Abstract:

Objective　To investigate the detection rate of nipple-areola complex (NAC) involvement in patients with breast cancer, and to construct a nomogram predictive model based on conventional ultrasound and contrast-enhanced ultrasound (CEUS) signs to guide preoperative evaluation. Methods　A total of 108 patients with breast cancer who were pathologically diagnosed and eligible for mastectomy in Shaanxi Nuclear Industry 215 Hospital from October 2022 to October 2023 were retrospectively summarized. They were (44.8±10.4) years old. They were divided into an NAC positive group (35 cases) and an NAC negative group (73 cases) according to the postoperative routine pathological results. Before the operation, conventional two-dimensional color Doppler ultrasound was used to measure the tumor maximum diameter, tumor to nipple distance (TND), tumor location, type, shape, direction, edge, boundary, echo pattern, internal uniformity, blood vessels, microcalcification, and echo pattern extending from lesion to catheter. CEUS was used to measure the maximum tumor diameter, TND, catheter enhancement extension to nipple (DEEN), focal nipple enhancement, enhanced uniformity, edge enhancement, contour enhancement, sequence of enhancement, perfusion defect, and extent of enhancement. The independent-sample t test and χ² were used to compared tha data between the groups. Multivariate logistic regression was used to screen the risk factors of NAC involvement and establish a nomogram model. The predictive performance was verified by the receiver operating characteristic curves (ROC), calibration, and decision analysis (DCA). Results　Compared with the negative group, the positive group showed increases of peripheral tumors, microcalcifications, ductal echo extending from lesion, as well as CEUS measurements of DEEN and focal nipple enhancement, while the TND measured by two-dimensional ultrasound and CEUS were significantly shorter (all P<0.05). The regression analysis showed that ductal echo extending from lesion (full course: OR=2.452, 95%CI 1.590-3.782, P<0.001; segmental: OR=1.992, 95%CI 1.252-3.169, P<0.001), DEEN (OR=1.385, 95% CI 1.018-1.885, P=0.003), and focal nipple enhancement (OR=3.435, 95%CI 1.654-7.135, P<0.001) were the main risk factors of NAC involvement. After establishing the nomogram using the R software, the AUC calculated by ROC was 0.876 (95%CI 0.811-0.935, P<0.001), indicating good predictive accuracy of the model. The calibration and DCA curves showed good agreement and clinical net benefit ratio of the model. Conclusion　The nomogram model developed by conventional ultrasound and CEUS typical signs has important potential for predicting the risk of NAC involvement in patients with breast cancer taking breast conserving resection.

Key words:

Breast cancer, Breast conserving resection, Nipple-areola complex, Contrast-enhanced ultrasound, Nomogram

摘要：

目的　调查乳腺癌患者乳头-乳晕复合体（NAC）受累的检出率，并基于常规超声和对比增强超声（CEUS）征象构建列线图预测模型以指导术前评估。方法　回顾性总结2022年10月至2023年10月陕西省核工业二一五医院病理确诊乳腺癌并符合保乳切除术的患者共108例，年龄（44.8±10.4）岁，根据术后常规病理结果分为NAC受累阳性组35例和阴性组73例。术前常规二维彩色多普勒超声测量肿瘤最大径、肿瘤到乳头的距离（TND）、肿瘤位置、类型、形状、方向、边缘、边界、回声模式、内部均匀性、血管、微钙化、从病变延伸到导管的回声模式，CEUS测量肿瘤最大径、TND、导管增强延伸至乳头（DEEN）、局灶性乳头增强、增强均一性、边缘增强、外形增强、增强顺序、灌注缺损和增强范围。组间比较采用独立样本t检验、χ²检验，多因素logistic回归筛选NAC受累的危险因素，并建立列线图模型。经受试者操作特征曲线（ROC）、校准以及决策分析（DCA）曲线进行预测性能的验证。结果　与阴性组相比，阳性组二维超声测量周围型肿瘤、微钙化、从病变延伸到导管的回声及CEUS测量DEEN、局灶性乳头增强明显增多，而二维超声和CEUS测量TND缩短（均P<0.05）。回归分析显示，从病变延伸到导管的回声（全程：OR=2.452，95%CI 1.590～3.782，P<0.001；节段：OR=1.992，95%CI=1.252～3.169，P<0.001）、DEEN（OR=1.385，95%CI 1.018～1.885，P=0.003）和局灶性乳头增强（OR=3.435，95%CI 1.654～7.135，P<0.001）是NAC受累的主要危险因素。R软件建立列线图后ROC计算AUC为0.876（95%CI 0.811～0.935，P<0.001），提示模型的预测准确度较好。校准和DCA曲线显示模型有较好的吻合度和临床净获益比。结论　利用常规超声和CEUS典型征象开发的列线图模型对预测保乳切除术乳腺癌患者NAC受累发生风险有重要的应用潜力。

关键词:

乳腺癌, 保乳切除术, 乳头-乳晕复合体, 对比增强超声, 列线图