Linear accelerator-based stereotactic radiosurgery for limited, locally persistent, and recurrent nasopharyngeal carcinoma: efficacy and complications.

PURPOSE: To evaluate the efficacy and complication of linear accelerator-based stereotactic radiosurgery (SRS) when used as salvage treatment for early-stage persistent and recurrent nasopharyngeal carcinoma (NPC) after primary radiotherapy (RT).
MATERIALS AND METHODS: Between March 1998 and June 2001, 18 patients (15 men and 3 women; median age 46 years, range 32-84) with locally persistent or recurrent NPC confined to the nasopharynx (rT1) or with limited extension to the nasal fossa or parapharyngeal space (rT2) were treated by SRS. Thirteen patients had rT1 disease and 5 had rT2 disease. Most patients had disease not amenable to surgery or brachytherapy. All patients had undergone previous radical RT. Persistent disease was defined as tumor relapse within 4 months of completion of primary RT, and recurrence as tumor relapse beyond 4 months. Seven patients were treated for persistent disease, eight for a first recurrence, and three for a second recurrence. SRS was performed using multiple noncoplanar arcs of photons delivered to the target volume, which was defined by axial CT at a 3 mm thickness, supplemented by MRI in selected patients (67%). The median target volume was 5.3 cm(3) (range 2.2-16.9). The median SRS dose was 12.5 Gy (range 11-14) delivered to the 80% isodose line. All patients underwent serial nasopharyngoscopy and imaging after SRS. The median follow-up was 26 months (range 11-48).
RESULTS: After SRS, 16 (89%) of 18 patients had complete regression of tumor as assessed by nasopharyngoscopy and biopsy. Four patients with an initial complete response to SRS subsequently developed local relapse again, with one recurrence developing outside the target volume 8 months after SRS and three within the target volume at 6-26 months after SRS. Two patients with local disease controlled by SRS developed relapse in other sites (neck node and liver metastases). The actuarial 2-year local control rate after SRS was 72%. Patients treated for persistent disease had a better local control rate (100%; 7 of 7) than those treated for recurrent disease (46%; 5 of 11). Patients with rT1 disease also had a better outcome after SRS compared with those with rT2 disease, with a control rate of 77% (10 of 13) for rT1 disease and 40% (2 of 5) for rT2 disease. Treatments were well tolerated, with no acute side effects. One patient had radiologic evidence of temporal lobe necrosis, although the right temporal lobe had already received a high dose during prior RT. That patient also developed additional local recurrence and liver metastases and died. The actuarial 2-year survival rate was 86%.
CONCLUSIONS: Our preliminary results indicate that SRS is an effective treatment modality for persistent and recurrent early-stage NPC, with early control rates comparable to other salvage treatments such as brachytherapy and nasopharyngectomy. A modest SRS dose at 12.5 Gy also appears to be effective and is associated with minimal morbidities. More clinical experience and longer follow-up are needed to validate our results and to address fully the role of SRS in salvaging local failures of NPC.