of RLNs in healthy subjects by MRI, a minimal axial diameter of 5 mm or longer was often used as the diagnostic criterion [19, 20]. However, it is controversial to extrapolate the MRI findings of healthy individuals to the diagnostic criteria of cancer patients. Regarding RLN for MSC, there are reports of diagnostic criteria of 5 mm or more for the minimal axial diameter and 8 mm or more for the major axial diameter, and the diagnostic criteria are not constant [14, 21]. Zhang et al. reported that the optimal diagnostic criterion was a minimal axial diameter of 6 mm or longer for metastatic RLNs in nasopharyngeal cancer on the basis of the treatment response by MRI before and after treatment [22]. Pathological findings in the RLN region are difficult to obtain due to the risk of adverse events such as vessel and nerve damage and dysphagia by dissection, and it is particularly useful to determine the diagnostic criteria for metastatic RLNs by imaging f/u [23]. Therefore, we revealed the optimal diagnostic criteria for RLNs for MSC from MRI f/u.

Judgment of metastatic RLNs by MRI f/u was determined according to RECIST ver. 1.1. Notably, in the case of continuous SD, the pretreatment image and the most changed image were compared to prevent an erroneous SD judgment. As a result, there were no cases of SD to PR or PD. On the basis of the therapeutic response, seven in 34 RLNs were classified as positive for metastasis. RLNs are classified as lateral and medial, and medial RLNs are usually judged to be metastatic by visual recognition, regardless of size [24]. There was no medial RLN, and thus there was no effect on the size in the radiologic criteria. Although not statistically significant, the maximal axial diameters of metastatic positive RLNs tended to be greater than those of negative RLNs, which reflected the findings of Zhang et al. Metastatic RLNs were more common at the occipital bone level than in previous studies [20], which may be because the primary site of this study included skull base invasion in many cases. The predictive value between anatomical location was not found (data not shown). All these metastatic RLNs were present on the ipsilateral side despite being advanced cases, and this result may indicate that prophylactic irradiation of the RLN region is sufficient on the ipsilateral side alone.

From the ROC curve analysis, metastatic RLNs tended to be diagnosed more accurately with the minimal axial diameter than with the maximal axial diameter. This was consistent with previous studies of RLNs in head and neck cancer [22, 24]. The most appropriate radiologic criterion was a minimal axial diameter of 5 mm or more, similar to the results reported by Zhang et al. in nasopharyngeal cancer. Thus, the diagnostic criterion revealed by this study was the same as those commonly used for head and neck cancer from the examination of RLNs in healthy subjects by MRI [19, 20]. Regarding the case with false positive in this criterion, the therapeutic response was judged to be SD. However, the f/u was performed only once, and caution is required in its interpretation. There were three RLNs with central necrosis, all of which were metastatic RLNs by MRI f/u and the specificity was 100%. The minimal axial diameter in one of three cases with central necrosis was 4.4 mm, and the addition of this finding to the diagnostic criterion by minimal diameter increased sensitivity slightly. However, the only findings of central necrosis of a small LN with a minimal axial diameter of approximately 5 mm were considered to have low sensitivity. That is, it was determined that the findings of central necrosis could be replaced by the criterion of minimal diameter, because only one case of central necrosis deviated from the criterion by minimal diameter.

This study has several limitations associated with its retrospective design. First, the modality and protocol of MRI performed before and after treatment were not standardized. Second, no pathological examination was performed, and positive metastasis was only identified by MRI f/u.

In this study, no RLN recurrence was observed, and the necessity of irradiation in the RLN region remains unknown. In recent years, there have been several reports of RLN recurrence (Table 5) [25‐29], and in the future, the requirement for irradiation in the RLN region should be examined by evaluating the recurrence type using our radiologic criteria, taking into consideration any adverse events.

In conclusion, the most appropriate radiologic criterion of metastatic RLNs in patients with MSC is a minimal axial diameter of 5 mm or more detected by MRI f/u.