[目的]探讨鼻咽癌综合治疗策略。[方法] 回顾性分析无远处转移初治鼻咽癌749例，按'92分期标准对所有病例重新分期。根据患者不同T、N组合的死亡风险比及生存情况，将病人分为5组：早期N0组（T1～N0）早期，N1组（T1～2N1）、局部晚期组（T3～4N2～3）区域晚期组（T1～2N2～3）局部区域晚期组（T3～4N2～3） 。比较各组的5年总生存率、5年累积局部区域复发率、远处转移率和死亡风险比等。[结果]早期N1组较早期N0组，5年总生存率明显下降（84.7%vs95.4%,P=0.0050）5年远处转移率明显升高（10.8%vs0.1%,P=0.0004），死亡风险比分别为1和3.8；局部晚期组与区域晚期组比较，5年总生存率，累积远处转移率及累积局部区域复发率差异均无统计学意义（61.4%vs63.9%、20.3%vs24.6%19.1%vs15.6%，P>0.05），死亡风险比在同一层次，分别为10.1及10.0；局部区域晚期组5年总生存率仅48.2，死亡风险比高达16.4。[结论] 鼻咽癌的治疗应根据患者具体的T、N分期组合进行分层分析治疗，T1～2N0预后好，可予单纯放疗；T1～2N1病人应考虑放疗联合化疗的综合治疗T3～4和/或N2～3患者有较高的复发及转移率，应增加放化疗综合治疗的强度。

To report our initial experience on the feasibility, toxicity, and tumor control using intensity-modulated radiotherapy (IMRT) for retreatment of recurrent nasopharyngeal carcinoma (NPC). Methods and Materials: A total of 49 patients with locoregional recurrent carcinoma in the nasopharynx were treated with IMRT between January 2001 and February 2002 at the Sun Yat-Sen University Cancer Center, Guangzhou, China. The average time to the nasopharyngeal recurrence was 30.2 months after initial conventional RT. The median isocenter dose to the nasopharynx was 70 Gy (range 60.9-78.0) for the initial conventional RT. All patients were restaged at the time of recurrence according to the 1992 Fuzhou, China staging system on NPC. The number of patients with Stage I, II, III and IV disease was 4, 9, 10, and 26, respectively. T1, T2, T3, and T4 disease was found in 4, 9, 11, and 25 patients, respectively. N0, N1, N2, and N3 disease was found in 46, 2, 0, and 1 patient, respectively. Invasion of the nasal cavity, maxillary sinus, ethmoid sinus, sphenoid sinus, and cavernous sinus and erosion of the base of the skull was found in 8, 1, 3, 8, 15, and 20 patients, respectively. The gross tumor volume (GTV) was contoured according to the International Commission on Radiation Units and Measurements (ICRU) Report 62 guidelines. The critical structures were contoured, and the doses to critical structures were constrained according to ICRU 50 guidelines. The GTV in the nasopharynx and positive lymph nodes in the neck received a prescription dose of 68-70 Gy and 60 Gy, respectively. All patients received full-course IMRT. Three patients who had positive lymph nodes were treated with five to six courses of chemotherapy (cisplatin+5-fluorouracil) after IMRT. The treatment plans showed that the percentage of GTV receiving 95% of the prescribed dose (V95-GTV) was 98.5%, and the dose encompassing 95% of GTV (D95-GTV) was 68.1 Gy in the nasopharynx. The mean dose to the GTV was 71.4 Gy. The average doses of the surrounding critical structures were much lower than the tolerable thresholds. At a median follow-up of 9 months (range 3-13), the locoregional control rate was 100%. Three cases (6.1%) of locoregional residual disease were seen at the completion of IMRT, but had achieved a complete response at follow-up. Three patients developed metastases at a distant site: two in the bone and one in the liver and lung at 13 months follow-up. Acute toxicity (skin, mucosa, and xerostomia) was acceptable according to the Radiation Therapy Oncology Group criteria. Tumor necrosis was seen toward the end of IMRT in 14 patients (28.6%). Conclusion: The improvement in tumor target coverage and ignificant sparing of adjacent critical structures allow the feasibility of IMRT as a retreatment option for recurrent NPC after initial conventional RT. This is the first large series using IMRT to reirradiate local recurrent NPC after initial RT failed. The treatment-related toxicity profile was acceptable. The initial tumor response/local control was also very encouraging. In contrast to primary NPC, recurrent NPC reirradiated with high-dose IMRT led to the shedding of tumor necrotic tissue toward the end of RT. More patients and longer term follow-up are warranted to evaluate late toxicity and treatment outcome.