Establishment of sentinel lymph node metastasis prediction model for early breast cancer
Objective Based on the results of color Doppler ultrasound,molybdenum target X-ray and pathological immuno-histochemistry,to establish a nomogram model by screening independent risk factors forsentinel lymph node metastasis in early breast cancer,and to explore the feasibility of diagnosing sentinel lymph node metastasis through preoperative ultrasound,mo-lybdenum target examination and pathological characteristics of the primary tumor.Methods The clinicopathological data,col-or Doppler ultrasound features and molybdenum target features of 200 patients with early breast cancer admitted to our hospital from Jan.2022 to June 2023 were analyzed retrospectively.Apply logistic regression analysis to analyze the risk factors of sen-tinel lymph node metastasis and establish a column chart prediction model.Use internal validation to draw subject working curve(ROC),calibration curve,and clinical decision curve(DCA)for model evaluation.Results Univariate analysis showed that age,tumor diameter,histological grade,EGFR,CK5/6,Ki-67,AR,color ultrasound(tumor margin,calcification,spicule sign,blood flow signal,BI-RADS grade,lymph node cortex),molybdenum target(calcification,BI-RADS grade)were associated with sentinel lymph node metastasis(all P＜0.05),tumor number,location,quadrant,menopausal status,pathological classification,pathological type,ER,PR,Her-2 status Color ultrasound(posterior echo attenuation)and mo-lybdenum target(tumor margin,spicule sign,lobulation sign)had no effect on sentinel lymph node metastasis(P＞0.05).Multivariate logistic analysis showed that tumor diameter,histo-logical grade,ultrasound(tumor margin,calcification,BI-RADS grade),molybdenum target(BI-RADS grade),EGFR,and CK5/6 were independent influencing factors for sentinel lymph node metastasis(all P＜0.05).Using independent influ-encing factors to construct a column chart model and conducting internal validation,the AUC under the ROC curve was 0.898.The calibration curve shows that the standard curve fits well with the calibration curve.The clinical decision-making DCA curve shows that the model can generate good clinical benefits.Conclusion The sentinel lymph node metastasis predic-tion model based on the results of color ultrasound,molybdenum target and pathological detection has good performance and can help some patients with early breast cancer exempt from sentinel lymph node biopsy.
Breast cancerSentinel lymph nodesRisk factorsColumn chart model
NETL
NSTL
万方数据
