中文摘要：

目的探讨成釉细胞瘤临床特征、病理类型及治疗方法与其复发之间的关系。方法对109例成釉细胞瘤患者的临床病理特征及术后随访资料进行回顾性研究。采用Kaplan-meier方法及Cox回归分析,分析各相关因素与患者复发之间的相关关系,并绘制相关因素下的生存曲线。结果发生于上颌骨者的复发率为50%,高于下颌骨的26.7%（P=0.004）。病程大于12个月者的复发率高于小于12个月者（P=0.002）。刮除术治疗组的复发率为50.9%（27/53）;切除术（包括方块切除、节段性切除及颌骨半切除）治疗组的复发率为7.1%（4/56）,不同术式间的复发率有显著性差异（P〈0.001）。109例成釉细胞瘤患者中,实性型复发率为28.8%（23/80）,单囊型复发率为27.6%（8/29）,两者无显著性差异（P〉0.05）;在实性型成釉细胞瘤中,病理类型与复发有关,其中,丛状型预后较好（P〈0.05）。结论发病部位、病程长短、手术方式及病理类型是影响成釉细胞瘤复发的相关因素,设计治疗方案时,应根据患者的临床特征及病理类型等因素综合考虑。

英文摘要：

Objective To evaluate the relations between the clinical characteristics,pathological types,treatment modalities and the recurrence of all cases of ameloblastoma treated in the Affiliated Stomatological Hospital of Nanjing Medical University from 1985 to 2010.Methods 109 cases of ameloblastoma were retrieved from the medical files.Data were reviewed and statistical analysis was performed using Kaplan-Meier method and Cox＇ regression to check relative significance,predicted recurrence and obtain the disease-free survival function curves of the patients with or without recurrence.Results There was predominant difference in prognosis with regard to the tumor location,with the recurrence rate 50% in maxilla and 26.7% in mandible respectively.The differences between the curettage and wide resection treatment groups in terms of disease-free survival were highly significant（P0.001）.The difference in recurrence rates between the solide ameloblastoma（28.8%） and unicyst ameloblastoma（27.6%） was not significant,but in solide ameloblastoma,the plexiform type showed relatively low potential for recurrence,and the follicular and the other types of ameloblastoma（including granular cell,desmoplastic and acanthomatous type of ameloblastoma） had relatively high likelihood of recurrence（P0.05）.Conclusions The treatment modality,pathological type,complaining time and the location may be associated with the recurrence of ameloblastoma.When we select the treatment modality,the patients＇ age,the location and histopathological type of tumor should be considered comprehensively.