Predicting the expression level of Ki-67 in breast cancer using multi-modal ultrasound parameters

Abstract

Objective: This study investigated the feasibility of predicting the expression levels of Ki-67 in breast cancer using ultrasonographic findings and clinical features. Methods: Fifty-eight breast cancer patients, with 82 lesions confirmed by surgical pathology, were selected retrospectively for this study. Conventional ultrasound examination and elastography examination were performed before surgery. Clinical features (age, estrogen receptor (ER), progesterone receptor, and human epidermal growth factor receptor-2 expression levels), ultrasonographic findings, and elastography scores, including the maximum size, location, number, margin, borderline, blood flow, and elastography score of the mass, were collected. The expression of Ki-67 was recorded using immunohistochemical staining, and the patients were divided into a high (≥ 20%) expression group and a low (< 20%) expression group. SPSS 23.0 software was used for statistical analysis. An independent sample t-test was used for measurement data, and a χ2 test was used for enumeration data. Logistic regression analysis was performed for meaningful indicators, and the receiver operating characteristic curve was used to calculate the best diagnostic cut-off point. Results: Monofactorial analysis showed that there was a statistically significant difference (p < 0.05) between the high expression of Ki-67 and the maximum diameter of the mass, the margin of the mass, the color Doppler flow imaging of the blood flow, and the resistance index of the blood flow. There were no significant differences in age, mass location, number, morphology, borderline, microcalcification, and elastography score (p > 0.05). Multiple factor regression analysis showed that a large mass and a mass with a rich blood flow had an independent predictive value for Ki-67. When the diameter was > 21.5 mm, the diagnostic sensitivity and specificity were 91.9% and 71.3%, respectively. The expression level of Ki-67 was negatively correlated with that of ER. Conclusion: The tumor size and blood flow of breast cancer is correlated with the expression level of Ki-67 and, thus, it could be used to predict the expression level of Ki-67 in ultrasound diagnosis. The margin condition and the expression level of ER of an ultrasonic mass could also indirectly reflect the Ki-67 expression level of the mass.