第一部分：TSC-RAML患者基因型与临床表型的相关性研究

目的：评估目标序列捕获二代测序技术在结节性硬化症相关肾脏血管平滑肌脂肪瘤（TSC-RAML）患者基因诊断中的应用价值，分析TSC-RAML患者基因型与临床表型的相关性。

材料与方法：2014年4月至2017年2月北京协和医院泌尿外科收治的共52例患者，其中48例患者临床诊断为TSC-RAML，另4例诊断为散发性肾脏血管平滑肌脂肪瘤（S-RAML），进行目标序列捕获二代测序检测TSC1和TSC2基因，应用经典Sanger测序进行验证，收集患者临床资料，分析TSC-RAML患者基因型与临床表型的相关性。应用SPSS 19.0统计软件处理数据，以P﹤0.05为有统计学意义。

结果：结节性硬化症患者中TSC1或TSC2基因突变检出率为91.7%（44/48）， TSC1基因突变5例，TSC2基因突变39例，TSC1：TSC2为1：7.8，4例S-RAML患者未检测到TSC1或TSC2基因突变。我们共发现了15种新的基因突变，包括9种框移突变，3种大片段缺失突变，2种无义突变及1种剪接突变。与非TSC2基因突变患者相比，TSC2基因突变患者发生色素脱失斑、牙釉质点状凹陷频率显著增加，差异具有统计学意义。TSC2基因突变患者RAML最大直径显著高于非TSC2基因突变患者，差异具有统计学意义。相比于非TSC2基因突变患者，TSC2基因突变患者RAML破裂出血及接受侵入性治疗风险有增加的趋势。

结论：目标序列捕获二代测序技术可准确检测出TSC1和TSC2基因突变，TSC2基因突变患者更易出现色素脱失斑及牙釉质点状凹陷等临床特征，RAML临床表型更为严重，对于TSC2基因突变患者应给予更为积极的监测及治疗。

关键词：结节性硬化症；二代测序；基因型；表型；血管平滑肌脂肪瘤

第二部分：依维莫司治疗中国TSC-RAML患者的疗效及安全性研究

目的：评估mTOR抑制剂依维莫司治疗中国TSC-RAML患者的疗效及安全性，分析影响依维莫司治疗疗效的相关指标。

材料与方法：本研究为一项单臂、开放、单中心的II期临床试验，研究对象为18例确定诊断为TSC-RAML的患者，按照纳入标准及排除标准严格筛选后，符合标准的患者给予依维莫司10mg/d口服，常规随访时间为开始服药后第12、24、48及96周。主要研究终点患者RAML靶病灶总体积缩小≥50%。应用受试者工作特征曲线（ROC曲线）计算灵敏度，特异性和诊断准确度。应用SPSS 19.0统计软件处理数据，以P﹤0.05为有统计学意义。

结果：依维莫司治疗12周、24周和48周时分别有55.6%, 58.8%和62.5%的患者RAML靶病灶总体积缩小≥50%。最常见的不良事件包括口腔黏膜炎、月经失调和腹痛。靶病灶基线最大瘤径（D_L）和靶病灶基线平扫CT平均值（HU_mean）与依维莫司治疗疗效相关。以靶病灶基线HU_mean预测依维莫司治疗疗效，靶病灶基线HU_mean预测靶病灶反应的最佳分界点为-24.50 HU，在12周、24周、48周和治疗最佳反应时间点，ROC曲线下面积分别为0.99、1.00、0.94和1.00，敏感度和特异性均＞90%，而且似然比＞10，显示了良好的预测价值。

结论：依维莫司治疗中国TSC-RAML患者显示了良好的疗效和安全性，但是仍需警惕巨大RAML病灶破裂出血的可能性以及存在恶性潜能的上皮样血管平滑肌脂肪瘤的可能性。此外，靶病灶基线平扫CT平均值（HU_mean）可准确预测依维莫司治疗TSC-RAML的疗效。

关键词：结节性硬化症；依维莫司；疗效；安全性；亨斯菲尔德单位

Part 1: The correlation between genotype and phenotype in patients with TSC-RAML

Objective：To evaluate the application value of targeted next-generation sequencing (NGS) in the genetic diagnosis of tuberous sclerosis complex associated with renal angiomyolipoma (TSC-RAML) and analyze the correlation between genotype and phenotype.

Material and methods：The clinical data of 52 patients, 48 patients of TSC-RAML and 4 patients of sporadic renal angiomyolipoma (S-RAML), were analyzed in PUMCH between April 2014 and February 2017. Mutations of TSC1 and TSC2 gene detected by NGS technology and subsequently confirmed by Sanger sequencing, analyzing the correlation between genotype and phenotype. All analyses used a significance level of 0.05 and were generated in SPSS19.0 software.

Results：91.7% (44/48) patients of TSC-RAML identified TSC1 or TSC2 mutants and the rate of TSC1:TSC2 was 1:7.8, while four S-RAML patients could not detect any mutation in TSC1 or TSC2 gene. Fifteen kinds of novel mutations were reported in our study, including 9 kinds of frameshift mutation, 3 kinds of deletion mutation, 2 kinds of nonsense mutation and 1 splice mutation. Comparing to those with TSC1 gene mutation and no mutation identified (NMI), patients with TSC2 gene mutation have significantly higher frequency with hypomelanotic macules and dental enamel pits. The maximum diameter of RAML was significantly longer than patients in TSC2 gene mutation group, along with a trend of higher frequency of AML bleeding risk and accepted invasive treatment.

Conclusions：NGS is an accurate technology in detecting mutant alleles for patients of TSC-RAML. The patients with TSC2 gene mutations presented higher frequency with hypomelanotic macules and dental enamel pits than those with non-TSC2 gene mutations. Comparing to those with TSC1 gene mutation and NMI group, the phenotype in patients of TSC2 gene mutations were severer than those with non-TSC2 gene mutations.

Key words: tuberous sclerosis complex；NGS；genotype；phenotype；angiomyolipoma

Part 2: Efficacy and safety of everolimus in Chinese patients with TSC-RAML

Objective：To evaluate the efficacy and safety of the mTOR inhibitor everolimus in treating Chinese patients with TSC-RAML and analyze the markers to predict the response of everolimus.

Material and methods：In this single-center open-label phase 2 trial, 18 patients were treated with everolimus (10mg/d). AML volume and computed tomography (CT) parameters were measured at baseline, 12, 24, 48 and 96 weeks. The primary efficacy endpoint was the proportion of patients with confirmed AML response of at least a 50% reduction in total volume of target AMLs relative to baseline. Receiver operating characteristic curve (ROC) method was used to calculate the sensitivity, specificity and diagnostic accuracy. All analyses used a significance level of 0.05 and were generated in SPSS19.0 software.

Results：The response rate were 55.6%, 58.8% and 62.5% at weeks 12, 24 and 48, respectively. The most common adverse events were mucositis oral, irregular menstruation and abdominal pain. Mean HU on unenhanced CT and the longest diameter (D_L) at baseline significantly correlated with percentage change of target AML at different indicated time (p<0.05). The AUC for mean hounsfield units (HU_mean) on unenhanced CT were 0.99, 1.00 and 0.94 at weeks 12, 24, 48 and the time of best reponse, respectively. A threshold of -24.50 HU for predicting AML response showed high sensitivity and specificity (＞90%).

Conclusions：Everolimus is effective and safety for TSC-RAML in Chinese patients.  Mean HU on unenhanced CT of target AML at baseline predicts response to everolimus and it may be useful to select treatment.

Key words: tuberous sclerosis complex, everolimus, efficacy, safety, hounsfield units