Detailed analysis of recovery process of cranial nerve palsy after IMRT-based comprehensive treatment in nasopharyngeal carcinoma

Background Cranial nerve (CN) palsy due to cancer involvement has been considered as an unfavorable prognostic factor for patients with nasopharyngeal carcinoma (NPC). We assessed the role of IMRT based treatment on the recovery of CN palsy and investigated the prognostic value of complete recovery of CN palsy. Methods A total of 115 NPC patients with cancer-related CN palsy were included in the study. We referred CTCAE version 5.0 to evaluate the grade of CN palsy. Results All patients with grade 1 CN palsy recovered completely during the 2 years of follow-up after definite treatment. Most grade 2 palsy could change gradually to grade 1 palsy or complete recovery during 2 years of follow-up. Patients with more than 2 symptoms of CN palsy had poor 3-year disease-free survival (DFS) than these with 1 or 2 symptoms (60.3% vs. 84.9%, HR 0.25, 95% CI 0.07–0.89, P = 0.001). There were no significant differences for PFS, OS, DMFS and LRFS between patients with complete recovery and non-complete recovery from CN palsy after receiving IMRT based comprehensive treatment. Conclusions IMRT based comprehensive treatment could effectively promote the recovery of tumor-related CN palsy for NPC patient. More than 2 symptoms of CN palsy was a poor prognostic factor for DFS of NPC patients. The prognostic role of complete recovery of CN palsy was not identified in our study. Supplementary Information The online version contains supplementary material available at 10.1186/s13014-021-01846-x.


Introduction
Nasopharyngeal carcinoma (NPC) is a cancer arising from the nasopharyngeal mucosal lining, from where the tumor can invade adjacent anatomical organs. Hence, depending on the involvement of anatomical structures, the clinical symptoms varies considerably from person to person, including epistaxis, nasal obstruction, hearing impairments, headache and cranial nerve (CN) palsy. The incidence of CN palsy is approximate 8 to 12.4% in NPC [1]. CN palsy is classified as T4 stage according to the American Joint Committee on Cancer/Union for International Cancer Control (AJCC/UICC). It has been considered as an unfavorable prognostic factor for NPC patients [2,3]. Several studies reported patients with complete recovery from CN palsy had better overall survival (OS) than these without complete recovery after receiving conventional radiation technique or 3-dimensional conformal radiation therapy (3D-CRT) based comprehensive treatment. Moreover, multiple CN palsy and Zang et al. Radiat Oncol (2021) 16:118 long recovery duration were associated with poor OS [4,5]. Therefore, the recovery of CN palsy seems to predict a better survival outcome before intensity modulated radiation therapy (IMRT) era. However, the detail of recovery process of CN palsy has not been investigated by these studies.
Undoubtedly, IMRT is a "Game-Changing" radiation technique for NPC due to the improved dosimetric properties. Locoregional control and survival have been improved by IMRT, and toxicity has been reduced [6,7]. Compared with conventional radiotherapy and 3D-CRT, IMRT could improve the local control from 84.7 to 90.5% for NPC [7]. Induction chemotherapy followed by IMRT based concomitant chemoradiotherapy (IC + CCRT) and concomitant chemoradiotherapy (CCRT) are alternative treatment options for patients with CN palsy according to National Comprehensive Cancer Network (NCCN) guidelines and guidelines of Chinese Society of Clinical Oncology (CSCO). These treatment modalities result in more than 90% local control rate for patients with locoregionally advanced NPC [8,9]. However, it is still unknown whether the superior local control rate could lead to a high relief rate of CN palsy. In this study, we investigate the recovery process of different symptoms of CN palsy, and explore the influence of treatment-related factors on the recovery of CN palsy and the prognostic value of CN palsy recovery in the era of IMRT. To our knowledge, this is the first study focus on the detailed recovery process of CN palsy in NPC patients receiving IMRT-based treatment modalities.

Patients
We reviewed the NPC database from cancer-specific database platform in XiJing Hospital between January 1, 2012 and December 31, 2016 and 514 patients with histologically proved nasopharyngeal squamous cell carcinoma treated with radiotherapy were identified. Of these patients, 121 (22.4%) had CN palsy symptoms. Six patients were excluded: 2 received radiotherapy alone, 3 did not complete the treatment course, and 1 received palliative treatment. The remaining 115 NPC patients with CN palsy were included in the study. This retrospective review was approved by our institutional review board and was in accord with the Helsinki Declaration of 1975 as revised in 1983.

Diagnosis and evaluation
The diagnosis of CN palsy depended on clinical symptoms physical examination, and magnetic resonance imaging (MRI) with or without thin slice computerized tomography imaging (CT) of skull base. CN invasion was detected on MRI that met one of the following criteria: (1) enhancing soft-tissue tumor along the course of the ipsilateral-related nerve, replacing the normal structures of the CN on gadolinium-enhanced T1-weighted images; or (2) perineural spread, defined as an enlargement or abnormal enhancement of the nerve, obliteration of the neural fat pads adjacent to the neurovascular foramina, and neuroforaminal enlargement [10,11]. Physical examinations of CN palsy were conducted by two radiation oncologists every time from pretreatment to the whole follow-up period. Two radiologists reviewed all the imaging records and disagreements were resolved by consents. CN palsy was defined as cranial nerve invasion detected by imaging combined with nerve palsy symptoms resulted by tumor invasion. Because of lack of criteria for CN palsy grade, we referred Common Terminology Criteria for Adverse Events (CTCAE) version 5.0 to evaluate the grade of CN palsy. Grade 1 palsy was defined as asymptomatic, clinical or diagnostic observation only and intervention not indicated. Grade 2 palsy was defined as moderate symptoms with limiting instrumental activities of daily living. Graded 3 palsy was defined as sever symptoms with limiting selfcare.

Treatment
All patients were treated with definitive radiotherapy. The detail of radiotherapy process was described in our previous studies [12,13]. The selections of radiation dose for patients depend on tumor volume, response to induction chemotherapy, age and (Eastern Cooperative Oncology Group, ECOG) score. Patients with higher primary tumor volume (> 55.5 ml) and less than 50% tumor regression after induction chemotherapy would receive higher radiation dose. In this study, 34 of 115 patients (29.6%) received fixed-gantry IMRT technique, 81 (70.4%) patients received volumetric modulated arc therapy (VMAT) technique.

Definition and statistical analysis
Complete recovery was defined as complete disappearing of CN palsy symptoms. The follow up time was calculated from the date of treatment completion to either the date of death or the last review date. Patients were reexamined at least every 3 months in the first 2 years, then every 6 months for up to 5 years or until death. The grade of CN palsy symptoms was recorded in detail by two radiation oncologists. The survival endpoints included overall survival (OS), progression-free survival (PFS), distant metastasis-free survival (DMFS) and locoregional recurrence-free survival (LRFS). OS was measured from the date of treatment completion to death; PFS was measured from the date of treatment completion to the date of disease progression or death from any causes; DMFS was defined as the time from end of treatment to first detection of distant metastasis; LRFS was defined as the time from end of treatment to first detection of local or lymph node region relapse.
The variation of measurement data in groups were compared by means of Student's t test, and data was presented as means ± standard error of the mean (SEM) values. Differences in proportions between groups were evaluated by χ 2 test. Numerical variable was transformed to categorical variable using median as cut-off if it was not Gaussian distribution, such as tumor volume. The Kaplan-Meier method was used to estimate the endpoints, survival curves were compared using the log-rank test. Cox proportional hazard model was used to identify potentially independent prognostic factors, and the proportional hazards assumption was tested with Schoenfeld residuals. The hazard ratio (HR) and its 95% confidence interval (95% CI) were used to indicate the prognostic value of risk factors. A two-sided P value of less than 0.05 was considered significant.

Patient characteristics and incidence of CN palsy
Patient characteristics are summarized in Table 1. The male/female ratio was 3.8:1. Median age was 49 years (range 18-83 years); 83.5% patient ages were less than 60 years. Most patients (77.4%) had regionally metastatic lymph nodes at the time of diagnosis of NPC. All patients had non-keratinizing squamous cell carcinoma. Sixtythree patients (54.8%) received induction chemotherapy. The median radiation dose was 72.62 Gy.
Trigeminal nerve (TN, CN V) disorder was the most common CN palsy symptom (90.4%), followed by abducens nerve (AN, CV VI) disorder (36.5%), oculomotor nerve (ON, CN III) disorder (22.6%), hypoglossal nerve (HN, CN XII) disorder (6.1%) and facial nerve (FN, CN VII) disorder (3.5%). Of all patients, 68 patients (59.1%) had one symptom of CN palsy, 30 patients (26.1%) had two symptoms of CN palsy and 17 patients (14.8%) had more than three symptoms of CN palsy. Because the incidences of HN disorder and FN disorder were very low, we just focused on the recovery of TN disorder, AN disorder and ON disorder in this study.

Correlation between duration of CN palsy and symptom grade
According to the grade definition of CN palsy, no patients had grade 3 symptom in this cohort. Patients with grade 2 CN palsy had significantly longer duration of symptom before clinical diagnosis. The mean duration time were 4.22 ± 0.29 months versus 1.32 ± 0.19 months between patients with grade 2 and grade 1 TN disorder, 3.72 ± 0.41 months versus 1.86 ± 0.37 months between grade 2 and grade 1 AN disorder, 3.14 ± 0.57 months versus 1.41 ± 0.35 months between grade 2 and grade 1 ON disorder, with significantly statistical differences (Fig. 1).  Fig. S1). The grade of CN palsy before treatment had an impact on the recovery of CN symptom. All patients with grade 1 CN palsy recovered completely during the 2 years of follow-up after definite treatment (Fig. 2a-c). Most grade 2 TN, AN and ON palsy could change gradually to grade 1 palsy or complete recovery during 2 years of follow-up ( Fig. 2d-f ). The complete recovery rate of grade 2 TN, AN, and ON palsy were 67.5%, 58.8%, and 61.1%, respectively. However, there was still a small percentage of patients who had persistent grade 2 TN (13.3%), AN (13.8%), and ON (6.6%) palsy ( Fig. 2d-f ).

Influence of treatment-related factors on complete recovery of CN palsy
Treatment-related factors associated with complete recovery of CN palsy were explored including radiation doses and different treatment modalities. Using median radiation dose (72.62 Gy) as dichotomy, patients were divided into two groups: low radiation dose groups (≤ 72.62 Gy) and high radiation does group (> 72.62 Gy). There were no significant differences in number of CN palsy symptom between the groups of radiation dose (data was not shown). For grade 2 of TN palsy, there were more patients received radiotherapy with low radiation dose than those received high radiation dose (58.1% vs. 41.9%, P = 0.049). For grade 2 of AN palsy, there were more patients received radiotherapy with low radiation dose than those received high radiation dose (65.5% vs. 34.5%, P = 0.047). For grade 2 of ON palsy, the proportions of patients received low radiation dose and high radiation dose were not different (44.4% vs. 55.6%, P = 0.749). After comparison of recovery rate between the groups of radiation dose, there were no differences of complete recovery rates of TN palsy and ON palsy between two radiation dose groups (Fig. 3a, c). However, for AN palsy, patients received higher radiation dose (> 72.62 Gy) had higher complete recovery rate than those received lower radiation dose (≤ 72.62 Gy), with significantly statistical differences between two groups in 12 months (28.3% vs. 51.2%) and 18 months (14.2% vs. 38.3%) after treatment (Fig. 3b). There were no significant differences in complete recovery rates of TN palsy and ON palsy between patients treated with IC + CCRT versus CCRT. Compared with CCRT, IC + CCRT had a trend to improve the recovery rate of AN palsy, but not reached significant difference ( Fig. 3d-f ).
The prognostic value of pretreatment duration of CN palsy was analyzed in different types of CN palsy. A significant difference was observed for DFS (P = 0.046) in patients with pretreatment duration within 2 months of TN palsy compared with those with longer than 2 months (Additional file 2: Fig. S2a), but no significant differences were observed for OS (P = 0.252), DMFS (P = 0.09) and LRFS (P = 0.052) between two groups (Additional file 2: Fig. S2b-d). Prognostic value of pretreatment duration of AN and ON palsy were not detected for every survival endpoint (Additional file 3: Fig. S3, Additional file 4: Fig. S4). There were no significant differences for PFS, OS, DMFS and LRFS between patients with complete recovery and non-complete recovery from CN palsy after receiving IMRT based comprehensive treatment (Additional file 5: Fig. S5).

Discussion
CN palsy due to tumor infiltration of cranial nerve implies the locally extensive lesion in NPC. According to the 8th edition AJCC staging system, CN palsy is classified as T4 disease. In the era of conventional radiotherapy and 3D-CRT, several studies reported number of CN palsy, palsy duration before treatment, recovery duration of palsy and complete recovery of CN palsy were associated with prognosis of patients with locoregionally advanced NPC [4,5]. But there were lack of detailed information about the recovery process of CN palsy. This is the first study focus on the recovery process of CN palsy in patients received IMRT based definitive treatment.
Consistent with previous study, we found the most common CN palsy were trigeminal nerve palsy (90.4%), abducens nerve palsy (36.5%), and oculomotor nerve palsy (22.6%) in local advanced NPC [14]. We found that the longer duration of symptom was associated with the more severe symptoms, indicating more damage to the nerve. In this study, up to 77% of patients with TN, AN, or OS palsy could completely recovered within 2 years after treatment, indicating the efficacy of IMRT in complete recovery of upper CN palsy in NPC. AN and ON palsy in NPC were often due to intracranial involvement of tumor, such as involvement of cavernous sinus. Compared to IMRT, it was difficult to precisely deliver radiation dose to this area using conventional or 3D-CRT technique. IMRT has been shown to be associated with better 5-years local control compared with conventional RT and 3D-CRT [7,[15][16][17]. Therefore, the recovery of TN, AN and ON palsy might result from tumor ablation by IMRT. The recovery of HN and FN palsy were not investigated in this study due to low incidence of these symptoms. So it was still unclear whether IMRT could lead to better recovery rate for the two types of palsy compared with conventional or 3D-CRT technique.
We showed that the pretreatment grade of CN palsy was significantly associated with recovery process of CN palsy. All grade 1 palsy achieved complete recovery within 2 years after treatment. However, only 70% of patients with grade 2 palsy had complete recovery, although all of them had some improvement. The  radical radiation dose of NPC was routinely up to 70 Gy in the tumor target for patients received IMRT [18,19]. Even complete response was achieved after induction chemotherapy, the dose of radiotherapy was still up to more than 66 Gy in the target of primary tumor and lymph nodes [20]. In this study, the radiation dose, using median dose as dichotomy, was not associated with the recovery of TN and ON palsy. But for AN palsy, the recovery rate was much higher in patients received higher than 72.62 Gy. Recently, IC followed by CCRT was confirmed to be a more effective regimen than CCRT alone for locoregionally advanced NPC. The 3-year OS was improved by 5% with IC + CCRT [8,9]. Therefore, we compared the recovery of CN palsy in patients treated with IC + CCRT to those treated with CCRT. We found that except for recovery of AN palsy in which IC + CCTRT had a trend to improve recovery, there were no significant correlation between recovery of CN palsy and treatment regimens. These results implied AN palsy might be relieved more easily in NPC patients after IMRT based comprehensive modality. Previous study reported multiple CN palsy and long duration of pretreatment symptom were worse prognostic factors for OS in locoregionally advanced NPC in the era of conventional radiotherapy [5]. In this study, we found patients with more than 2 symptoms of CN palsy had worse prognosis for DMFS but not for OS. Several previous studies suggested that intracranial involvement of NPC might be potential routes for hematogenous dissemination due to anatomically rich venous plexus [21,22]. In this study, 75% of patients with more than 2 symptoms of CN palsy had intracranial involvement (data was not shown), which may contribute to a higher risk of distant metastasis. Except the pretreatment duration of TN palsy was a prognostic factor for DFS and had potential predictive role to impact on the OS, the prognostic value of pretreatment duration of AN and ON were still unclear due to small sample size. Different from previous studies, we did not find prognostic value of complete recovery of CN palsy in this study. Advances in radiation technique and comprehensive treatment modalities might weaken the prognostic role of complete recovery for survival.
In multivariate analysis, consistent with previous studies, N stage was found to be a significant prognostic factor for DFS, DMFS and OS. We failed to detect prognostic value of CN palsy parameters for patient survival in multivariate analysis. Moreover, we could not find association of tumor volume with the number of CN palsy in this study (Additional file 6: Fig. S6), indicating tumor volume is not necessarily related to the extent of CN involvement in locoregionally advanced NPC.