探讨超声光散射乳腺诊断系统(简称乳光超)对T₁期乳腺癌的诊断价值。

本研究共纳入2016年4月至2018年4月到成都市妇女儿童中心医院治疗的超声检查3级及以上(排除6级)患者共2 245例。患者同时完成了超声及乳光超检查，其中，经活组织检查且有完整病理资料的患者共1 157例，对其资料进行回顾性分析。按肿瘤最大直径将1 157例分为3组：S1组(≤1.0 cm)共452例，S2组(1.1～2.0 cm)共489例，S3组(>2.0 cm)共216例。采用受试者工作特征(ROC)曲线分析2种检查方法的诊断效能，计算ROC曲线下面积及95%CI, ROC曲线下面积的比较采用Medcalc V19.0.4软件行Z检验。2种方法敏感度及特异度的比较采用χ²检验或Fisher确切概率法；2种方法分别与病理检查结果比较，采用配对χ²检验(McNemar检验)。

1 157例中，恶性144例(12.4%)，全部为浸润性导管癌，良性1 013例(87.6%)，其中纤维腺瘤852例，导管内乳头状瘤49例，腺病112例。超声诊断乳腺癌的敏感度和特异度分别为90.2%(130/144)和67.7%(686/1 013)，诊断准确率为70.5(816/1 157)，乳光超诊断乳腺癌的敏感度和特异度分别为94.4% (136/144)和81.8%(829/1 013),诊断准确率为83.4% (965/1 157)。2种方法比较，乳光超诊断的敏感度与超声相比，差异无统计学意义(χ²＝1.772，P＝0.183)，但乳光超的诊断准确率及特异度比超声高(χ²＝55.148，P<0.001，χ²＝53.515，P<0.001)。ROC曲线结果显示：对0～1.0 cm及1.1～2.0 cm肿块，乳光超诊断效能优于超声(Z＝2.384，P＝0.034；Z＝3.962, P<0.001)；而对于>2.0 cm的肿块，乳光超的诊断效能与超声比较，差异无统计学意义(Z＝1.911, P＝0.056)。亚组分析显示：对于0～1.0 cm小肿块，乳光超特异度为85.7%(359/419)，高于超声特异度72.3%(303/419)(χ²＝22.555，P<0.001)。对于1.1～2.0 cm小肿块，乳光超特异度为80.4%(349/434)，高于超声特异度66.1%(287/434)(χ²＝22.613，P<0.001)。对于>2.0 cm肿块，乳光超特异度为73.1%(117/160)，超声特异度为58.1%(93/160)，乳光超特异度优于超声(χ²＝7.979，P＝0.005)。各亚组中，乳光超与超声的敏感度比较，差异均无统计学意义(P＝0.427; χ²＝0.374，P＝0.541; P＝1.000)。

乳光超对T₁期乳腺癌的诊断优于超声，可作为早期乳腺癌诊断的有效手段之一。

To investigate the application of the ultrasound diffusion breast imaging(UDBI) system in the diagnosis of T₁ stage breast cancer.

Totally 2 245 patients with ultrasound grade 3 and above (the patients in grade 6 were excluded) in the Chengdu Women’s and Children’s Central Hospital from April 2016 to April 2018 were included in this study. The patients underwent both ultrasonography and UDBI. Among them, 1 157 patients underwent biopsy, with complete pathological data, so they were retrospectively analyzed. According to the maximum diameter of the tumor, 1 157 cases were divided into 3 groups: 452 cases in S1 group (≤1.0 cm), 489 cases in S2 group (1.1-2.0 cm), and 216 cases in S3 group (>2.0 cm). The receiver operating characteristic (ROC) curve was used to analyze the diagnostic performance of the two methods. The area under the ROC curve and 95%CI were calculated accordingly. The area under the ROC curve was compared by Z test using the Medcalc V19.0.4 software. The sensitivity and specificity of the two methods were compared by χ² test or Fisher exact test. The paired χ² test (McNemar test) was used to compare the two methods with pathological results.

Among the 1 157 cases, there were 144 cases of malignant tumor (12.4%), confirmed as invasive ductal carcinoma, 1 013 cases of benign disease (87.6%), including 852 cases of fibroadenoma, 49 intraductal papilloma and 112 adenosis. The sensitivity, specificity and diagnostic accuracy in the diagnosis of breast cancer were 90.2% (130/144), 67.7% (686/1 013) and 70.5 (816/1 157) by ultrasound, 94.4% (136/144), 81.8% (829/1 013) and 83.4% (965/1 157) by UDBI, respectively. The diagnostic accuracy and specificity of UDBI were significantly higher than those of ultrasound (χ²=55.148, P<0.001, χ²=53.515, P<0.001), while the sensitivity of the two methods showed no significant difference(χ²=1.772, P=0.183). The ROC curve showed that for the tumor with the diameter of 0-1.0 cm or 1.1-2.0 cm, the diagnostic performance of UDBI was superior to that of the ultrasound (Z=2.384, P=0.034; Z=3.962, P<0.001); for the tumor with the diameter >2.0 cm, there was no significant difference in the diagnostic performance between UDBI and ultrasound (Z=1.911, P=0.056). Subgroup analysis showed that for the tumor with the diameter of 0-1.0 cm, the specificity of UDBI was significantly higher than that of ultrasound [85.7% (359/419) vs 72.3% (303/419), χ²=22.555, P<0.001]. For the tumor with the diameter of 1.1-2.0 cm, the specificity of UDBI was significantly higher than that of ultrasound [80.4% (349/434) vs 66.1% (287/434), χ²=22.613, P<0.001]. For the tumor with the diameter >2.0 cm, the specificity of UDBI was significantly higher than that of ultrasound [73.1% (117/160) vs 58.1% (93/160), χ²=7.979, P=0.005]. There was no significant difference in the sensitivity between UDBI and ultrasound in each subgroup (P=0.427; χ²=0.374, P=0.541; P=1.000).

The UDBI is superior to ultrasound in the diagnosis of breast cancer of T₁ stage, which can be used as an effective method for early breast cancer diagnosis.