Objective:

To evaluate the diagnostic value of histogram analysis using B-mode gray-scale ultrasound histogram analysis (HA) to determine the infiltration degree in breast cancer.

Methods:

A total of 224 patients with surgically confirmed breast cancer were enrolled in this study. Ultrasound feature

s of patients were reviewed retrospectively. Eighteen HA parameters were derived using Omni-Kinetics software (GE Healthcare). The reproducibility of those parameters was evaluated using two independent delineations conducted by two observers. In order to minimize the potential influence of individual differences

the relative value of each HA parameter was used for statistical analysis. These relative value of HA parameters were compared among the three different infiltration degree in breast cancer (non-invasive

early invasive and invasive breast cancer) using Kruskal-Wallis H test and multivariable ordered logistic regression analysis.

Results:

All the 18 parameters presented excellent reproducibility

with intraclass correlation coefficient (ICC) values from 0.799 to 0.997. Except for Min intensity and Mean deviation

there were statistically difference in the other 16 parameters relative values. Multivariate logistic regression showed that the relative values of Mean intensity

Skewness

Uniformity

Quantile5

Quantile10

Quantile75

and Quantile90 were helpful to determine the infiltration degree in breast cancer.

Conclusion:

HA parameters derived from ultrasound can be used as a reliable quantitative tool to determine the infiltration degree in breast cancer.