Breast cancer diagnosis based on ultrasound RF echo modeling and physician's level of confidence

A number of researchers have shown that the ultrasound RF echo from tissue exhibits (1/f ) /sup /spl beta// characteristics and developed tissue characterization methods based on the fractal parameter of the received signal. In this paper a new model for the received ultrasound RF data has been proposed, namely the fractional differencing auto regressive moving-average (FARMA) model, whose parameters were investigated for their ability to differentiate between benign and malignant tumors. Along with the FARMA model parameters, the patient's age and the radiologist's pre-biopsy level of suspicion (LOS) were used as additional features to increase the characterization performance. 120 in vivo B-scan images obtained from 90 patients were used during the modeling and estimation procedures. The area under the receiver operator characteristics (ROC) curve yields a value of 0.8723, with a confidence interval of [0.8512, 0.8945], at a significance level of 0.05. These results indicate that the proposed tissue characterization method can be used as a second opinion to aid the radiologist decision criteria.

[1]  Keinosuke Fukunaga,et al.  Introduction to Statistical Pattern Recognition , 1972 .

[2]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1972 .

[3]  F.S. Cohen,et al.  WOLD decomposition of the backscatter echo in ultrasound images of soft tissue organs , 1997, IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control.

[4]  R. Kashyap,et al.  ESTIMATION IN LONG‐MEMORY TIME SERIES MODEL , 1988 .

[5]  Raj Acharya,et al.  Fractal analysis of self-similar textures using a Fourier-domain maximum likelihood estimation method , 1996, Proceedings of 3rd IEEE International Conference on Image Processing.

[6]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1971 .

[7]  Athina P. Petropulu,et al.  ROC analysis of ultrasound tissue characterization classifiers for breast cancer diagnosis , 2003, IEEE Transactions on Medical Imaging.

[8]  M. Giger,et al.  Computerized analysis of lesions in US images of the breast. , 1999, Academic radiology.

[9]  George E. P. Box,et al.  Time Series Analysis: Forecasting and Control , 1977 .

[10]  M. A. Kutay,et al.  On modeling biomedical ultrasound RF echoes using a power-law shot-noise model , 2001, IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control.

[11]  B. Onaral,et al.  Modeling of scattering dominated ultrasonic attenuation using power-law function , 2000, 2000 IEEE Ultrasonics Symposium. Proceedings. An International Symposium (Cat. No.00CH37121).