【作者】 廖宁；

【导师】 吴一龙；

【作者基本信息】 南方医科大学 ， 肿瘤学， 2010， 博士

【摘要】 乳腺癌在世界范围内是最常见的恶性肿瘤之一,在我国,乳腺癌在大城市女性恶性肿瘤发病率中已占第一位,且发病率呈持续上升趋势。长期以来国内外乳腺癌治疗的成熟模式一直是多学科的标准化综合治疗。然而由于恶性肿瘤具高度异质性的特点,乳腺癌患者之间存在着明显的个体差异,即使是组织学类型、TNM分期、甚至激素受体状态都相同的乳腺癌,使用临床指引所推荐的同一标准化治疗方案,疗效及预后常有很大差异,因此对乳腺癌患者进行个体化治疗显得极其重要。但传统的临床病理分期的方法并不能有效地区别不同生物学行为的乳腺癌,从而令个体化治疗难以实现。随着分子生物学在乳腺癌领域的发展及应用,已经能够通过基因表达谱在生物学行为层面上把不同类型的乳腺癌区分开来,形成乳腺癌的分子亚型。根据不同亚型乳腺癌各异的生物学行为结合传统的临床病理分期来制定相应的治疗策略、形成个体化治疗的模式正在对以临床试验为基础的传统的标准化治疗模式发起挑战。与此同时,由于新辅助化疗本身是最佳的体内药敏实验,根据肿瘤对化疗的反应可以确定后续的有效治疗方案,是早中期可手术乳腺癌实现个体化治疗的重要途径。另一方面,传统的乳腺癌临床病理分期亦未能有效预测各种治疗手段在乳腺癌中疗效。近年来随着多基因组合检测技术的发展,已经形成了数种能够有效预测乳腺癌患者从化疗中获益程度的多基因组合,成为乳腺癌专科医生制定个体化治疗策略时的重要辅助工具。因此,在中国人群乳腺癌患者中建立基于分子分型的个体化新辅助化疗体系,同时应用多基因组合检测进行疗效的预测以进一步完善并指导个体化治疗,是对提高中国乳腺癌治疗水平的有益探索,具有重要的临床意义。在本研究第1部分中,我们进行了一项单中心开放Ⅱ期临床试验,通过新辅助化疗前对中国乳腺癌患者进行分子分型,对不同亚型乳腺癌选择目前循证医学证据所提供的最佳方案进行新辅助化疗,进而评价其疗效及安全性,初步建立中国人群乳腺癌基于分子分型的个体化新辅助化疗的治疗新模式。在这部分研究中,我们根据肿瘤组织的受体表达状态对中国人群浸润性乳腺癌进行了分子分型。根据乳腺癌细胞受体的免疫组化检测结果,把乳腺癌分为Luminal A, Luminal B, HER2阳性及三阴性四种亚型,并根据各种亚型的生物学行为特点,在循证医学证据的基础上采用个体化的新辅助化疗方案。Luminal A型患者采用卡培他滨联合多西他赛方案,过表达HER2的Luminal B与HER2阳性两种亚型,采用紫杉醇联合曲妥珠单抗方案,而三阴性亚型则采用紫杉醇周疗联合顺铂方案。结果显示,在入组的102例患者中,通过对不同分子亚型的乳腺癌采用相应的新辅助化疗方案,获得了高达36.3%(37/102)的pCR率,提示此种治疗模式可以使DFS与OS获得改善的患者的比例在标准化的基础上至少增加10%。而高达94.1%(96/102)的OR率则提示此种治疗模式可令绝大部分患者获得肿瘤降期提高保乳成功率与改善局部控制效果。进一步的统计分析结果提示,与Ⅲ期的乳腺患者相比,Ⅰ期获得pCR的可能性大幅度提高,而Ⅱ期患者也有提高的趋势,但未达到统计学意义。与Luminal A亚型患者相比,Luminal B亚型获得pCR的可能性是其6.4倍(95%CI 1.66—24.8, P=0.007), HER2阳性亚型获得pCR的可能性是其11.8倍(95%CI 3.05—46.0,P<0.001),三阴性亚型获得pCR的可能性是其3.88倍(95%CI 0.874—17.2,P=0.075)。三阴性患者pCR率与Luminal A亚型患者之间的差异未达到统计学意义,原因是三阴性乳腺癌原本占所有乳腺癌的比例最少,导致本研究入组的三阴性患者仅15例,数据尚不充分。与基于临床病理分期的标准化新辅助化疗模式相比,基于分子分型的个体化新辅助治疗进一步提高了乳腺癌新辅助化疗的疗效,长期疗效尚有待未来的随访数据加以验证。XT、TH、PT三个新辅助化疗方案在中国人群相应的乳腺癌分子亚型中疗效明显,耐受性良好。生物学行为恶性程度较高的非luminal A型乳腺癌以及肿瘤负荷较小的早期乳腺癌患者更容易从个体化新辅助化疗中获益,进一步改善长期的预后。进一步采用多基因表达谱检测对占乳腺癌总数66.7%(67／102)且Luminal A和Luminal B两型患者加以分层区分新辅助化疗获益程度大小,为制定更加精确的个体化治疗策略提供依据。ER阳性乳腺癌(包括Luminal A和Luminal B两种亚型)与HER2阳性及三阴性乳腺癌最大的不同之处在于管腔型乳腺癌除了化疗之外,内分泌治疗对全身疾病的控制也占据重要的地位。而且这两种亚型的乳腺癌占本研究入组患者的66.7%,其中luminal A型患者仅17.0%获得pCR,而luminal B型患者的pCR率与HER2阳性型及三阴性相比也要略低一些(分别为47.6%,57.9%和53.3%)。因此,进一步区分ER阳性乳腺癌患者对化疗的获益程度,使患者避免不必要的化疗,对于完善个体化治疗意义重大,而多基因表达谱的检测是实现分层的可行手段。21基因表达谱检测RS评分在激素受体阳性的乳腺癌中已被证实可用于预测患者自辅助化疗或辅助内分泌治疗中的获益程度,评价局部复发、远处转移或死亡的风险,已经成为激素受体阳性乳腺癌个体化辅助治疗策略的重要参考指标。然而仅有少数小样本量研究将其用于新辅助化疗疗效的预测,并且目前并未有来自中国人群的数据加以验证。因此,我们在本研究的第2部分内容中,对第1部分研究的临床研究中纳入研究的ER阳性患者的肿瘤组织进行21基因表达谱的检测,以获得中国人群乳腺癌21基因表达谱的新鲜组织的检测数据,进一步验证21基因RS评分是否可以用于预测中国人群ER阳性乳腺癌自新辅助化疗中的获益程度,进一步完善个体化治疗体系。目前全球范围内的21基因表达检测大体上使用21-gene oncogype DX标准检测试剂盒,其检测标本来源于石蜡包埋的组织块,技术难度大。我们通过严格的质量控制建立起一套以新鲜冻冻组织为检测来源标本的多基因表达谱检测体系,并获得良好的效果。新鲜组织冰冻病理切片严格质控使肿瘤组织比例大于80%的标本方能进入检测流程,减少了非肿瘤细胞对检测结果的干扰。而从核酸提取、逆转录、荧光PCR检测都建立了一套多基因联合检测标准化流程和质量控制方案。我们所获得的高质量21基因表达谱的数据也说明,本研究所建立采用新鲜冰冻组织提取RNA进行基因表达谱检测的方法与技术平台稳定可靠,可以进一步推广应用。我们随后对65例激素受体阳性的中国人群乳腺癌患者的新鲜冻存肿瘤组织进行了21基因表达谱的检测,获得了每例患者的RS评分。RS评分的计算结果显示,本研究中65例激素受体阳性乳腺癌患者的RS评分平均值为46.3±2.26,RS最低分为11.3,最高分为100,中位RS分数为45分。而且仅有9例患者的RS评分小于31分,其余86.7%的患者RS评分均大于31分。统计分析结果表明RS评分可以预测患者对新辅助化疗的疗效,其显著性优于临床病理分期及HER2受体状态这两个传统的预后预测指标。相对于多数基于21 gene Oncogype DX检测的RS=31的中位分值,本研究所获得的RS分值明显偏高。可能的原因有：首先本研究采用的是来源于新鲜冰冻组织标本的RNA进行多基因表达谱的检测,技术平台体系本身就存在着较大的差异。其次,本研究纳入的患者数尚少,且所有患者非随机入组患者,其RS分布本身可能存在偏移。再次,文献报道的数据基本上来源于西方人群,而我们的数据来源于中国患者,此种人种差异本身就可能导致基因表达谱的特点存在着较大的不同。另一方面,本研究RS分级的截断值也不同于21基因oncotype DX。由于基于我们所开发的技术平台检测获得的基因表达谱的计算结果显示大于85%的患者的RS值大于31,继续采用RS=31分作为截断值显然不适用于中国患者。通过对RS值与pCR状态作受试者工作特征曲线(receive operating characteristic curve, ROC),结果发现RS截断值在44附近时其敏感性及特异性均较好。当RS取44时,ROC图显示对PCR的判断敏感性为83.3%,特异性为57.4%。在将截断值调整为44后,以RS>44为高RS组,RS≤44为低RS组,则低RS组占44.6%,高RS组占55.4%。进一步的Cross Table检验结果显示,高RS组患者获得pCR的可能性是低RS组的6.19倍,95%CI 1.58-24.28,χ2=7.869,P=0.005,与文献报告的5.0的OR值相近,提示我们以RS=44分作为截断值可以很好的区分患者自新辅助化疗获益程度的差异。pCR从本质上来讲仍然是一个临床病理学指标,而RS则是一个基于基因表达谱的分子生物学指标。在行新辅助化疗的激素受体阳性的患者中,RS与pCR对于长期疗效的预测效果何者更佳是一个有趣的问题。传统的临床病理学指标是否会被新兴的分子生物学指标所代替尚需长期的随访数据来进行检验。本部分的研究已经明确基于新鲜冰冻组织RNA来源的21基因表达谱检测技术稳定可靠,根据该平台检测所得结果计算出的RS评分可以用于预测中国人群ER阳性乳腺癌患者自新辅助化疗中的获益程度。对于RS非高分的患者,由于自新辅助化疗获益程度有限。更多还原

【Abstract】 Breast cancer is the most common malignant disease worldwide. In China, the incidence of breast cancer is the highest in all cancers for females in big cities with an tendency of inclination. For a long time, the mature treatment model for breast cancer either domestic or abroad is multiple disciplinary standardized comprehensive treatment. However, for the highly heterogeneous characteristic for malignant disease, there are obvious individual differences between breast cancer patients. Even patients with the same histologic type, TNM staging or even the same hormonal receptor status undergoing the same standard treatment recommended by clinical guidance display quite different treatment efficacies and final outcomes. Thus individualized treatment for breast cancer is of great importance. Nonetheless, traditional clinicopathologic staging is limited in effectively stratifying breast cancers with various biological behaviors drawing back individualizing treatments.With the development and application of molecular biology in breast cancer, this disease can be classified in the level of malignant biological behavior via multiple gene expression detection into molecular subtypes. The novel treatment model that making treatment decisions based on various biological behavior of different breast cancer with traditional clinopathologic staging translated into individualized treatment is now challenging the conventional standardized treatment model based on clinical trials.Meanwhile, neoadjuvant chemotherapy itself is the best in vivo drug sensitive experiment. According to response to chemotherapy help deciding effective treatment regimens later on. Therefore, it is an important pathway towards individualized treatment for operable breast cancer. On the other hand, conventional clinopathologic staging is not effective in the prediction for the efficacies of various treatments in breast cancer. In recent years, with the development of multiple gene expression detection, several gene set that can effectively predict how much benefit breast cancer patients can get from chemotherapy have been developed and become important aiding tools for specialists in the making of individualized treatment strategies.Therefore, establishing individualized neoadjuvant chemotherapy system based on molecular subtype classification in Chinese breast cancer patients and using multiple gene expression detection for efficacy prediction to further improve individualized treatment is of importance in clinical practice.In the first part of this study, we carried out a single center open-labeled stageⅡclinical trial. By performing molecular subtype classification for Chinese breast cancer patient before neoadjuvant chemotherapy and provide best regimen based on updated evidences, we established a treatment model of individualized treatment.In this part of the study, we designated molecular subtypes according to the receptor status of tumor cells. Luminal A, luminal B, HER2 positive and triple negative subtypes were established and individualized neoadjuvant chemotherapy regimens were used. For luminal A subtype, capecitabine combined with docetaxel were used. For the luminal B and HER2 positive subtypes that overexpressing HER2, trastuzumab combined with paclitaxel was used. For triple negative breast cancer, weekly paclitaxel combined with cisplatinum was used. Breast conserving surgery or mastectomy with complete axillary lymph node dissection were performed after 4 cycles of therapy. Pathologic complete remission rate and overall response rate were calculated accordingly.Our results showed that, in the 102 patients included, individualized neoadjuvant chemotherapy obtained pCR rate as high as 36.3%(37/102), suggesting this novel treatment model can at least gain a 10% increase in patients that have improved DFS and OS. Statistical analyzes showed that, stage I patients were more likely to achieve pCR. Compared to Luminal A subtype, other subtypes were more like to achieve pCR, odds ratio were 6.4(95%CI 1.66—24.8, P=0.007),11.8(95%CI 3.05—46.0,P<0.001) and 3.88 (95%CI 0.874—17.2,P=0.075), respectively.Compared to standardized neoadjuvant chemotherapy based on clinopathologic staging, the novel treatment model established in this study further improved the efficacy of neoadjuvant chemotherapy. Long term survival benefit is waiting for future follow-up data for validation. XT, TH and PT neoadjuvant regimens have good efficacy and well tolerability in Chinese breast cancer patients. More aggressive subtypes other than luminal A and early stage patients with less tumor burden are more likely to benefit from individualized neoadjuvant chemotherapy. Further study using multiple gene expression detection to stratifying the two subtypes that account for 66.7% of all breast cancer and sensitive to hormonal treatment is important for making more delicate indivicualized treatment strategies.The distinct difference between luminal type breast cancers is hormonal receptor positive and sensitive to hormonal therapy. Luminal A and B subtypes accounts for the majority of breast cancer in this study while their pCR rate is lower than HER2 positive subtype and triple negative breast cancer. Thus to further stratify patients of luminal types who benefit little from therapy and treating with single hormonal therapy for luminal A patients or combined with HER2 targeted therapy will greatly improved the individualized treatment system established in the previous part of this study.21 gene recurrence score RS has been demonstrated to be powerful in the prediction of benefit from adjuvant chemotherapy or hormonal therapy or local recurrence, distant metastasis or death risk in hormonal receptor positive breast cancer. However, only few small scaled report using this RS score in the prediction for neoadjuvant chemotherapy benefit without and data based on Chinese patients.Therefore, in the second part of this study, we performed 21-gene expression detection for hormonal receptor positive patients included in the first part of this study to obtain data based on Chinese breast cancer patients using RNA originated form fresh frozen tissues. We tried to further validate whether RS can be used to predict benefit from neoadjuvant chemotherapy in Chinese HR positive breast cancer patients.Currently, the 21-gene OncotypeDX kit has been widely used for the detection of 21 gene expression using RNA generated from paraffin-embedded tissues. By strict quality control, we established a multiple gene expression system using RNA generated from fresh frozen tissues with satisfactory and reliable results.21-gene expression detection was performed in 65 HR positive Chinese patients using fresh frozen tumor tissues. RS was calculated according to methods reported in other studies. The mean RS in our study was 46.3±2.26, ranging from 11.3 to 100, with the median RS 45. Only 9 patients had a RS lower than 31. Statistic analyzes showed that RS can independently predict pCR after neoadjuvant chemotherapy with significance level better than staging and HER2 status. Compared to the median RS 31 in most other studies using 21-gene OncotypeDX, RS in our study was obviously higher. Possible reasons were:different detection system, small case number with non-randomization of our study and different races. On the other hand, the cutoff point for RS in our study was different from other studies. Obvious, the distribution of RS in our cohort is different that most patients had a RS greater than 31. After performing ROC figure, we found that if the RS cutoff point is set at 44, the sensitivity and specificity would be both acceptable. If patients were divided into high RS group with RS>44 and low RS group with RS≤44, patients in the high RS group were more like to achieve pCR (odds ratio 6.19, 95%CI 1.58-24.28,χ2=7.869, P=0.005). This result was similar to the 5.0 odds ratio reported by other studies, suggesting that cutoff point set at 44 can effectively decide benefit from neoadjuvant chemotherapy.Originally, pCR is still a clinopathologic parameter while RS is a biological marker based on multiple gene expression. It is interesting that which one of RS and pCR is more powerful in the prediction for long term survival benefit patients obtained from neoadjuvant chemotherapy in HR positive patients. To answer this question, we need long term follow-up data in the future.The second part of this study demonstrated that 21-gene expression detection using RNA generated from fresh frozen tissues is reliable and RS calculated based on data obtained from this detection system can predict benefit from neoadjuvant in Chinese HR positive patients. For non high RS score patients, little benefit will be obtained from chemotherapy. Whether neoadjuvant single hormonal therapy or combined with targeted therapy can still benefit these patients deserves further study.更多还原