Does MRI-Detected Cranial Nerve Involvement Affect the Prognosis of Locally Advanced Nasopharyngeal Carcinoma Treated with Intensity Modulated Radiotherapy?

Nasopharyngeal carcinoma (NPC) is one of the common cancers in South China. It can easily invade into cranial nerves, especially in patients with local advanced disease. Despite the fact that the magnetic resonance imaging (MRI) findings are not always consistent with the symptoms of CN palsy, MRI is recommended for the detection of CN involvement (CNI). However, the prognostic impact of MRI-detected CNI in NPC patients is still controversial. To investigate the prognostic value of MRI detected CNI, we performed a retrospective analysis on the clinical data of 375 patients with NPC who were initially diagnosed by MRI. All patients had T3-4 disease and received radical intensity modulated radiation therapy (IMRT) as their primary treatment. The incidence of MRI-detected CNI was 60.8%. A higher incidence of MRI-detected CNI was observed in T4 disease compared with T3 disease (96.8% vs. 42.8%, P<0.001), and a higher incidence was also found in patients with Stage IV disease compared with those with Stage III disease (91.5% vs. 42.3%; P<0.001). The local relapse-free survival (LRFS), distant metastasis-free survival (DMFS), and overall survival (OS) of patients with T3 disease, with or without MRI-detected CNI, was superior to that of patients with T4 disease (P<0.05). No significant differences in LRFS, DMFS or OS were observed between T3 patients with or without MRI-detected CNI. The survival of Stage III patients with or without MRI-detected CNI was significantly superior to that of Stage IV patients (P<0.01), but there was no significant difference between Stage III patients with or without MRI-detected CNI for all endpoints. Therefore, when treated with IMRT, MRI-detected CNI in patients with NPC does not appear to affect the prognosis. In patients with clinical T3 disease, the presence of MRI-detected CNI is not sufficient evidence for defining T4 disease.


Introduction
Nasopharyngeal carcinoma (NPC) is one of the common cancers in South China. NPC can easily infiltrate into the surrounding structures, such as parapharyngeal space, cavernous sinus and the cranial nerve (CN) [1][2][3]. The reported incidence of CN palsy in NPC is about 8.9%-10.4% [4][5][6]. Upon development of CN palsy, the prognosis is poor and these patients are staged as T4 disease according to the 7th edition American Joint Committee on Cancer (AJCC) TNM staging system [7,8].
Recently, some researchers reported that Magnetic Resonance Imaging detected (MRI-detected) CN involvement (CNI) in NPC patients, with or without CN palsy, had a poor prognosis and proposed taking MRI-detected CNI into account in a future staging system [4]. Once the proposal is confirmed, it would modify the current staging criteria and affect the clinical treatment strategy, especially in Stage T3 patients with MRIdetected CNI.
With the extensive use of intensity modulated radiation therapy (IMRT), the local control rate has exceeded 90%, even in locally advanced NPC [9][10][11]. However, the prognosis of MRI-detected CNI NPC treated with IMRT has rarely been previously reported. In the present study, the clinical data of 375 local advanced NPC patients treated with IMRT were collected and retrospectively reviewed, the outcomes were evaluated and the prognosis of patients with MRI-detected CNI were analyzed to provide reference for a future staging system revision.
We collected the data of 375 newly diagnosed NPC patients  without distant organ metastasis, who were classified as Stage T3/  T4 according to the 7th edition AJCC TNM staging system and  treated at Fujian Provincial Cancer Hospital between June 1, 2005 and July 31, 2009. The retrospective analysis was approved by Fujian Provincial Cancer Hospital Institutional Review Board. Although consent was not specifically obtained for this retrospective review, all information had been anonymized and deidentified prior to its analysis. The clinical characteristics of patients are summarized in Table 1

Pretreatment workup and diagnosis of MRI-detected CN involvement
Pretreatment work-up included a complete history and physical examination, standard laboratory tests, nasopharynx fiberoptic examination, head and neck MRI scan according to the protocol [12], chest radiograph or CT scan, abdominal ultrasound, bone scan, and dental evaluation.
The diagnostic criteria for MRI-detected CN involvement were as follows [13][14][15]: i) abnormal CN thickening with enhancement after intravenous administration of contrast material; ii) cavernous sinus expansion and abnormal enhancement after enhanced scanning; iii) asymmetrical effacement of the Meckel's cave or gasserian ganglion by enhancing soft tissue; iv) widening and/or destruction of and/or excessive enhancement within neural foramina (Figs. 1, 2, 3).

Treatment
All patients were initially treated with definitive IMRT. A detailed description of the IMRT had been published previously

Follow up and statistical analysis
The median follow-up time was 57 months (range, 6-96 months). Overall survival (OS) was calculated from the first day of diagnosis to the date of death or the last follow-up. Local relapsefree survival (LRFS) was calculated from the first day of diagnosis to the date of local relapse, and the distant metastasis-free survival (DMFS) from the first day of diagnosis to the date of distant metastasis. The outcomes were evaluated in August, 2013.
The survival data were analyzed with SPSS software, version 18.0 (SPSS, Inc., Chicago, IL, USA). The MRI-detected CNI incidence in different T and N classifications were analyzed and compared with the chi-square test. Survival curves were created with the Kaplan-Meier method and compared with the log-rank test. Two-tailed R values ,0.05 were considered statistically significant.

Incidence and prognosis of MRI-detected CN involvement
A total of 228 (60.8%) patients were diagnosed with MRIdetected CNI. The CN V involvement was the most frequent followed by CN III, IV, and XII, sequentially. Only 41 (18%) of MRI-detected CNI patients presented with CN palsy. However, there were a total of five cases of CN palsy that we failed to discover by MRI in four patients, including one patient with CN VII, one patient with CN XI and XII, and two patients with CN V.
The incidence of MRI-detected CNI in different T and N classifications are summarized in Table 2. A higher incidence of MRI-detected CNI was observed in Stage T4 compared to Stage T3 (96.8% vs. 42.8%; P,0.001), as well as in Stage IV compared to Stage III (91.5% vs. 42.3%; P,0.001). These differences were statistically significant. While the difference between different N stages (N0-1 to N2 -3) was not significant (59.8% vs.62.8%, P = 0.567).

MRI-detected CN involvement impact in the current AJCC staging system
To evaluate the impact of MRI-detected CNI in the current AJCC staging system, the Stage T3 and Stage III patients were classified into different categories by MRI-detected CNI: Stage T3 with and without MRI-detected CNI, Stage III with and without MRI-detected CNI.
The survival of Stage III patients with or without MRI-detected CNI was significantly superior to that of Stage IV patients (P, 0.01), whereas the difference in survival between Stage III with and Stage III without MRI-detected CNI patients was not significant for any endpoints (LRFS x 2 = 1.722, P = 0.189; DMFS x 2 = 0.065, P = 0.798; OS x 2 = 0.407, P = 0.524; Fig. 5A

Discussion
CN palsy is common in locally advanced NPC. Although MRI has better soft tissue discrimination than CT, and has been recommended for detection of perineural involvement, the MRI findings are not always consistent with the symptoms of CN palsy [1][2][3]. The incidence of MRI-detected CNI was 60.8% (228/375) and only 18.0% (41/228) patients presented with CN palsy in the current study, which were similar to the reports from Su and Liu [2,4]. A reasonable explanation could be that the process of CNs involvement with tumors was insidious without symptoms or the CNs are resistant to tumor invasion.
In addition, there were a total of five cases of CN palsy that we failed to discover by MRI in four patients. MRI images of these four patients were reviewed, and the failure to detect CNI may be attributed to the following reasons: i) some small branches of CN had been infiltrated by tumor but were too small to be displayed on MRI; ii) although the carotid sheath involvement can be easily detected by MRI, it is difficult to distinguish whether there is accompanying lower CN (CN IX-XII) involvement or not.
The current study revealed that .90% of Stage T4/IV patients presented with MRI-detected CNI, whereas, fewer than 43% of Stage T3/III patients had MRI-detected CNI ( Table 2). The difference was statistically significant. This is because [3]: i) NPC, which originates from the pharyngobasilar fascia, can invade the cavernous sinus and middle cranial fossa through the skull base and infringe on CN II to VI. ii) NPC may also involve the carotid space invading CN XII which exits through the hypoglossal canal, and CN IX to XI which emerge from the jugular foramen (lower CN).
There are numerous available studies regarding patients with CN palsy have a poorer prognosis than those without CN palsy [8,16]. Therefore, the 7th edition AJCC TNM staging system considered CN palsy as Stage T4, and emphasized the importance of clinical evaluation of CN palsy but not the radiology signs of CNI in staging assessments [4,6]. Particularly, the role of radiology findings of CNI in the current staging system is still controversial [4,17].
In the present study, the 5-year LRFS, DMFS, and OS of Stage T3 patients with or without MRI-detected CNI were similar, and were superior to those of Stage T4 patients (Fig. 4A, B, C), suggesting that Stage T3 patients with MRI-detected CN involvement should not be defined as Stage T4 and, thus, avoid excessive treatment. In addition, the survival curves of LRFS, DMFS, and OS for Stage III with or without MRI-detected CN involvement were close, but clearly separated from those for Stage IV (Fig. 5A, B, C), indicating that, regardless of MRI-detected CN involvement, Stage III was associated with a favorable prognosis. Liu and his colleague reported that, the 3-year OS and DMFS of Stage T3 between patients with and without MRI-detected CNI were significantly different (P = 0.0087, and P = 0.0112), whereas, the difference of LRFS between Stage T3 with MRI-detected CNI and Stage T4 was not statistically significant (P = 0.0797) [4]. The inconsistency with the current study may be due to: i) conventional radiotherapy was the main treatment in that study, which may lead to a lower LRFS for T3 patients with MRI-detected CNI; ii) the OS and DMFS may be influenced by N classification, as it is generally accepted, that N classification is a primary independent prognostic factor for the OS and DMFS in the IMRT era.
The IMRT can improve local control and reduce radiationinduced toxicities by providing better tumor target coverage and is significantly better at sparing sensitive normal structures. It has been widely accepted as the standard treatment for locally advanced NPC, although it is still controversial whether early or advanced T Stages of NPC can benefit from IMRT [18,19]. Here we demonstrate for the first time that when treated with IMRT, the survival of NPC patients in Stage T3 or Stage III with and without CNI was not significantly different, and was distinctly superior to that of Stage T4 and Stage IV. However, due to the limitations of this being a single center treatment, the results require confirmation through the accumulation of studies from more cancer centers.
Another limitation in the current study is the lack of unified chemotherapy regimens. Concurrent chemoradiation therapy has been repeatedly proven to improve survival [20,21]. However, only 37.9% (142/375) of the patients received concurrent chemotherapy in the study. Future prospective studies of IMRT combined with concurrent chemotherapy with appropriate chemotherapy regimens and number of chemotherapeutic cycles, may be needed to verify the results.
In conclusion, the current study confirmed the high incidence of MRI detected CNI in local advanced NPC. When treated with IMRT, the LRFS, DMFS, and OS of Stage T3/III patients with or without MRI detected CNI were similar, but significantly better than patients with Stage T4/IV. Patients in Stage T3 with MRIdetected CNI should not be classified as Stage T4. This supports physical examination rather than the radiology imaging to evaluate CN involvement in the current AJCC T4 classification.