中文摘要

第一部分

超声及超声造影在导管良恶性病变中的诊断价值研究

目的

研究乳腺导管病变的的超声及超声造影特点，探索超声及超声造影在乳腺导管良恶性病变中的诊断价值。

材料和方法

收集2019年4月至2020年1月在中国医学科学院肿瘤医院就诊的可疑乳腺导管病变患者82例（82个病灶），所有患者均具有穿刺活检或术后病理结果，所有患者均于手术或穿刺前行超声及超声造影检查，记录常规超声中的大小、形态、边界、内部及后方回声、钙化以及血流信号（Adler分级）特征；记录超声造影中的血流灌注方式（wash in/out）、增强强度、增强模式、增强范围、灌注缺损、边缘高增强，绕穿血管、边界、到达时间(AT，arrive time)，上升斜率k，达峰时间(TTP，time to peak)，峰值强度(PI，peak intensity)，曲线下面积(AUC，area under curve)。以穿刺活检或术后病理结果为金标准，将病变分为良性组与恶性组，对比分析良恶性组间常规超声特点及超声造影特点的差异，筛选出具有统计学差异的特征作为协变量，使用多因素Logistic回归分析识别出预测恶性导管病变的具有独立危险因素的特征变量，并对该特征变量进行受试者工作特征曲线分析，检验其对乳腺导管内良性病变的鉴别诊断价值。

结果

1、本研究共纳入82例可疑导管病变患者。其中良性病变44个，包括导管内乳头状瘤16个，纤维腺瘤12个，腺病10个，导管内沉积物3个，导管上皮增生2个，脂肪坏死1个；恶性病变38个，包括导管原位癌18个，导管内乳头状癌1个，浸润性导管癌16个，粘液癌1个，浸润性小叶癌2个。

2、常规超声：良、恶性组在大小、形态、边界、内部回声、微钙化、血流信号方面有显著性差异（P<0.05）；在后方回声中良恶性组没有统计学差异（P<0.05）。

3、超声造影：4个病灶（3个导管沉积物和1个粘液癌）无血流灌注，其余78个病灶（良性病变41个，恶性病变37个）在wash in 模式、增强强度、增强模式、增强范围、灌注缺损、边缘高增强、边界上差异有统计学意义（P<0.05）；定量分析结果表明，恶性组和良性组之间的AT和AUC有显著差异（P<0.05）。良性组平均AT值为11.3±7.01s，恶性组为8.27±2.25s，AUC良性组为477.66±364.46，恶性组为760.38±478.79。

4、二元多因素Logistic回归分析表明微钙化（OR=8.96，P<0.05）和增强范围扩大（OR=27.42，P<0.05）为恶性导管病变的独立危险因素。

5、微钙化诊断恶性导管内病变的敏感性、特异性、阳性预测值、阴性预测值、准确率和受试者操作特征曲线下面积分别为0.421、0.955、0.889、0.656、0.707、0.690；增强范围扩大分别为0.784、0.927、0.901、0.826、0.859、0.860；钙化联合增强范围扩大分别为0.895、0.886、0.872、0.907、0.890、0.920。

结论

1、乳腺良恶性导管病变常规超声及超声造影特点不尽相同。恶性病变主要表现为形态不规则，边界不清，导管内微钙化，以1-2级血流为主的低回声实性肿物；超声造影表现表现为增强范围扩大的“快进快出或快进慢出”的不均匀高增强。

2、微钙化和增强范围扩大是预测恶性导管病变的独立危险因素。

3、超声及超声造影在乳腺导管病变的诊断中具有重要价值：微钙化与增强范围扩大诊断乳腺导管良恶性病变的敏感性、特异性、阳性预测值、阴性预测值、准确率均较高。二者联合诊断可大大提高诊断性能。

第二部分

超声造影在乳腺导管病变BI-RADS分类中的再分类研究

目的

探讨超声造影在常规超声基础上对病变再次分类的价值。

材料和方法

收集2020年8月至2021年3月间，在中国医学科学院肿瘤医院就诊的可疑乳腺导管病变患者55例（60个病灶），所有患者均具有穿刺活检或术后病理结果，所有患者均于手术或穿刺前行超声及超声造影检查。所有患者均获得口头和书面的知情同意。记录常规超声中的大小、形态、边界、内部及后方回声、钙化以及血流信号（Adler分级）特征；记录超声造影中的血流灌注方式（wash in/out）、增强强度、增强模式、增强范围、灌注缺损、边缘高增强，绕穿血管、边界、到达时间(AT，arrive time)，上升斜率k，达峰时间(TTP，time to peak)，峰值强度(PI，peak intensity)，曲线下面积(AUC，area under curve)。由两名具有5年以上超声及超声造影诊断工作经验的医师根据美国放射学会（America College of Radiology，ACR）2013 版乳腺超声影像报告及数据系统（Breast Imaging Reporting and Data System，BI-RADS）分类标准对病变进行分类。分类意见不一致时，则互相讨论并达成一致意见。在超声造影方面，根据本研究第一部分结果，对病变进行再分类，如果病变在造影过程中出现增强范围扩大则在常规超声分类基础之上升1级；若出现无增强或低增强，则降1级；其余情况则保持分类不变。

结果

1、本研究共纳入55例（60个病灶）可疑导管病变患者。其中良性病变37个，包括腺病10个，导管内乳头状瘤9个，纤维腺瘤8个，导管内沉积物3个，导管上皮普通增生3个，导管上皮不典型增生2个，大汗腺腺瘤1个，乳腺内淋巴结1个；恶性病变23个，浸润性导管癌15个，包括导管原位癌5个，导管内乳头状癌2个，粘液癌1个。

2、在常规超声分类中，4a类病变中恶性百分比为 11.76%（4/34）；4b类病变中，恶性百分比为 57.1%（8/14）；4c类病变中，恶性百分比为100%（8/8）；均高于恶性理论风险值范围，4个5类病变，恶性百分比为75%（3/4），低于恶性理论风险值范围（＞95%）。

超声造影再次分类后，3 类病灶恶性百分比为5.0%（1/20），略高于理论风险值（理论风险值＜2%）；4a类病灶中，恶性百分比为 6.7%（1/15）；4b类病灶中，恶性百分比为 66.7%（2/3）；4c类病灶中，恶性百分比为80.0%（8/10）；除4b类高于理论风险值外，其余均在理论风险值范围内；5类病灶中，恶性百分比为 91.67%（11/12），略低于理论风险值（理伦风险值＞95%）。超声与超声造影之间各分类的对比中，3、4a类的分级中，超声与超声造影再分类差异有统计学意义（P<0.05）。

3、以BI-RADS 4b类作为判断良恶性病灶的分界点时，常规超声准确判断常规超声诊断恶性导管病变的敏感性 82.6%，特异性 81.1%，阳性预测73.1%，阴性预测值 88.2%，准确率 81.7%；超声造影再次分类诊断乳腺导管恶性病变的敏感性 91.3%，特异性89.2%，阳性预测值 84.0%，阴性预测值 94.2%，准确率 90.0%。超声与超声造影之间的敏感性、特异性、阳性预测值、阴性预测值、准确率的对比均无统计学意义（P>0.05）。以BI-RADS 4类作为判断良恶性病灶的分界点时，常规超声诊断恶性导管病变的敏感性100%，特异性 0%，阳性预测值 38.3%，准确率 38.3%。超声造影再次分级的敏感性 95.6%，特异性51.4%，阳性预测值 55.0%，阴性预测值 95.0%，准确率 68.3%。超声与超声造影之间的特异性、准确率的对比差异有统计学意义（P<0.05）。

4、造影前常规超声分类后，26个良性病变病灶中（除外导管内乳头状瘤与不典型导管上皮增生），26个（被分类为4及5类）病变推荐穿刺活检，穿刺活检率为100%（26/26）。

而超声造影分类后，26个良性病变中（除外导管内乳头状瘤与不典型导管上皮增生），超声造影再次分类后仅8个（被分类为4及5类）病变推荐穿刺活检，穿刺活检率仅为30.8%（8/26）；相对于常规超声，超声造影可明显降低良性病变不必要的穿刺活检，差异有统计学意义（P<0.05）。

结论

1、超声造影在常规超声基础上再分类能够有效提高BI-RADS分类准确性，使其恶性百分比更加接近于其理论风险值。

2、超声造影在常规超声基础上能有效提高良恶性病变诊断准确率。主要体现在对常规超声4a类病变中的再分类。

3、超声造影能够有效降低良性病灶的穿刺活检率。

第三部分

超声造影与X线、MRI在乳腺导管病变诊断中的对比研究

目的

探讨超声造影与乳腺X线摄影、MRI在乳腺导管病变诊断中的差异

材料和方法

收集2020年8月至2021年3月在中国医学科学院肿瘤医院就诊的可疑乳腺导管病变患者55例（64个病灶），所有患者均具有穿刺活检或术后病理结果，所有患者均于手术或穿刺前行超声及超声造影检查。所有患者均获得口头和书面的知情同意。记录常规超声中的大小、形态、边界、内部及后方回声、钙化以及血流信号（Adler分级）特征；记录超声造影中的血流灌注方式（wash in/out）、增强强度、增强模式、增强范围、灌注缺损、边缘高增强，绕穿血管、边界、到达时间(AT，arrive time)，上升斜率k，达峰时间(TTP，time to peak)，峰值强度(PI，peak intensity)，曲线下面积(AUC，area under curve)。超声及超声造影分类方法同第二部分研究。由两位具有5年及以上工作经验的放射科医师分析患者的乳腺X线及MRI图像，根据ACR 2013年乳腺X线摄影BI-RADS标准进行阅片和分类。分类意见不一致时，则互相讨论并达成一致意见；同时根据ACR 2013版 MRI BI-RADS分类标准进行分类，分类意见不一致时，则互相讨论并达成一致意见。

结果

1、本研究共纳入55例（64个病灶）可疑导管病变患者。其中良性病变40个，包括腺病12个，导管内乳头状瘤9个，纤维腺瘤8个，导管内沉积物3个，导管上皮普通增生3个，导管上皮不典型增生3个，大汗腺腺瘤1个，乳腺内淋巴结1个；恶性病变24个，浸润性导管癌15例，包括导管原位癌6个，导管内乳头状癌2个，粘液癌1个。

2、超声及超声造影检出60个病灶，检出率为93.8%（60/64）；MRI检出病灶63个病灶，检出率为98.4%（63/64）；乳腺X线摄影检查的患者检出病灶31个，检出率为65.9%（31/47）。乳腺X线摄影检出率低于超声和MRI，差异有统计学意义（P<0.05）。超声与MRI在检出率之间差异无统计学意义（P>0.05）。超声可以发现乳腺X线中64.3%（9/14）的钙化灶。

3、超声及超声造影中，3类病灶中，恶性百分比为5.0%（1/20），略高于理论风险值（理论风险值＜2%）；4类病灶中，良性百分比为 60.7%（17/28）；恶性百分比为 39.3%（11/28）；5级病灶中，恶性百分比为 91.67%（11/12），略低于理论风险值（理伦风险值＞95%）。

MRI分类中，3 类病灶中，恶性百分比为8.3%（1/12），略高于理论风险值（理论风险值＜2%）；4类病灶中，良性百分比为 71.0%（27/38）；恶性百分比为 28.9%（11/38）；5级病灶中，恶性百分比为 92.3%（12/13），略低于理论风险值（理伦风险值＞95%）。

对乳腺X线分类，3 类病灶中，恶性百分比为50.0%（3/6），高于理论风险值（理论风险值＜2%）；4类病灶中，良性百分比为 57.9%（11/19）；恶性百分比为 42.1%（8/19）；5级病灶中，恶性百分比为 84.3%（5/6），低于理论风险值（理伦风险值＞95%）。

对3类病变的分类中，超声优于乳腺X线，在4类的分类中超声优于MRI，差异有统计学意义。其余差异无统计学意义（P>0.05）。

4、超声造影诊断恶性导管病变的敏感性91.6%，特异性51.4%，阳性预测值 55.0%，阴性预测值95.0%，准确率68.3%；

MRI诊断恶性导管病变的敏感性95.8%，特异性28.2%，阳性预测值45.1%，阴性预测值91.7%，准确率53.9%；

乳腺X线摄影恶性导管病变的敏感性81.3%，特异性20.0%，阳性预测值52.0%，阴性预测值50.0%，准确率51.6%。

     超声及超声造影特异性高于乳腺X线摄影与MRI，差异有统计学意义（P<0.05）。

5、MRI分类后，28个良性病变病灶中（除外导管内乳头状瘤与不典型导管上皮增生），17个（被分类为4及5类）病变推荐穿刺活检，穿刺活检率为60.7%（17/28）。

而超声造影分类后，26个良性病变中（除外导管内乳头状瘤与不典型导管上皮增生），超声造影再次分类后仅8个（被分类为4及5类）病变推荐穿刺活检，穿刺活检率仅为30.8%（8/26）；超声造影可明显降低良性病变不必要的穿刺活检，差异有统计学意义（P<0.05）。

结论

1、超声及超声造影能够检出绝大部分乳腺导管病变，其检出率与MRI基本相当，高于乳腺X线。

2、超声及超声造影对BI-RADS4类病变中的分类更加准确，其恶性百分比更高。

3、超声及超声造影在良恶性病变鉴别中的特异性高于MRI与乳腺X线，能更好的识别出良性病变。

4、相对于MRI，超声及超声造影能够减少良性病变的穿刺活检率，减少不必要的穿刺活检，同时并不会减少恶性病变的穿刺活检率。

Abstract

Part I

Diagnostic value of ultrasound and contrast-enhanced ultrasound in benign and malignant ductal lesions

Objective

To study the characteristics of conventional ultrasound and CEUS in breast ductal lesions, and to explore the diagnostic value of conventional ultrasound and CEUS in benign and malignant breast ductal lesions.

Materials and methods

82 patients (82 lesions) with suspected breast ductal lesions in Cancer Hospital of Chinese Academy of Medical Sciences from April 2019 to January 2020 were collected. All patients had biopsy or postoperative pathological results. All patients were examined by ultrasound and contrast-enhanced ultrasound before operation or puncture. The size, shape, boundary, internal and posterior echogenicity, echo intensity and echo intensity of conventional ultrasound were recorded Calcification and blood flow signal (Adler grade) characteristics; The flow perfusion pattern (wash in/out), enhancement intensity, enhancement mode, enhancement range, perfusion defect, edge high enhancement, pass through vessel, boundary, arrival time (at, arrival time), rise slope k, time to peak (TTP, time to peak), peak intensity (PI, peak intensity), area under curve (AUC, area under curve) were recorded. According to the gold standard of biopsy or postoperative pathological results, the lesions were divided into benign group and malignant group. The differences of conventional ultrasound characteristics and contrast-enhanced ultrasound characteristics between benign and malignant groups were compared and analyzed. The characteristics with statistical differences were selected as covariates. The characteristic variables with independent risk factors were identified by multivariate logistic regression analysis. The characteristic variable was analyzed by receiver operating characteristic curve (ROC) to test its value in the differential diagnosis of breast ductal benign lesions.

Result

1. A total of 82 patients with suspected ductal lesions were included in this study. There were 44 benign lesions, including 16 intraductal papilloma, 12 fibroadenoma, 10 adenosis, 3 intraductal deposits, 2 ductal epithelial hyperplasia and 1 fat necrosis; There were 38 malignant lesions, including 18 ductal carcinoma in situ, 1 intraductal papillary carcinoma, 16 invasive ductal carcinoma, 1 mucinous carcinoma and 2 invasive lobular carcinoma.

2. Conventional ultrasound: there were significant differences in size, shape, boundary, internal echo, microcalcification and blood flow signal between benign and malignant groups (P<0.05); In the posterior echo, there was no significant difference between benign and malignant groups (P<0.05).

3. CEUS: 4 lesions (3 ductal deposits and 1 mucinous carcinoma) had no blood perfusion, and the remaining 78 lesions (41 benign lesions and 37 malignant lesions) showed significant differences in wash in mode, enhancement intensity, enhancement mode, enhancement range, perfusion defect, high edge enhancement and boundary (P<0.05); The results of quantitative analysis showed that there were significant differences in AT and AUC between malignant group and benign group (P<0.05). The mean AT value of benign group was 11.3±7.01s and 8.27s±2.25s in malignant group; The mean AUC value of benign group 477.66 ±364.46s , 760.38±478.79 in malignant group.

4. Binary multivariate logistic regression analysis showed that microcalcification (OR=8.96, P<0.05) and enhancement range enlargement (OR=27.42, P<0.05) were independent risk factors for malignant ductal lesions.

5. The sensitivity, specificity, positive predictive value, negative predictive value, accuracy and area under the receiver operating characteristic curve were 0.421, 0.955, 0.889, 0.656, 0.707 and 0.690, respectively; The enhancement range was 0.784, 0.927, 0.901, 0.826, 0.859 and 0.860, respectively; 0.895, 0.886, 0.872, 0.907, 0.890, 0.920 respectively.

Conclusion

1. The characteristics of conventional ultrasound and CEUS are different between benign and malignant breast ductal lesions. The main manifestations of malignant lesions were irregular shape, unclear boundary, microcalcification in the catheter, hypoechoic solid tumor with grade 1-2 blood flow; CEUS showed heterogeneous high enhancement of "fast in, fast out or fast in, slow out" with enlarged enhancement range.

2. Microcalcification and enhancement range enlargement are independent risk factors for predicting malignant ductal lesions.

3. Conventional ultrasound and CEUS have important value in the diagnosis of breast ductal lesions: the sensitivity, specificity, positive predictive value, negative predictive value and accuracy of microcalcification and enhanced range in the diagnosis of benign and malignant breast ductal lesions are high. The combined diagnosis can greatly improve the diagnostic performance.

Key words: Breast, Ductal lesions, Ultrasound, Contrast-enhanced ultrasound

Part II

Re classification of contrast-enhanced ultrasound in BI-RADS of breast ductal lesions

Objective

To explore the value of contrast-enhanced ultrasound (CEUS) in the classification of lesions on the basis of conventional ultrasound.

Materials and methods

55 patients (60 lesions) with suspected breast ductal lesions in Cancer Hospital of Chinese Academy of Medical Sciences from August 2020 to March 2021 were collected. All patients had biopsy or postoperative pathological results. All patients underwent ultrasound and contrast-enhanced ultrasound before operation or puncture. All patients obtained informed oral and written consent. The size, shape, boundary, internal and posterior echo, calcification and blood flow signal (Adler grade) were recorded; The flow perfusion pattern (wash in / out), enhancement intensity, enhancement mode, enhancement range, perfusion defect, edge high enhancement, pass through vessel, boundary, arrival time (at, arrival time), rise slope k, time to peak (TTP, time to peak), peak intensity (PI, peak intensity), area under curve (AUC, area under curve) were recorded. Two doctors with more than 5 years of experience in ultrasound and contrast-enhanced ultrasound diagnosis classified the lesions according to the American College of Radiology (ACR) 2013 breast imaging reporting and data system (BI-RADS) classification standard. When there are different opinions on classification, they will discuss with each other and reach an agreement. In contrast-enhanced ultrasound, according to the results of the first part of this study, the lesions were reclassified. If the enhancement range of the lesions was enlarged during contrast-enhanced ultrasound, it would be increased by one grade on the basis of conventional ultrasound classification; If there is no enhancement or low enhancement, it will be reduced by one grade; In other cases, the classification remains unchanged.

Result

1. 55 patients (60 lesions) with suspected ductal lesions were included in this study. There were 37 benign lesions, including 10 adenoses, 9 intraductal papillomas, 8 fibroadenomas, 3 intraductal deposits, 3 ductal epithelial hyperplasia, 2 ductal epithelial atypical hyperplasia, 1 apocrine adenoma and 1 breast lymph node; There were 23 malignant lesions, 15 invasive ductal carcinoma, including 5 ductal carcinoma in situ, 2 intraductal papillary carcinoma and 1 mucinous carcinoma.

2. In the conventional ultrasound classification, the malignant percentage of 4A lesions was 11.76% (4/34); The percentage of malignant lesions was 57.1% (8/14); Among 4C lesions, the malignant rate was 100% (8/8); The percentage of malignant lesions was 75% (3/4), which was lower than the range of malignant theoretical risk value (> 95%).

After re classification of CEUS, the malignant percentage of category 3 was 5.0% (1 / 20), slightly higher than the theoretical risk value (theoretical risk value < 2%); The malignant rate was 6.7% (1/15); 7% (2/3); Among 4C lesions, 80.0% (8/10) were malignant; Except for category 4B which is higher than the theoretical risk value, the others are in the range of the theoretical risk value; Among the category 5, the malignant rate was 91.67% (11/12), which was slightly lower than the theoretical risk value (theoretical risk value > 95%). There was a significant difference in the classification between ultrasound and contrast-enhanced ultrasound (P<0.05).

3. The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of BI-RADS 4b were 82.6%, 81.1%, 73.1%, 88.2% and 81.7%, respectively; The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of CEUS were 91.3%, 89.2%, 84.0%, 94.2% and 90.0%, respectively. There was no significant difference in sensitivity, specificity, positive predictive value, negative predictive value and accuracy between ultrasound and contrast-enhanced ultrasound (P>0.05). The sensitivity, specificity, positive predictive value and accuracy of conventional ultrasound in the diagnosis of malignant ductal lesions were 100%, 0%, 38.3% and 38.3%, respectively. The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of CEUS grading were 95.6%, 51.4%, 55.0%, 95.0% and 68.3%, respectively. The difference of specificity and accuracy between ultrasound and contrast-enhanced ultrasound was statistically significant (P<0.05).

4. Conventional ultrasound classification before CEUS, 26 benign lesions (except intraductal papilloma and atypical ductal epithelial hyperplasia) were classified into 4 and 5 categories. The biopsy rate was 100% (26 / 26).

Among 26 benign lesions (except intraductal papilloma and atypical ductal epithelial hyperplasia) after CEUS classification, only 8 lesions (classified as 4 and 5) after CEUS classification were recommended for biopsy, and the biopsy rate was only 30.8% (8 / 26); Compared with conventional ultrasound, CEUS can significantly reduce the unnecessary biopsy of benign lesions, the difference was statistically significant (P<0.05).

Conclusion

1. Re classification of contrast-enhanced ultrasound based on conventional ultrasound can effectively improve the accuracy of BI-RADS classification and make the malignant percentage closer to its theoretical risk value.

2. CEUS can effectively improve the diagnostic accuracy of benign and malignant lesions on the basis of conventional ultrasound. It is mainly reflected in the classification of 4A lesions by conventional ultrasound.

3. CEUS can effectively reduce the biopsy rate of benign lesions.

Part III

Comparative study of contrast-enhanced ultrasound, mammography and MRI in the diagnosis of breast ductal lesions

Objective

To explore the difference of contrast-enhanced ultrasound, mammography and MRI in the diagnosis of breast ductal lesions.

Materials and methods

55patients (64 lesions) with suspected breast ductal lesions in Cancer Hospital of Chinese Academy of Medical Sciences from August 2020 to March 2021 were collected. All patients had biopsy or postoperative pathological results. All patients underwent ultrasound and contrast-enhanced ultrasound before operation or puncture. All patients obtained informed oral and written consent. The size, shape, boundary, internal and posterior echo, calcification and blood flow signal (Adler grade) were recorded; The flow perfusion pattern (wash in / out), enhancement intensity, enhancement mode, enhancement range, perfusion defect, edge high enhancement, pass through vessel, boundary, arrival time (at, arrival time), rise slope k, time to peak (TTP, time to peak), peak intensity (PI, peak intensity), area under curve (AUC, area under curve) were recorded. The classification of ultrasound and contrast-enhanced ultrasound is the same as the second part. Two radiologists with more than 5 years' working experience analyzed the mammography and MRI images of the patients, and read and classified them according to the ACR 2013 BI-RADS standard. When there are different opinions on classification, they will discuss with each other and reach an agreement; At the same time, according to ACR 2013 MRI BI-RADS classification standard, when the classification opinions are different, they will discuss with each other and reach an agreement.

Result

1. A total of 55 patients (64 lesions) with suspected ductal lesions were included in this study. There were 40 benign lesions, including 12 adenoses, 9 intraductal papillomas, 8 fibroadenomas, 3 intraductal deposits, 3 ductal epithelial dysplasia, 3 ductal epithelial dysplasia, 1 apocrine adenoma and 1 breast lymph node; There were 24 malignant lesions and 15 invasive ductal carcinoma, including 6 ductal carcinoma in situ, 2 intraductal papillary carcinoma and 1 mucinous carcinoma.

2. 60 lesions were detected by ultrasound and contrast-enhanced ultrasound, and the detection rate was 93.8% (60/64); MRI detected 63 lesions, the detection rate was 98.4% (63/64); 31 lesions were detected by mammography, and the detection rate was 65.9% (31/47). The detection rate of mammography was lower than that of ultrasound and MRI (P < 0.05). There was no significant difference in the detection rate between ultrasound and MRI (P>0.05). Ultrasound could detect 64.3% (9/14) calcification in mammography.

3. In ultrasound and contrast-enhanced ultrasound, the percentage of malignant lesions was 5.0% (1/20) in category 3, slightly higher than the theoretical risk value (theoretical risk value < 2%); Among category 4, the percentage of benign lesions was 60.7% (17/28); 3% (11/28); In category 5, the malignant rate was 91.67% (11/12), slightly lower than the theoretical risk value (theoretical risk value > 95%).

3% (1/12), slightly higher than the theoretical risk value (theoretical risk value < 2%); Among the 4 types of lesions, 71.0% (27/38) were benign; 9% (11/38); In Grade 5 lesions, the malignant rate was 92.3% (12/13), which was slightly lower than the theoretical risk value (theoretical risk value > 95%).

According to the classification of mammography, the percentage of malignant lesions was 50.0% (3/6), which was higher than the theoretical risk value (theoretical risk value < 2%); Among the 4 types of lesions, the percentage of benign lesions was 57.9% (11/19); The malignant rate was 42.1% (8/19); In Grade 5 lesions, the malignant rate was 84.3% (5/6), which was lower than the theoretical risk value (theoretical risk value > 95%).

In category 3, ultrasound is better than mammography, in the classification of four kinds of lesions, ultrasound is better than MRI, the difference is statistically significant. The other differences were not statistically significant (P > 0.05).

4. The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of CEUS in the diagnosis of malignant ductal lesions were 91.6%, 51.4%, 55.0%, 95.0% and 68.3%, respectively;

The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of MRI in the diagnosis of malignant ductal lesions were 95.8%, 28.2%, 45.1%, 91.7% and 53.9%, respectively;

The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of mammography for malignant ductal lesions were 81.3%, 20.0%, 52.0%, 50.0% and 51.6%, respectively.

The specificity of ultrasound and contrast-enhanced ultrasound was higher than that of mammography and MRI (P < 0.05).

5. After MRI Classification, in 28 benign lesions (except intraductal papilloma and atypical ductal epithelial hyperplasia), 17 lesions (classified as 4 and 5) were recommended for biopsy, and the biopsy rate was 60.7% (17 / 28).

Among 26 benign lesions (except intraductal papilloma and atypical ductal epithelial hyperplasia) after CEUS classification, only 8 lesions (classified as 4 and 5) after CEUS classification were recommended for biopsy, and the biopsy rate was only 30.8% (8 / 26); Contrast enhanced ultrasound can significantly reduce the unnecessary biopsy of benign lesions, the difference was statistically significant (P < 0.05).

Conclusion

1. Conventional ultrasound and CEUS can detect the vast majority of breast ductal lesions, the detection rate is almost the same as MRI, higher than mammography.

2. The classification of category 4 lesions by conventional ultrasound and CEUS is more accurate, and the malignant percentage is higher.

3. The specificity of conventional ultrasound and CEUS in the differential diagnosis of benign and malignant lesions is higher than that of MRI and mammography, which can better identify benign lesions.

4. Compared with MRI, conventional ultrasound and CEUS can reduce the biopsy rate of benign lesions, reduce unnecessary biopsy, and will not reduce the biopsy rate of malignant lesions.