乳腺癌前哨淋巴结术中分子诊断及其对非前哨淋巴结转移的预测研究

发布时间：2018-08-25 11:06

【摘要】：目的:本研究旨在评估乳腺癌前哨淋巴结(Sentinel Lymph Node,SLN)术中分子诊断技术一步核酸扩增法(One-Step Nucleic Acid Amplification,OSNA)的诊断价值,并根据OSNA检测结果分析其对非前哨淋巴结(Non-Sentinel Lymph Node,NSLN)转移的预测能力,进一步建立乳腺癌术中快速预测NSLN转移的预测模型,以期有效指导乳腺癌后续治疗。方法:本研究包括2010年OSNA临床试验入组的552例患者及此次试验本中心前瞻性入组的348例患者,共计入组患者900例,有效病例为870例。术中根据SLN的质量及短轴长度进行切分。若SLN质量100 mg,则不进行切分,仅术后行逐层切片病理检测;若SLN质量为100~1200 mg,则垂直于短轴进行切分。短轴长度≤4mm时,则切分为a和b两个组织块,a组织块术中行OSNA检测,b组织块行病理检测;短轴长度4 mm时,则切分为a,b,c,d四个组织块,a、c组织块术中行OSNA检测,b、d组织块行病理检测。其中a,b,c,d组织块均行术中印片细胞学(Touch Imprint Cytology,TIC)检测。一方面,OSNA检测结果同术后逐层连续切片病理作比较,另一方面根据OSNA结果建立NSLN转移风险的术中快速预测模型。结果:1、OSNA检测的诊断价值(1)以病例数为统计对象分析,以术后病理为金标准,OSNA与FS、TIC比较以病例数为统计对象分析,OSNA的敏感性、准确性、特异性、阳性预测值、阴性预测值分别为88.7%、89.0%、89.1%、72.3%、96.1%。FS的敏感性、准确性、特异性、阳性预测值、阴性预测值分别为77.9%、94.0%、100.0%、100.0%、92.3%。TIC的敏感性、准确性、特异性、阳性预测值、阴性预测值分别为83.0%、93.5%、96.8%、89.3%、94.7%。OSNA的敏感性显著优于FS(88.7%vs.77.9%,P=0.0095),与TIC相比并无统计学差异(88.7%vs.83.0%,P=0.1254)。(2)以淋巴结数为统计对象分析,以术后病理为金标准,OSNA与FS、TIC比较以淋巴结数为统计对象分析,OSNA的敏感性、准确性、特异性、阳性预测值、阴性预测值分别为85.4%、91.4%、92.6%、68.4%、97.1%。FS的敏感性、准确性、特异性、阳性预测值、阴性预测值分别为68.6%,94.3%,100.0%,100.0%,93.5%。TIC的敏感性、准确性、特异性、阳性预测值、阴性预测值分别为81.0%,94.4%,96.9%,83.2%,96.5%。OSNA的敏感性显著优于FS(85.4%vs.68.6%,P=0.0021),与TIC相比并无统计学差异(85.4%vs.81.0%,P=0.82)。(3)OSNA与FS、TIC对于宏转移及微转移的敏感性比较对于术后伴有宏转移灶的SLN,OSNA检出的敏感性是93.1%,FS检出的敏感性是88.5%,TIC检出的敏感性是92.2%,OSNA的敏感性与FS及TIC相似(P=0.1907,P=0.8543),并无统计学差异。对于术后伴有微转移灶的SLN,OSNA检出的敏感性是65.3%,FS检出的敏感性是40.4%,TIC检出的敏感性是44.0%,OSNA的敏感性均显著优于FS及TIC(P=0.0121,P=0.0139)。2、NSLN转移风险的术中快速预测模型(1)建立基于分子诊断预测NSLN转移的列线图模型对建模组103例患者行单因素分析发现,原发肿瘤大小(P=0.001)、脉管浸润(P=0.007)、SLN总肿瘤负荷(Total Tumor Load,TTL)(P=0.000)、最大转移灶大小(P=0.000)、SLN阳性数(P=0.000)、SLN阴性数(P=0.000)、SLN阳性数/SLN总数(P=0.000)均与NSLN转移相关。Logistic多因素回归分析后发现TTL(P=0.002)、原发肿瘤大小(P=0.013)、SLN阳性数(P=0.000)、SLN阴性数(P=0.001)是NSLN转移的独立相关因素,利用TTL、原发肿瘤大小、SLN阳性数及SLN阴性数建立预测乳腺癌NSLN转移的列线图模型。建模组患者NSLN转移预测值的受试者工作特征曲线(Receiver Operating Characteristic Curve,ROC)下面积(AUC)为0.814。(2)本模型的预测性验证研究本研究利用验证组61例患者对建立的预测模型进行了验证,得出AUC为0.842。(3)影像学评估的肿瘤大小与原发肿瘤大小对模型预测性差异的比较本研究利用验证组患者影像学(超声、钼靶、乳腺磁共振)评估的肿瘤大小数据替代术后病理评估的原发肿瘤大小数据,对模型进行了验证,得出AUC为0.838,与前述验证性ROC曲线下面积相比并无统计学差异,P=0.7406。(4)本模型对腋窝淋巴结分期pN1及≥pN2患者的预测性研究本模型区分患者pN1及≥pN2腋窝淋巴结分期的预测风险临界值为45.4%,AUC为0.861,P0.0001。另外,为探究术中利用影像学肿瘤大小快速预测腋窝淋巴结分期情况,利用本模型得出验证组患者中pN1及≥p N2的预测风险临界值为33.5%,AUC为0.839,P0.0001。(5)本模型与其他预测模型的验证性比较利用入组患者分别对美国安德森癌症中心(MD Anderson,MDA)的模型及法国Tenon医院的模型进行了验证,得出AUC分别为0.745及0.623。另外,本模型对西班牙Isabel等建立的模型进行了验证,得出AUC为0.834。通过对比分析,本模型与MDA及Tenon模型相比有统计学差异,与Isabel等人的模型相比并无统计学差异。结论:OSNA检测作为一种客观的标准化技术,可以术中快速、准确的对SLN做出诊断,且操作简单,大大降低了病理医师的主观性和工作负担,可作为SLN术中诊断乃至术后诊断的首选,适合在中国推广。基于OSNA术中分子诊断结果建立的术中快速预测NSLN转移的模型明显优于其他预测模型,对后续腋窝的处理、放疗靶区勾画及全身治疗具有更好的指导价值。
[Abstract]:Objective: To evaluate the diagnostic value of one-step Nucleic Acid Amplification (OSNA), a molecular diagnostic technique for sentinel lymph node (SLN) metastasis in breast cancer, and to analyze its predictive ability for non-Sentinel Lymph Node (NSLN) metastasis according to the results of OSNA detection. Methods: This study included 552 patients enrolled in the 2010 OSNA clinical trial and 348 prospective patients enrolled in the center of this trial. A total of 900 patients were enrolled in the study and 870 effective patients were enrolled in the study. If the length of SLN is less than 4 mm, it is divided into a and B tissue blocks. When the length of the short axis is less than 4 mm, OSNA detection is performed in a tissue block, and pathological examination is performed in B tissue block. When the length of the short axis is 4 mm, it is divided into a, B and B tissue blocks. The intraoperative Touch Imprint Cytology (TIC) was performed in all the tissues of a, b, C and D. On the one hand, the results of OSNA were compared with the pathology of successive slices after operation, on the other hand, the risk of NSLN metastasis was established according to the results of OSNA. Results: 1. The diagnostic value of OSNA (1) was analyzed by case number and pathology after operation as golden standard. The sensitivity, accuracy, specificity, positive predictive value and negative predictive value of OSNA were 88.7%, 89.0%, 89.1%, 72.3%, 96.1% respectively. Specificity, positive predictive value, negative predictive value were 77.9%, 94.0%, 100.0%, 100.0%, and 92.3%, respectively. Sensitivity, accuracy, specificity, positive predictive value and negative predictive value of TIC were 83.0%, 93.5%, 96.8%, 89.3%, 94.7% respectively. OSNA was significantly more sensitive than FS (88.7% vs. 77.9%, P = 0.0095), and had no statistical difference compared with TIC (88.7% vs. 83.0%, P = 0.1254). (2) The sensitivity, accuracy, specificity, positive predictive value and negative predictive value of OSNA were 85.4%, 91.4%, 92.6%, 68.4%, 97.1% respectively. The sensitivity, accuracy, specificity, positive predictive value and negative predictive value of FS were 85.4%, 91.4%, 92.6%, 68.4% and 97.1% respectively. The sensitivity, accuracy, specificity, positive predictive value and negative predictive value of TIC were 81.0%, 94.4%, 96.9%, 83.2% and 96.5%, respectively. OSNA was significantly superior to FS (85.4% vs. 68.6%, P = 0.0021), and had no statistical difference compared with TIC (85.4% vs. 81.0%, P = 0.82). (3) OSNA and FS, TIC were more sensitive to macrometastasis and microrotation. The sensitivity of OSNA was 93.1%, FS 88.5%, TIC 92.2%, OSNA was similar to FS and TIC (P = 0.1907, P = 0.8543), and there was no significant difference between them. Sensitivity was 40.4%. Sensitivity of TIC was 44.0%. OSNA was significantly superior to FS and TIC (P = 0.0121, P = 0.0139). Intraoperative rapid prediction model of NSLN metastasis risk (1) Establishment of a molecular diagnosis-based contour map model for predicting NSLN metastasis in 103 patients in the modeling group by univariate analysis found that primary tumor size (P = 0.001), vascular vessels. Infiltration (P = 0.007), total Tumor Load (TTL) (P = 0.000), maximum metastasis size (P = 0.000), SLN positive (P = 0.000), SLN negative (P = 0.000), and SLN positive / SLN total (P = 0.000) were all associated with NSLN metastasis. Negative number (P = 0.001) was an independent risk factor for NSLN metastasis. TTL, primary tumor size, SLN positive number and SLN negative number were used to establish a contour map model for predicting NSLN metastasis in breast cancer. Predictive validation of the model This study used 61 patients in the validation group to validate the predictive model. AUC was 0.842. (3) Comparison of predictive differences between the size of the tumor and the size of the primary tumor in the imaging evaluation. This study used validation group imaging (ultrasound, mammography, MRI) to assess tumor size. The AUC of the model was 0.838, which was not significantly different from the area under the validated ROC curve (P = 0.7406). (4) The model predicted axillary lymph node staging in patients with pN1 and < pN2. The model distinguished between pN1 and < pN2 axillary lymph nodes. The predictive risk threshold for staging was 45.4%, AUC was 0.861, P 0.0001. In addition, the predictive risk threshold for pN1 and < P N2 was 33.5%, AUC was 0.839, P 0.0001. (5) The validity of this model with other predictive models was verified. Compared with the model of Anderson Cancer Center (MDA) in USA and Tenon Hospital in France, the AUC was 0.745 and 0.623 respectively. In addition, the model was validated by Isabel in Spain and the AUC was 0.834. Conclusion: OSNA, as an objective standardized technique, can diagnose SLN quickly and accurately during operation, and the operation is simple, which greatly reduces the subjectivity and workload of pathologists, and can be used as the first diagnosis of SLN during operation and even after operation. The model based on OSNA intraoperative molecular diagnosis is superior to other predictive models in predicting NSLN metastasis. It has better guiding value for the follow-up axillary treatment, radiotherapy target delineation and systemic treatment.
【学位授予单位】：济南大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R737.9

【参考文献】