多模态超声特征构建列线图预测乳腺癌Ki-67表达

Prediction of Ki-67 expression in breast cancer by nomogram constructed with multi-modal ultrasound features

目的基于乳腺全容积扫描(ABVS)、辐射力脉冲(ARFI)、二维及彩色多普勒超声为基础的多模态超声特征构建列线图预测乳腺癌Ki-67表达的早期无创诊断效能。方法回顾性分析2021年1月—2023年12月本院277例乳腺癌患者超声及病理资料,分为Ki-67阴性组和阳性组,单因素及多因素Logistic筛选预测Ki-67阳性的独立影响因素以构建列线图。Hosmer-Lemeshow(HL)检验、校准曲线及决策曲线(DCA)分析评价列线图的校准度及临床应用价值。结果年龄(OR=0.974,95%CI:0.949~0.999)、肿块混合性回声(OR=746035163.7,95%CI:0.001~21347.012)、边缘微小分叶(OR=2.683,95%CI:1.133~6.352)为筛选后Ki-67阳性的独立影响因素(P<0.05),列线图曲线下面积(AUC)为0.669(95%CI:0.596~0.742),灵敏度87.1%,特异度69.5%;列线图=2.464~0.027×年龄-2.234×肿块回声(混合回声)+0.987×细分叶(有)。HL检验、校准曲线显示模型具有良好的校准;DCA表明在相同的预测概率下列线图的临床效能最大。结论多模态超声特征构建列线图可以在一定程度上预测乳腺癌Ki-67的表达,为临床早期采取治疗决策提供依据。

Objective To construct a nomogram based on multi-modal ultrasound features of automated breast volume scanning(ABVS),acoustic radiation force impulse(ARFI),two-dimensional,and color Doppler ultrasound to predict the early non-invasive diagnostic efficacy of Ki-67 expression in breast cancer.Methods The ultrasonic and pathological data of 277 breast cancer patients admitted to our hospital from January 2021 to December 2023 were retrospectively analyzed.Patients were divided into Ki-67 negative and positive groups.The independent influencing factors of Ki-67 positive were screened by single factor and multi factor Logistic regression to construct a nomogram.Hosmer-Lemeshow(HL)test,calibration curve and decision curve(DCA)analysis were used to evaluate the calibration and clinical application value of nomogram.Results Age(OR=0.974,95%CI:0.949-0.999),mixed echo of the mass(OR=746035163.7,95%CI:0.001-21347.012),and marginal lobulation(OR=2.683,95%CI:1.133~6.352)were identified as independent influencing factors for Ki-67 positivity after screening(P<0.05).The area under the curve(AUC)of the nomogram was 0.669(95%CI:0.596~0.742),with a sensitivity of 87.1%and specificity of 69.5%.The nomogram formula was as follows:Nomogram=2.464~0.027×Age-2.234×Mass Echo(for mixed echo)+0.987×Marginal Lobulation(if present).The HL test and calibration curve indicated that the model had good calibration.DCA showed that the nomogram had the highest clinical efficacy under the same prediction probability.Conclusions The nomogram constructed based on multi-modal ultrasound features can predict the expression of Ki-67 in breast cancer to a certain extent and provide a basis for early clinical treatment decisions.