【作者】 谭红娜；

【导师】 彭卫军；

【作者基本信息】 复旦大学 ， 影像医学与核医学， 2011， 博士

【摘要】 第一部分VX2兔乳腺癌动物模型的建立及其在乳腺癌前哨淋巴结显像中的应用研究目的：建立VX2兔乳腺癌动物模型,并评价其在乳腺癌前哨淋巴结(Sentinel Lymph Node, SLN)显像中应用的可行性。材料和方法：采用VX2肿瘤软组织块悬液注射法,在30只纯种雌性新西兰大白兔左侧胸壁第三乳头乳垫下注入1ml悬液,建立原位乳腺癌动物模型；在VX2肿瘤种植后2-3周行影像学检查。实验大白兔随机编号,分成奇数组和偶数组。所有成瘤动物均行MR-LG检查,而奇数组实验大白兔先后行CT-LG和99Tcm标记的同位素检查。在三维(3D)重建后的CT或MR图像上,白注射部位指向腋窝方向的引流淋巴管通路上最先显像的一个或数个淋巴结定义为SLN,并按照有无SLN或/和引流淋巴管显像将成像效果分为良好(SLN和淋巴管均显像)和较差(SLN和/或淋巴管不显像)两个等级。结果：(1)新西兰大白兔原位乳腺癌成瘤率为100%,其中53.3%(16/30)的肿瘤发生囊变坏死,36.7%(11/30)的肿瘤表面皮肤发生破溃。(2)80%(24/30)的荷瘤兔完成MR-LG检查,且在MR-LG检查时62.5%(15/24)的荷瘤兔腋窝可扪及淋巴结肿大。(3)在奇数组实验动物中,完成CT-LG和99Tcm同位素检查的荷瘤兔的百分比分别为73.3%(11/15)和66.7%(10/15)。(4)3D重建后图像的等级为良好和较差的百分比在MR-LG和CT-LG上分别为79.2%(19/24)、20.8%(5/24)和81.8%(9/11)、18.2%(2/11)。结论：VX2兔乳腺癌动物模型制作简单,成瘤率高,对实验操作耐受性好；VX2兔乳腺癌模型可较好地应用于乳腺癌SLN显像的影像学成像研究。第二部分VX2兔乳腺癌前哨淋巴结MR-LG显像的动物实验研究目的：探讨马根维显(Magnevist)对比剂增强的MR淋巴管造影(Magnetic Resonance Lymphpgraphy, MR-LG)检查用于VX2兔乳腺癌SLN显像的可行性,并探索较合适的成像条件。材料和方法：30只新西兰大白兔在VX2肿瘤接种之前先行基线MR-LG检查,在肿瘤生长至2-3周时行第二次MR-LG检查。皮下注射对比剂并快速按摩前后分别行短反转时间的反转恢复(Short TI Inversion Recovery,STIR)序列扫描和T1加权3D容积扫描(包括轴位和冠状位),并计算对比剂增强后图像上淋巴结及其引流淋巴管和同层肌肉的SNR值(SNR=绝对信号强度(SI)/背景噪声标准差(SD))。记录STIR序列扫描和MR-LG显示的所有淋巴结,并与前哨淋巴结活检(Sentinel Lymph Node Biopsy)结果对照。结果：(1)30只新西兰大白兔均完成第一次MR-LG检查,80%(24/30)的荷瘤兔完成第二次MR-LG检查；图像3D重建后成像等级为良好和较差的百分比分别为79.2%(19/24)和20.8%(5/24)。(2)对比剂皮下注射并快速按摩后,所有扫描序列图像上淋巴结(包括SLN和n-SLN)SNR值均高于其引流淋巴管和同层肌肉(P值均<0.05),但在SLN和n-SLN组中淋巴结、其引流淋巴管及同层肌肉的SNR值均无统计学差异；且轴状位扫描图像较冠状位扫描清晰(P值均<0.05)。(3)24只荷瘤兔STIR序列扫描图像上均可见大小不一的淋巴结影；MR-LG重建图像上均可清晰显示SLN和n-SLN(共26枚)的17只荷瘤兔,其STIR图像上共观察到28枚淋巴结,且两者观察到的淋巴结大小无明显差异(t=0.124,P=0.902)。(4)SLNB手术成功率可达95.5%(21/22)；17只MR-LG成像效果良好的荷瘤兔均顺利完成SLNB检查,且获得SLN18枚,n-SLN14枚。SLNB获得的淋巴结个数及大小与MR-LG检查和STIR序列扫描显示的淋巴结无明显统计学意义。结论：马根维显对比剂增强的MR-LG检查可较好用于VX2兔乳腺癌SLN显像,且轴状位扫描图像较冠状位扫描图像清晰。第三部分VX2兔乳腺癌前哨淋巴结CT-LG显像的动物实验研究目的：探讨碘帕醇(Iopamidol)对比剂增强的CT淋巴管造影(Computed Tomography Lymphpgraphy, CT-LG)检查用于VX2兔乳腺癌SLN显像的可行性,并探索较合适的成像条件。材料和方法：11只存活奇数组荷瘤兔在MR-LG检查结束24小时后行CT-LG检查,并于对比剂皮下注射并按摩后即刻、5min和10min时行薄层CT扫描。按照淋巴管及SLN有无显示将CT-LG成像效果分成良好(SLN及淋巴管均显示)和较差(SLN和/或淋巴管未显像)两个等级。记录CT-LG检查显示的所有淋巴结,并与前哨淋巴结活检(Sentinel Lymph Node Biopsy)结果对照。结果：(1)11只奇数组荷瘤兔CT扫描图像容积重建后,成像效果为良好和较差的百分比分别为81.8%(9/11)和18.2%(2/11)。(2)在9只成像效果为良好的荷瘤兔的CT-LG图像上,对比剂注射并快速按摩后即刻扫描可清楚显示20枚淋巴结影,而5min和10min扫描时淋巴结的显示率分别为90%和65%；淋巴结的平均CT值均高于引流淋巴管及同层肌肉,且随着时间的延长CT值逐渐下降,尤其是在淋巴结及其引流淋巴管CT值下降明显,二者均有统计学意义(P=0.002和P=0.017)。(3)对比剂皮下注射并快速按摩后即刻扫描CT-LG显像共显示20枚淋巴结,其中SLN和n-SLN各10枚。椭圆形在SLN中多见,但二者无统计学差异(P=0.168)；SLN虽然比n-SLN稍大,但二者并无统计学意义(P=0.164和0.210)。(4)CT-LG结束存活的10只荷瘤兔均顺利完成SLNB手术(活检成功率为100%),活检共获得13枚SLN和9枚n-SLN,其中19枚(86.4%,19/22)为蓝染淋巴结；CT-LG获得淋巴结较SLNB获得淋巴结稍小,且二者有统计学意义(P=0.021)。结论：CT-LG检查可较好地用于VX2兔乳腺癌SLN的显像,且往往在对比剂皮下注射并快速按摩后即刻扫描就可清楚显示SLN及其引流的淋巴管。第四部分淋巴管造影方法评价转移性和炎症性SLN的动物实验研究目的：建立兔乳腺肿瘤及炎性病变动物模型,并探讨CT及MR淋巴管造影检查定位和评价转移性及炎症性SLN的可行性。材料和方法：采用VX2肿瘤组织块悬浊液和新鲜鸡蛋黄乳胶皮下注射法分别制作VX2兔乳腺癌(30只)和炎症性兔动物模型(10只)；且肿瘤组实验动物随机编号,分为奇数组和偶数组。所有存活实验兔均行MR-LG检查(对比剂皮下注射前行STIR序列扫描),且肿瘤组中奇数组实验动物和炎症组实验动物同时行CT-LG检查。在MR/CT-LG图像上SLN是指自对比剂注射部位至腋窝方向上最先引流的一个或数个淋巴结,其余淋巴结为n-SLN,并与STIR序列扫描图像上自肿瘤至腋窝方向上的第一个淋巴结和SLN活检结果对照。结果：(1)在CT-LG检查中,荷瘤组SLN及n-SLN的长短径均比炎症组的大,但二者均无统计学差异,P值依次为0.154、0.283、0.346和0.629；淋巴结边缘出现明显充盈缺损在转移性SLN(100%,2/2)中较炎症性SLN(10.5%,2/19)多见。(2)在MR-LG检查中,荷瘤组和炎症组中SLN的长短径及形状均无统计学差异(P=0.311,0.914和0.677)；病理学证实为转移的SLN的长径大于炎症性SLN的,且边缘环状强化多出现在炎症性SLN中(2/10 vs 13/18),二者在区分转移性及炎症性SLN中均有统计学意义(P=0.046和0.016)。(3)在荷瘤组和炎症组实验动物中,MR-LG显示的SLN与STIR序列扫描显示的自肿瘤至腋窝方向上的第一个淋巴结在长短径上均无统计学差异。(4)对照SLN活检手术结果,荷瘤组和炎症组实验兔的长短径均较SLNB获得的SLN小,但MR-LG所获得SLN的长短径与SLNB所获得SLN的长短径差别较大,且在荷瘤组中三者均有统计学差异(P=0.011和0.004)。结论：转移性及炎症性SLN在CT-LG及MR-LG图像上有一定的特点,且CT-LG检查显像SLN的效果较MR-LG检查稍好。第五部分乳腺癌前哨淋巴结CT-LG显像的临床应用研究目的初步评价CT定位SLN的准确性及CT淋巴管造影(CT-LG)显示早期乳腺癌前哨淋巴结(SLN)的可行性。材料和方法选取34例穿刺或局切手术证实且腋窝触诊为阴性的乳腺癌患者行CT-LG检查,其中19例患者同时行CT-LG引导下SLN穿刺定位检查。自注射部位指向腋窝方向的引流淋巴管上最先显像的1个或数个淋巴结为SLN,与前哨淋巴结活检(SLNB)结果相对照,数目相等者为符合,多于和少于分别为高估和低估。显像效果根据容积重建后有无SLN及淋巴管显像分为良好(Ⅰ级：SLN和淋巴管均显像)和较差(Ⅱ级：SLN和/或淋巴管无显像)；并把身高体重指数(BMI)≥25时定义为肥胖。分析显像SLN的形态学特征,并与SLN活检结果对照。结果(1)CT定位乳腺癌SLN的准确度为89.5%,且联合染料法时准确度可达100%。(2)34例患者CT-LG均见SLN显像,其中88.2%(30例)患者显像效果为I级,11.8%(4例)为II级。肥胖患者CT-LG显像效果较差,两者差异有统计学意义(P＜0.05)。(3)34例患者共显示85枚SLN和63条淋巴管。与SLNB对照,47.1%(16例)患者两种结果符合,而高估和低估者分别为20.6%(7例)和32.3%(11例),造成两种结果不一致的原因主要与肥胖因素有关,两者差异有统计学意义(P＜0.05)。(4)SLNB证实23例(78枚)阴性SLN,11例(19枚)阳性SLN,对应CT-LG共85枚SLN显像,其中阴性67枚,阳性18枚。形状为圆形在阴性和阳性SLN的比例分别为22.4%(15/67)和66.7%(12/18),二者差异有统计学意义(P<0.05)。中央区出现充盈缺损在阴性和阳性SLN中的比例分别为9.0%(6/67)和27.8(5/18),但边缘区表现为不规则充盈缺损只在33.3%(6/18)的阳性SLN中出现。4枚(3例)SLN周围伴有多发小淋巴结,组织学显示有癌细胞浸润。结论CT引导下SLN穿刺定位准确度较高,且与染料法联合准确度可达100%；CT-LG可有效显示乳腺癌的SLN,但其准确性易受患者肥胖因素的影响；SLN为圆形,边缘出现虫蚀样充盈缺损,以及伴有多发小淋巴结者均可提示癌细胞浸润。更多还原

【Abstract】 Part I Establishment of Rabbit VX2 Breast Cancer Model and Evaluation of the Feasibility on Sentinel Lymph Node MappingObjective:To establish a rabbit VX2 breast cancer model, then to evaluate the feasibility for sentinel lymph node(SLN) mapping. Materials and Methods:30 New Zealand White Rabbits were injected lml VX2 tumor tissue mass suspension in the left and third breast, and then a rabbit VX2 breast cancer model was established; the imaging examinations were arranged when tumor growth was at the second to third weeks. All the experimental rabbits were gave numbers at random, which was classified odd number group and even number group; all live rabbits with bearing tumor underwent MR-LG examinations, but only odd number group bearing tumor rabbits completed CT-LG and 99Tcm labeled isotope exams orderly. On the 3D images which were obtained by reconstruction using the raw CT or MR imagings, the first one or several lymph node(s) in the lymphatic drainage pathway was/were defined as the SLN(s). According to the presentation condition of SLN and affected lymphatic drainage pathway, the imaging effect was classified as good grade (with both SLN and lymphatic drainage pathway presented) and poor grade (either SLN or lymphatic drainage pathway absented). Results:(1)The overall rate of successful establishment of the model is 100%, and the presentation of necrosis and cyst forming, big ulcer of tumor surface were in 53.3%(16/30) and 36.7%(11/30) implanted tumors, respectively. (2) The imaging examinations were performed at the 2-3 weeks after the VX2 tumor being implanted, and 80%(24/30) bearing tumor rabbits completed MR-LG exam when axillary lymph nodes enlargement were found by palpation in 62.5%(15/24) experimental rabbits. (3) In odd number rabbits, the achievement ratio for CT-LG and 99Tcm isotope exams were 73.3%(11/15) and 66.7%(10/15), respectively. (4) The imaging grade in good, and poor were 79.2%(19/24) and 20.8%(5/24) on MR-LG, and 81.8%(9/11) and 18.2%(2/11) on CT-LG, respectively. Conclusion:This VX2 breast cancer rabbit model is easy to establish, which has high rate of living tumor and good tolerance for the operations of experiment. PartⅡAnimal Experimental Study on CT-LG for SLN Mapping of VX2 Rabbit Breast CancerObjective:To study the fesibility of the interstitial magnetic resonance lymphography(MR-LG) using Magnevist contrast agent for sentinel lymph node mapping in rabbit VX2 breast cancer, as well as to find a suitable scanning parametes. Materials and Methods:MR-LG exams were completed on 30 New Zealand White Rabbits before VX2 tumor being implanted and after the tumor growing to the 2th-3th week. Short TI Inversion Recovery (STIR) sequence scanning and T1-weighted 3-dimensional (3D) scanning, including axial and coronal positions were underwent before and after the undiluted gadopentetate dimeglumine being injected, and the signal noise rate(SNR) was cunted by the ratios of signal intensity(SI) and background noise standard deviation (SD). All lymph nodes were recorded on the imagings of STIR scanning and MR-LG, then comparing with the results of SLN biopsy. Results:(1)A11 rabbits had completed the first MR-LG examinations, and 80%(24/30) bearing tumor rabbits completed the second MR-LG examinations; The 3D imaging grade in good, and poor grade were 79.2%(19/24) and 20.8%(5/24), respectively. (2) The SNR of lymph node were higher than that of affected lymph duct and muscle after the contrast agent being injected on all scanning phases(all P value<0.05), but there were no differences between SLN and n-SLN on affected lymph duct and muscle; the axial scanning images were more clear than coronal scanning images(P value<0.05). (3).There were 26 lymph nodes on CT-LG and 28 lymph nodes on STIR scanning in all, but there was no significant difference about the size of the two(t=0.124,P=0.902).(4)The achievement ratio of SLN Biopsy was 95.5%(21/22), and all 17 bearing tumor rabbits with good grade imaging underwent the SLN Biopsy, then got 18 SLN and 14 n-SLN. However, the number and size of the SLN, which were got from SLN Biopsy, MR-LG and STIR, were no differences. Conclusion:The Magnevist contrast enhanced MR-LG is useful for the SLN mapping of VX2 rabbit breast cancer, and the axial scanning image is better than that of coronal scanning. PartⅢAnimal Experimental Study on CT-LG for SLN Mapping of VX2 Rabbit Breast CancerObjective:To study the fesibility of the interstitial computed tomography lymphography(CT-LG) using Iopamidol contrast agent for sentinel lymph node mapping in rabbit VX2 breast cancer, as well as to find a suitable scanning parametes. Materials and Methods:Eleven living odd number bearing tumor rabbits were underwent CT-LG examination after 24 hours later of MR-LG exams, and scanning were achieved at instantly,5min and 10min after the contrast agent being injected and massaged. The imaging effect was classed two grades according to the display of SLN and affected lymph duct, and grade good was defined as lymph duct and SLN displaying, while grade poorwith no lymph duct or SLN displaying. All lymph nodes on CT-LG were recorded, then comparing with the result of SLN Biopsy. Results:(1) Eleven odd number bearing tumor rabbits completed CT-LG, and the percentage of CT-LG images were 81.8%(9/11) and 18.2%(2/11), respectively. (2)On the nine good grade CT-LG,20 lymph nodes were seen clearly when scanning instantely after the contrast agent injected and massaged, but only 90% and 65% of these 20 lymph nodes displayed when scannings were set at 5min and 10min after the injection and massage; and the average CT unit of lymph nodes were higher than that of affected lymph duct and muscle, falling-off sharply with the delay of time (P=0.002and 0.017). (3)There were 20 lymph nodes in all on CT-LG images(10 SLN and n-SLN,respectively); and oval shape and bigger size than n-SLN were found frequently, which had no differences between SLN and n-SLN. (4)Ten bearing tumor rabbits with CT-LG examinations had underwent SLN Biopsy successfully, then 13 SLN and 9 n-SLN were achieved; the lymph nodes on CT-LG were smaller than that of SLN Biopsy(P=0.021). Conclusion:CT-LG is useful for the SLN mapping of VX2 rabbits breast cancer, and the instant scanning after injection and massage of Iopamidol is enough for the SLN and its affected lymph duct displaying. Part IV Experimental Study on Lymphography in Evaluating the Sentinel Lymph Nodes of Rabbit’Breast Cancer and InflammationObjective:To establish rabbit breast cancer and inflammation model, then to evaluate the feasibility for sentinel lymph node(SLN) mapping using the method of lymphography. Materials and Methods:30 VX2 rabbit breast cancer and 10 inflammate rabbit models were established by tumor tissue mass suspension and the yolk colloid injection; and all tumor group rabbits were gave numbers at random, then odd number and even number group were divided. All live experimental rabbits underwent MR-LG examinations (STIR sequence scanning were underdone before the injection of contrast agent subcutaneously), and odd number bearing tumor and all inflammate rabbits also had CT-LG examinations. On the 3D images which were obtained by reconstruction using the raw CT or MR imagings, the first one or several lymph node(s) in the lymphatic drainage pathway was/were defined as the SLN(s), and the others were n-SLN, then comparing with the first lymph node from the direction of the tumor to the axillary space, as well as to the results of SLN Biopsy. Results:(1) On CT-LG images, the long and short diameter of SLN and n-SLN in bearing tumor rabbits were bigger than that of inflammate group rabbits, and there were no differences(P value were 0.154,0.283,0.346 and 0.629, respectively); irregular filling defect on the margin was more frequent in metastatic SLN (100%,2/2) than that of in inflammate SLN(10.5%,2/19). (2)On MR-LG images, the long and short diameter of SLN, as well as the shape were no significant different between the tumor and inflammate group rabbits(P value were 0.311,0.914 and 0.677, respectively); Comparing with the SLN biopsy pathology, the long diameter of metastatic SLN was bigger than that of inflammate SLN, and ring-liking enhancement was more frequent found in inflammate SLN(2/10 vs 13/18), which were significant differences between metastatic and inflammate SLN(P=0.04 and 0.016). (3)There was no difference in the long and short diameter of SLN on MR-LG and of the first lymph node of STIR between the tumor bearing and inflammate rabbits. (4)Comparing with the results of SLN Biopsy, both the long and short diameter of SLN in bearing tumor and inflammate rabbits were smaller than that of SLN achieved by SLN Biopsy, but the difference was bigger in MR-LG than that of CT-LG(P value were 0.011 and 0.004,respectively) in tumor bearing group rabbits. Conclusion:Metastatic and inflammate SLN has some characteristics on CT-LG and MR-LG images; and SLN in CT-LG images is more true reflection SLN in SLN Biospy than that of MR-LG.Part V The Clinical Application Study on CT-LG for Sentinel Lymph Node Mapping of Breast CancerObjective:To determine the accuracy of computed tomographic (CT) localization and the feasibility of Sentinel Lymph Node (SLN) mapping with CT lymphography (CT-LG), and then directing of sentinel lymph node biopsy(SLNB) in early-stage breast cancer patients. Materials and Methods:Thirty-four patients with confirmed breast cancer underwent 40-row CT scanning, and the first one or several lymph node(s) in the lymphatic drainage pathway was/were defined as the SLN(s). To accurately localize the SLN,19 patients (55.9%) underwent the percutaneous lymph node puncture procedure.Then comparing with SLNB, if the number was equal, the evaluation was coherence, more or less were respective high and low evaluation. The quality of imaging were classedⅠandⅡby lymph duct appearance or not on the image which were reconstructed by volume. Obesity was defined when the value of body mass index (BMI) was more 25. The morphologic features of all the SLNs on CT scans were analyzed and compared with the SLN biopsy pathologic diagnosis. Results:(1)The accuracy of CT localization for SLN was 89.5%, and the accuracy increased to 100% when the CT scan technique was combined with the blue dye method. (2)All 34 patients showed SLN, and 88.2%(n=30) cases mapping class were gradeⅠ,11.8%(n=4) were gradeⅡ. The obese patients were difficult to acquire high quality mapping, and the difference was significant (P<0.05). (3)Eighty-five SLN and sixty-three lymph duct in all of 34 patients were identified on CT-LG mapping. Comparing with the results of SLNB,47.1%(n=16) patients CT-LG mapping were equal evaluation,20.6%(n=7) and 32.3%(n=11) patients had high and low evaluation, respectively. The BMI was the main reasons, which had significant differences(P< 0.05). (4) Twenty-three patients with 78 SLNs were negative, and eleven patients with 19 SLNs were positive confirmed by pathology, but on CT-LG mapping,67 negative and 18 positive SLNs were acquired. The 22.4%(15/67)negative SLN and 66.7%(12/18) positive SLN were round shape, the difference was significant(P< 0.05). The 9.0%(6/67) negative SLNs and 27.8(5/18) positive SLNs demonstrated filling defect on the center, but irregular filling defect on the margin were found only in 33.3%(6/18) of positive SLNs. However,4 SLNs (3 patients) combined with multi-small lymph nodes on CT-LG mapping were also confirmed tumor cell invasive by pathology. Conclusion:CT-LG is useful for accurate localization for SLN, and the accuracy increased to 100% when the CT scan technique was combined with the blue dye method; CT-LG can sufficiently visualize breast lymphatic drainage and may have potential utility for breast SLN mapping, but can be influenced by the obese patients. On CT-LG mapping, SLN that showed round shape, filling defect on the margin, and combined with multi-small lymph nodes around, were likely to metastatic SLN.更多还原