Patient selection
From July 2010 to January 2017, NPC patients who had radical endoscopic necrectomy followed by reconstruction using a nasal septum and floor mucoperiosteum flap at the Sun Yat-sen University Cancer Center were selected for enrollment. This study was approved by the ethical committee of Sun Yat-sen University Cancer Center. Written informed consents were obtained from all patients.
The selection criteria were as follows: (1) a definitive history of radiotherapy, (2) diagnosed with PRNN based on clinical symptoms, imaging examinations, and preoperative biopsy, (3) their PRNN lesion was limited to a resectable region (i.e., the ICA was located at the outside margin of the lesion, the upper bound of the lesion consisted of the roof of the sphenoid sinus and ethmoid sinus, the posterior boundary of the lesion did not exceed the clivus, and the lower edge of axis was located at the lower bound of the lesion) so that the necrotic tissue could be completely removed with curative-intent.
The exclusion criteria for this study were as follows: (1) broad osteoradionecrosis of the skull base exceeding out of the resectable region (e.g. the lower edge of the axis) which would render the removal with curative-intent difficult and risky, (2) pathologically confirmed local recurrence with or without distant metastases before surgery, and (3) physical conditions unsuitable for surgery, such as heart failure, severe cardiac tamponade, unstable hemodynamics, severe nervous system abnormalities, clinically apparent malperfusion including lower limb, cerebral, coronary and renal malperfusion, and visceral ischemia or severe hepatic and renal abnormalities.
Surgery
Prior to surgery, a bacterial cultivation for purulent nasopharyngeal secretions and drug sensitivity assays were performed to identify sensitive antibiotics to be used postoperatively. For patients with ICA exposure in the PRNN lesions, a balloon occlusion test (BOT) was conducted to determine if the ICA could be safely occluded without causing cerebral ischemia [14, 15], and endovascular coiling embolization was required for some patients who suffered from vascular wall trauma or pseudoaneurysm, in order to avoid sudden fatal nasopharyngeal hemorrhage during surgery.
Radical endoscopic necrectomy was performed under general anesthesia using the endoscopic endonasal approach. The nasopharyngeal necrotic soft tissue (Figs. 1a, 2a) was completely removed using the XPS 3000 powered ENT system (Medtronic, Minneapolis, MN, USA) until a slight bleeding was observed (Fig. 1b), thereby exposing the underlying normal fresh tissue. Electrocoagulation was then performed to remove tough connective tissues surrounding the fresh tissues using a long head electrotome (Covidien, Saint Louis, MO, USA). If the skull base bone and anterior arch of atlas were observed to be necrotized, they were removed using a high-speed electric micro-drill (Stryker, Kalamazoo, MI, USA). Surgical specimens from the lateral and basal surgical margins, any tissue suspected to be a recurrent tumor, and the necrotic tissues removed were sent for pathological examination. One side flap, typically the ipsilateral side, was harvested after complete radical endoscopic necrectomy, as reported in our previous study [13]. According to the area of the nasopharyngeal defect, except for a narrow posterior pedicle above the nasal lumen, the flap which contained the posterior nasal septal artery was partially or completely separated and gently rotated backward to cover the defect (Figs. 1c, 2b). If the one-side flap was not large enough to cover the entire defect, the flap on the opposite side of the nasal septum was also simultaneously harvested with the inferior turbinate mucoperiosteum being separated to cover the oropharyngeal defect, especially when the necrotic anterior arch of the atlas was removed. Subsequently, an absorbable gelatin sponge was gently placed into the nasopharyngeal cavity to support the flap. A Rapid Rhino Nasal Pac with Gel Knit (Applied Therapeutics, Tampa, FL, USA) was used to hold up the absorbable gelatin sponge from the nasal cavity to avoid it falling into the oropharyngeal cavity. If the flap necrotized after being harvested, the flap on the opposite side of nasal septum was harvested as a remedial measure. If the flap with two bared sides of the nasal septum were harvested, the nasal septum bone would be removed.
Antibiotic therapy based on the result of bacterial cultivation before the surgery was administered for 5–7 days after surgery. Magnetic resonance imaging (MRI) of the nasopharynx was performed within 2 weeks after the operation to evaluate the flap blood supply (Fig. 2b). When signs of poor blood supply were observed on the MRI, balloon outgassing was performed to relieve the flap compression, and additional treatment with vasoactive drugs was prescribed. After removing the nasal packing on the 7th postoperative day, the patients were prescribed regional perfusion with recombinant human epidermal growth factor derivative liquid (Watsin Genetech Ltd, Shenzhen, Guangdong, P.R. China) to keep the wound humid, and endoscopic nasal cleanup was performed every 2–4 weeks until the flap was completely epithelialized and the entire defect was relined (Figs. 1d, 2c).
Evaluation of treatment efficacy
Locoregional recurrence in this study was defined as tumor recurrence identified in the PRNN lesion with positive pathology at the time of surgery or if the PRNN lesion had negative pathology at the time of surgery but the patient developed recurrence during follow-up examinations.
The primary endpoint of this study was the relining of the nasopharyngeal defect which was defined as the complete re-epithelialization of the nasopharyngeal cavity, upon nasal endoscopic examination, and the mucosa; upon MRI examination [8]. The observed changes in headache episodes and OS were defined as the secondary endpoints. The Numeric Rating Scales (NRS) of pain was used to evaluate headaches level before and within 3 months after surgery. The patients were requested to rate their level of headache on a scale from 0 to 10, with 0 being none and 10 being unbearable. OS was calculated from the date of PRNN surgery till the date of last follow-up (September 20, 2017) or death.
Follow-up
All patients were followed-up to assess the surgical outcomes, disease status, and performance status every 3 months in the first 3 years after the operation, every 6 months in the fourth and fifth years, and annually thereafter. Follow-up examinations included complete physical examinations, chest X-ray, abdominal ultrasonography, MRI of the head and neck, and/or whole-body bone scan.
Statistical analysis
For all cohorts, categorical data were described as numbers and percentages, and continuous data are presented as medians. Categorical data were analyzed using the Chi-square test or Chi-square test with continuity correction. Wilcoxon rank-sum test was conducted for ranked data. Univariate logistic regression was used to determine the associations between clinicopathological characteristics and complete re-epithelialization of the nasopharyngeal defect. Variables associated with severe necrosis (P < 0.1) in univariate logistic regression analysis were included in multivariate logistic regression models to adjust for confounding effects. Survival results were calculated using the Kaplan–Meier method, and their differences were compared using the log-rank test. Covariates significantly associated with prognosis detected in univariate analysis (P < 0.1) were included in a multivariate Cox proportional hazards model to identify independent prognostic factors. All analyses were performed using the SPSS software (version 16.0, SPSS, Chicago, IL, USA), and a two-tailed P value less than 0.05 was considered as statistically significant.