Figure 1. A, A T2-weighted magnetic resonance image showing evidence of a previous maxillary swing operation on the left side, with removal of the posterior nasal septum and the left inferior turbinate. Ipsilateral tumor recurrence (arrow) with substantial parapharyngeal extension is evident close to the petrosal internal carotid artery. B, Remaxillary swing operation using the previous facial incision (arrow) and the osteotomy sites. C, After the maxilla is swung laterally, the recurrent tumor is resected en bloc with the pterygoid muscles and the pharyngobasilar fascia, resulting in exposure of the petrosal internal carotid artery (arrow). D, To prevent a subsequent carotid artery blowout, a free vastus lateralis muscle flap is used to cover the exposed petrosal internal carotid artery. The vascular pedicle passes through a tunnel medial to the mandible for subsequent microvascular anastomosis in the neck.
Figure 2. A, A contrast magnetic resonance image showing that the recurrent tumor encases the left petrosal internal carotid artery. B, The first-stage extracranial-intracranial bypass uses the radial artery interposition (arrow) between the external carotid artery in the neck and the middle cerebral artery intracranially. C, A postoperative magnetic resonance angiogram showing that the vascular bypass graft (arrow) is patent. D, The second-stage maxillary swing operation with craniofacial resection was performed 1 week after the bypass surgery, removing the tumor in the nasopharynx with the petrosal internal carotid artery, without neurologic sequelae.
Jimmy Yu Wai Chan, William Ignace Wei. Recurrent Nasopharyngeal Carcinoma After Salvage Nasopharyngectomy. Arch Otolaryngol Head Neck Surg. 2012;138(6):572–576. doi:10.1001/archoto.2012.832
Author Affiliations: Division of Head and Neck Surgery, Department of Surgery, University of Hong Kong Li Ka Shing Faculty of Medicine, Queen Mary Hospital.
Objective To study the safety and oncological outcome of salvage surgery for local tumor recurrence after previous maxillary swing nasopharyngectomy.
Setting University hospital.
Patients Between 1998 and 2011, a total of 252 patients who had local tumor recurrence after previous nasopharyngectomy were recruited for the study.
Main Outcome Measures The locations of the recurrent tumor, operability, and surgical details of resection and reconstruction, as well as the complications and oncological results, were studied.
Results The local recurrence rate after nasopharyngectomy was 13.1%, the risk of which was significantly higher in patients with positive resection margins (39.6% vs 6.9%; P = .006). The chance of the development of local recurrence was significantly lower in patients who had received postoperative chemoradiation therapy than in patients who did not (29.4% vs 64.3%; P = .04). Overall, 63.6% of the patients with local recurrence were amenable to further surgery. Depending on the location of the tumor, remaxillary swing (n = 10), contralateral maxillary swing (n = 8), or central palatal resection (n = 3) was performed. Free-flap coverage of the exposed skull base and the petrosal internal carotid artery was required in 12 patients. There was no evidence of hospital mortality or major complications associated with the surgery. The mean duration of follow-up was 22.4 months, and the overall disease-specific survival in the group of patients who underwent surgery was 80.9%.
Conclusions Surgical salvage for local tumor recurrence after previous nasopharyngectomy is safe, with a good oncological outcome. In the presence of previous maxillary swing surgery, reswing or contralateral swing operation is feasible, without major complications.
Nasopharyngeal carcinoma (NPC) differs from other head and neck malignant neoplasms in its tumor characteristics and treatment outcome.1,2 It is endemic in southern China and Southeast Asia, affecting 10 to 50 individuals per 100 000 population per year.3 The primary treatment is radiotherapy or chemoradiotherapy depending on the stage of disease at presentation,4- 7 while surgery is reserved for persistent or recurrent tumors after the combined therapy.8 However, when the tumor recurs after nasopharyngectomy, the decision for further treatment is complicated, as surgery is difficult and chemoradiation therapy may not be effective. The aim of the present study was to report our experience in treating patients with recurrent NPC after nasopharyngectomy.
Before data collection, approval was obtained by the institutional review board of the University of Hong Kong–Hospital Authority Hong Kong West Cluster. From June 1998 to June 2011, at the University of Hong Kong, Queen Mary Hospital, we performed nasopharyngectomy using the maxillary swing approach in 252 patients with residual or recurrent NPC after previous chemoradiation therapy. The procedure has been described previously,9- 12 with a few modifications over the years.13 Through a Weber-Ferguson-Longmire incision, osteotomies were made on the anterior wall of the maxilla, the zygomatic arch, and the hard palate. Maxillary tuberosity was separated from the pterygoid plate with a curved osteotome. The maxilla ipsilateral to the tumor was then swung out to gain access to the tumor.
The NPC was resected under direct vision with 1.5-cm margins, including the pterygoid muscles, the posterior part of the nasal septum, and the whole posterior wall of the nasopharynx up to the medial edge of the contralateral fossa of Rosenmüller. The need for further resection was guided by the intraoperative frozen-section results. The maxillary osteocutaneous unit was then returned, and the osteotomy sites were fixed with titanium miniplates and screws. A prefabricated dental obturator was fitted to ensure accurate dental positioning.
The formal paraffin section histology report was available 1 week after surgery, and the results were reviewed by the surgeons together with the clinical oncologists. Patients with microscopic tumor involvement of the resection margins (R1 resection) would receive further adjuvant treatment if feasible. For the first year after surgery, the patients were followed up monthly, with clinical and endoscopic examination of the nasopharynx during every visit. Magnetic resonance imaging (MRI) of the nasopharynx was performed, and plasma samples were checked for Epstein-Barr virus DNA every 6 months. These examinations were spaced out progressively during the subsequent years after surgery.
When local tumor recurrence in the nasopharynx was confirmed, contrast MRI was mandatory to assess the location and extent of the tumor. Systemic workup was performed with fludeoxyglucose F 18 positron emission tomography–computed tomography (PET-CT). The tumor was considered inoperable if MRI demonstrated substantial intracranial extension or if there was evidence of systemic metastasis. Otherwise, surgical resection of the recurrent tumor was offered, and the location of tumor determined the surgical approach. Neurosurgeons were consulted if craniofacial resection was anticipated, and plastic surgeons were involved when free-flap reconstruction was required during surgery.
The demographic data of the patients, the information gained at the time of surgery, and the outcome of the patients were analyzed to evaluate the efficacy of the salvage nasopharyngectomy. The t test for quantitative variables and the χ2 test for qualitative variables were used to analyze any relationships. The results were analyzed with SPSS version 18.0 software (SPSS Inc). P < .05 was considered statistically significant.
Of the 252 patients, 180 (71.4%) were male. The median age was 50 years (age range, 24-81 years). Thirty-two patients received concomitant chemoradiation as the initial treatment for the primary tumor because of the advanced stage of disease, while the rest received external radiation only. All patients underwent subsequent nasopharyngectomy via the maxillary swing approach for recurrent tumor. Intraoperative macroscopic tumor clearance was achieved in every patient, although the final pathologic report confirmed microscopic clearance (R0 resection) in only 204 patients (81.0%). Of the patients with positive resection margins, 6 had an inadequate resection because of a discrepancy between the pathology reports at frozen section and subsequent paraffin section. In these patients, further removal of the microscopically involved margins would have been possible at the initial surgery if the frozen-section result was accurate, unlike the rest of the patients with R1 resection, in whom further excision was not possible owing to the involvement of the petrosal internal carotid artery (ICA) or the skull base. Further adjuvant treatment (chemotherapy, radiotherapy, or a combination of both) was eventually administered to 34 patients with positive resection margins.
The median duration of follow-up was 56 months. On postoperative surveillance, 33 patients were found to have local tumor recurrence, resulting in an overall local recurrence rate of 13.1% after nasopharyngectomy. The mean time from surgery to subsequent recurrence was 31 months. The risk of local tumor recurrence in the groups with R0 and R1 resections was 6.8% and 39.6%, respectively (P < .001). Among the patients with tumor recurrence after R1 resection, the risk of local tumor recurrence in the groups with and without postoperative adjuvant treatment was 29.4% and 64.3%, respectively (P = .04). According to the PET-CT findings, none of these patients had systemic metastasis.
The recurrent tumor was located on the ipsilateral side of the nasopharynx in 20 patients (60.6%), on the contralateral side of the nasopharynx in 10 patients (30.3%), and on the nasal side of the soft-hard palate junction in the remaining 3 patients (9.1%). After MRI assessment, only 21 patients (63.6%) with local tumor recurrence were amenable to further surgery. The most common reason for inoperability was substantial intracranial extension with invasion into the cavernous sinus (n = 8 [66.7%]) followed by encasement of the petrosal ICA with substantial paraspinal infiltration (n = 4 [33.3%]). According to the location of the recurrent tumor, the proportion of patients in whom further surgery was considered feasible was 50% for ipsilateral recurrence, 80% for contralateral recurrence, and 100% for nasal floor recurrence.
All patients whose tumors were considered operable agreed to further surgery. Because of the location or extensiveness of disease at the nasopharynx, they underwent salvage nasopharyngectomy via the maxillary swing approach. For those tumors that recurred at the same side of the previous surgery (n = 10), remaxillary swing was performed via the previous facial incision and osteotomy site. For tumors located at the other side of the nasopharynx (n = 8), a contralateral maxillary swing operation was performed. For tumors that recurred at the soft-hard palate junction (n = 3), central fenestration was performed, removing both the soft and hard palate, while preserving the bilateral upper alveolus. The resultant defects were then reconstructed with either an obturator (n = 2) or a microvascular free flap (n = 1). Twelve patients had tumor extension to the parapharyngeal space. Their tumors were resected en bloc with the pharyngobasilar fascia to ensure tumor clearance. Because of the extent of resection, all 12 patients required free-flap reconstruction for coverage of the skull base and the petrosal ICA, and the flaps that were used included the free radial forearm flap (n = 3) and the free vastus lateralis muscle flap (n = 9). In 1 patient, the petrosal ICA was already thrombosed as a result of tumor invasion. The vessel was resected with the tumor, and the patient had no neurologic sequelae after surgery.
All patients underwent macroscopic removal of the recurrent tumor. The rate of R0 resection for the second salvage nasopharyngectomy was 71.4%. All patients survived the operation. The mean hospital stay was 11.4 days. There were no major complications, such as carotid artery blowout or osteoradionecrosis of the skull base, associated with the surgery. One patient (10.0%) developed a palatal fistula after a remaxillary swing operation. The risk of the development of severe trismus, which was defined as an interincisor distance of less than 25 mm, after remaxillary swing and contralateral swing was 20% and 12.5%, respectively.
The mean duration of follow-up among patients with local tumor recurrence after previous nasopharyngectomy was 22.4 months. The overall disease-specific survival of patients who underwent surgery for isolated tumor recurrence at the nasopharynx after previous nasopharyngectomy was 80.9%. Among the patients with an inoperable recurrent tumor, the overall survival was 16.7%, and the mean duration to death was 9.6 months. The remaining 2 patients, who survived despite undergoing no treatment, demonstrated no obvious tumor progression radiologically over time.
Despite the apparent improved outcome of treatment for NPC with intensity-modulated radiotherapy14 and concurrent chemoradiation therapy,5- 7 the result of treatment for residual or recurrent disease remains unsatisfactory. Surgical salvage offers better local tumor control and survival,15- 27 as well as less posttreatment morbidity, than reirradiation.
Over the years, we have used the anterolateral approach for salvage nasopharyngectomy by swinging the maxillary antrum laterally. Not only is the ipsilateral nasopharynx visualized clearly, the contralateral side can also be approached by removing the posterior part of the nasal septum. With such adequate exposure, en bloc removal of the tumor with wide margins is feasible. We have demonstrated a statistically significant difference in the 5-year actuarial control of local disease as well as the 5-year disease-free survival for patients with curative and palliative resection (79% vs 57% and 63% vs 37%, respectively).8 A Cox regression model showed that the tumor size (<1.5 cm in diameter; P = .005) and the negative resection margin status at frozen section (P = .003) were the 2 independent factors that affected the control of disease in the nasopharynx as well as the disease-free survival.
Despite the ability to resect with wide margins, our current study showed that the local recurrence rate after maxillary swing nasopharyngectomy was 13.1%. As expected, the risk of developing local tumor recurrence is significantly higher in patients with positive resection margins from previous surgery (39.6% vs 6.8%; P < .001) (see the “Results” section). Attempts should be made to achieve microscopic tumor clearance during surgery. Intraoperative frozen-section analysis provides an invaluable tool to achieve this goal, although in some situations it may not be completely reliable, as seen in 6 of our patients. It is crucial for surgeons to send the most representable margins to the pathologists to obtain a meaningful frozen-section result. Difficulties arise when immunohistochemical staining is required to distinguish viable tumor cells from the background of atypical epithelial cells from previous radiotherapy. Currently, we are investigating the role of tumor molecular markers in improving the sensitivity and specificity of intraoperative frozen-section examination of the resection margins.
Among patients with positive resection margins after the initial nasopharyngectomy, our study shows that further adjuvant treatment may have a role in improving local tumor control, as evidenced by a lower risk of subsequent local tumor recurrence (29.4% vs 64.3%; P = .04). Currently, there is no consensus on the appropriate type and efficacy of adjuvant treatment in such a scenario, and the results of the current study have to be interpreted with caution because of the potential bias from the retrospective nature of the study, the small sample size included, and the heterogeneity of the adjuvant treatment that is administered. Further investigation is ongoing to delineate its role in such circumstances.
Regular follow-up is mandatory after surgery. There is no single investigation modality that is suitable for all patients, and the complementary use of nasoendoscopy, MRI, and plasma Epstein-Barr virus DNA as well as PET-CT can achieve early detection of tumor recurrence. Even with a stringent follow-up protocol, however, only 63.6% of our patients were amenable to surgery on confirmation of tumor recurrence. This may be partially explained by the fact that recurrent tumors are often characterized by submucosal extension, making early endoscopic detection difficult. The location of the recurrent tumor is one of the most important predictors of the feasibility of further surgery. Patients with recurrence at the ipsilateral nasopharynx are less likely to be candidates for further surgery, as tumors at this location usually recur at the deep margin of the previous operation and thus have a greater chance of infiltrating the parapharyngeal and paraspinal space as well as of extending into the intracranial cavity and involving the cavernous sinus or the brainstem. Tumors that recur in the contralateral nasopharynx or the nasal side of the palate are more likely to be operable, as they are the result of the proliferation of the residual tumor cells at the radial resection margin of the previous operation. These tumors are usually more superficial and easily detected endoscopically and are less likely to have paranasopharyngeal or skull base involvement.
With careful patient selection and detailed preoperative planning, surgical salvage for recurrent tumors after previous nasopharyngectomy is safe and feasible. The mean duration of follow-up of the study group was only 22.4 months. Clearly, a longer observation is required to make a definitive conclusion. However, early results show that the oncological outcome in such cases is clearly superior to those without surgery, demonstrating the importance of a stringent follow-up protocol and the early detection of tumor recurrence while it is still amenable to surgery. Depending on the location of the recurrent tumor, it can be resected via a remaxillary swing approach for ipsilateral tumor recurrence or via a contralateral maxillary swing approach for tumors involving the nasopharynx contralateral to the side of the previous operation. When the petrosal ICA is exposed after tumor ablation, free-flap coverage is mandatory to prevent life-threatening bleeding resulting from carotid artery blowout (Figure 1). For even larger tumors, extracranial-intracranial vascular bypass allows resection of the petrosal ICA en bloc without jeopardizing the cerebral perfusion (Figure 2). There was no hospital mortality in our series, and major potential complications such as carotid artery blowout and osteoradionecrosis of the skull base were effectively prevented by the use of microvascular free flaps. Free muscle flaps (vastus lateralis flap) are preferred over fasciocutaneous flaps (radial forearm flap) because the latter are associated with recurrent crust formation owing to the continuous desquamation of the cutaneous component. This crust is often difficult to remove especially when it is dried, and it can affect the hygiene in the nasopharynx and lead to repeated infection. The muscle flap, on the other hand, will be covered by mucosa within 2 to 3 weeks, and the muscle bulk will undergo atrophy in 3 to 4 months, restoring the patency of the nasal passage. The hospital stay is similar to that after the initial nasopharyngectomy. Wound healing after remaxillary swing and contralateral maxillary swing is satisfactory, and the risks of associated complications are similar to those of the initial operation, although the chance of severe trismus developing may be higher after remaxillary swing because of increased scar formation in the pterygoid region after repeated surgery. Early mouth-opening exercise is recommended for this group of patients
In conclusion, surgical salvage for local tumor recurrence after previous nasopharyngectomy is safe, with a good oncological outcome. Reswing or contralateral swing operation in the presence of previous maxillary swing surgery is feasible, without major complications.
Correspondence: Jimmy Yu Wai Chan, MBBS, MS, FRCS, FCSHK, Division of Head and Neck Surgery, Department of Surgery, University of Hong Kong Li Ka Shing Faculty of Medicine, Queen Mary Hospital, 102 Pokfulam Rd, Hong Kong (firstname.lastname@example.org).
Submitted for Publication: January 27, 2012; final revision received March 5, 2012; accepted March 20, 2012.
Author Contributions: Both authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Chan and Wei. Acquisition of data: Chan. Analysis and interpretation of data: Chan. Drafting of the manuscript: Chan. Critical revision of the manuscript for important intellectual content: Chan and Wei. Statistical analysis: Chan. Administrative, technical, and material support: Chan and Wei.
Financial Disclosure: None reported.