| Abstrakt: |
Objective: The goal of this study is to develop an efficient model to identify the factors that might help in predicting the status of the axillary lymph nodes(ALN)before sentinel lymph node biopsy(SLNB)in early breast cancer patients. Methods: This cross-sectional study was performed between2015and2017in Mashhad University Hospital, Iran. All female patients with early invasive breast cancer(T1-3and N0-1),having positive axillary ultrasound findings, undergoing a successful SLNB were included. Age at diagnosis, menopausal status, tumour size, tumour location, histological type, ultrasonographic findings, status of estrogen receptor (ER), progesterone receptor (PR), HER2 and Ki-67 were recorded. Logistic regression was performed using SPSS. Data and results: From 171 patients, 136 were randomly selected for the modeling group (82 had ALN positive disease), while another 35 were assigned to the validation group. In the univariate analysis, factors that were significantly associated with ALN metastasis included ER (P=0.022), PR (P=0.004), and Ki-67 (P=0.015) positivity, absence of hilum (P<0.001), higher maximum cortical thickness (P<0.001), and maximum transverse diameter (P<0.001). To avoid omitting significant indicators, factors with P<0.25 in univariate analysis were included in the multivariate analysis and predictive model. The model (p/1-p) = -9.376 - 0.34 × a- 0.88 × b + 1.26 ×c + 0.76 ×d + 0.44 × e, was generated (p = the probability of ALN metastasis; a = age at diagnosis > 35 years, b = postmenopausal status, c = absence of hilum, d = maximum cortical thickness of ALN as detected by ultrasound in mm, and e = maximum transverse diameter) and showed good performance for evaluation of ALN metastasis in the validation group with an AUC of 0.963. A cut-off value of 7.75mm was found for cortical thickness in predicating ALN metastasis. Conclusions: The predictive model presented here relies on readily available factors, indicating that it can be used to select patients who were more likely to have positive ALNs. However, this model should be applied prospectively to a large number of patients to verify its validity. [ABSTRACT FROM AUTHOR] |
