Salvage brachytherapy for patients with locally persistent nasopharyngeal carcinoma

Purpose: Locally persistent nasopharyngeal carcinoma (NPC) carries an increased risk of local failure if additional treatment is not given. This study was conducted to evaluate the outcomes of patients with locally persistent NPC as treated by high-dose-rate (HDR) intracavitary brachytherapy, and to explore whether routine brachytherapy boost could improve the local control. Methods and Materials: Eighty-seven patients with locally persistent NPC treated during 1990-1998 with HDR intracavitary brachytherapy were retrospectively analyzed. Fibreoptic nasopharyngoscopy was performed 3-6 weeks after completion of the primary external radiation therapy (ERT). Biopsies were only taken from suspicious areas. Those with complete regression of local disease were put on observation. Eighty-seven patients were shown to have persistent viable disease at a median time of 6 weeks post-RT. The distribution according to Ho's staging system at initial diagnosis was as follows: Stage I-8, II-33, III-41, IV-5; T1-19, T2-48, T3-20; N0-32, N1-22, N2-28, N3-5. CT scan for restaging was not performed after the documentation of persistent disease. Our policy was to treat all patients with persistent disease with brachytherapy irrespective of the extent of disease just prior to brachytherapy. They were treated with HDR intracavitary brachytherapy, with either cobalt sources or an iridium source, giving 22.5-24 Gy in 3 weekly sessions in all but 4 patients. This dose was prescribed at a distance of 1.5 cm from the center of the surface as defined by the sources in the first six patients and subsequently reduced to 1 cm for the others. Twelve patients were treated with neoadjuvant chemotherapy. To compare the efficacy of brachytherapy, another 383 consecutive nonmetastatic patients, treated with curative intent by ERT, during the years 1990-1993, were evaluated. Multivariate analysis was performed using the Cox regression proportional hazards model. Results: The 5-year actuarial local failure-free survival (LFFS) rates and disease-specific survival rates for the brachytherapy group and ERT group were 85% and 76.6% (p = 0.15), and 72% and 67.8% (p = 0.2), respectively. The corresponding 5-year actuarial LFFS rates for T1, T2, and T3 disease were 94.7%, 88.2%, 67.4%, and 84.1%, 79.8%, 62.6%. In assessing the local control, only the T staging was significant on multivariate analysis (p = 0.0004). Other parameters such as age, sex, and persistence of disease (giving brachytherapy) were all nonsignificant. Complications were comparable between the two groups. In the persistent group, the local failure rates of the patients treated with and without neoadjuvant chemotherapy were 17% (2/12) and 13% (10/75) respectively. When analyzed according to different brachytherapy sources, the 5-year LFFS rates of the T1, T2, and T3 patients treated with iridium and cobalt sources were 100% vs. 85.7 (p = 0.19), 93.6% vs. 70% (p = 0.04), and 67.7% vs. 60% (p = 0.72). The difference was statistically significant for the T2 groups. When early T-stage (T1 and T2) patients were grouped together for analysis, the iridium group again showed a statistically significant improvement in 5-year LFFS rate when it was compared with the cobalt group (95.3% vs. 76.5%, p = 0.03) and the ERT alone group (95.3% vs. 81.5%, p = 0.03). The improvement of local control is attributed to a higher nasopharyngeal mucosal dose that is achieved by using small-size flexible applicators with an iridium source. It is interesting to note that the 5-year LFFS rates for the ERT alone group (T1: 84.1%, T2: 79.8%, T3: 62.6%) are comparable to the corresponding rates of the cobalt group. This information supports our speculation that an adequate booster treatment could compensate for inadequate primary treatment. The prognosis of patients with locally recurrent NPC is grave. Maximizing the local control is therefore crucial for the survival of the patients. The good local control of early T-stage locally persistent NPC patients receiving brachytherapy supports the incorporation of brachytherapy into the primary treatment program. The 5-year LEES rates for T3 disease treated with and without brachytherapy were unsatisfactory (iridium group: 67.7%; cobalt group: 60%; ERT alone group: 62.6%). A more aggressive approach to improve the local control of Ho's T3 disease should be explored. Conclusion: Our experience suggests that locally persistent NPC can be effectively salvaged by brachytherapy. The local control of patients with early T-stage NPC at initial diagnosis is even better than those having clinical remission of local disease at the completion of ERT. Furthermore, it is possible that routine brachytherapy boost, after the completion of ERT, could improve the local control. The exact benefit, however, can only be elucidated by prospective randomized studies. (C) 2000 Elsevier Science Inc.