Augmenting detection of prostate cancer in transrectal ultrasound images using SVM and RF time series.

We propose a novel and accurate method based on ultrasound RF time series analysis and an extended version of support vector machine classification for generating probabilistic cancer maps that can augment ultrasound images of prostate and enhance the biopsy process. To form the RF time series, we record sequential ultrasound RF echoes backscattered from tissue while the imaging probe and the tissue are stationary in position. We show that RF time series acquired from agar-gelatin-based tissue mimicking phantoms, with difference only in the size of cell-mimicking microscopic glass beads, are distinguishable with statistically reliable accuracies up to 80.5%. This fact indicates that the differences in tissue microstructures affect the ultrasound RF time series features. Based on this phenomenon, in an ex vivo study involving 35 prostate specimens, we show that the features extracted from RF time series are significantly more accurate and sensitive compared to two other established categories of ultrasound-based tissue typing methods. We report an area under receiver operating characteristic curve of 0.95 in tenfold cross validation and 0.82 in leave-one-patient-out cross validation for detection of prostate cancer.

Detection of prostate cancer from RF ultrasound echo signals using fractal analysis.

In this paper we propose a new feature, average Higuchi dimension of RF time series (AHDRFT), for detection of prostate cancer using ultrasound data. The proposed feature is extracted from RF echo signals acquired from prostate tissue in an in vitro setting and is used in combination with texture features extracted from the corresponding B-scan images. In a novel approach towards RF data collection, we continuously recorded backscattered echoes from the prostate tissue to acquire time series of the RF signals. We also collected B-scan images and performed a detailed histopathologic analysis on the tissue. To compute AHDRFT, the Higuchi fractal dimensions of the RF time series were averaged over a region of interest. AHDRFT and texture features extracted from corresponding B-scan images were used to classify regions of interest, as small as 0.028 cm of the prostate tissue in cancerous and normal classes. We validated the results based on our histopathologic maps. A combination of image statistical moments and features extracted from co-occurrence matrices of the B-scan images resulted in classification accuracy of around 87%. When AHDRFT was added to the feature vectors, the classification accuracy was consistently over 95% with best results of over 99% accuracy. Our results show that the RF time series backscattered from prostate tissues contain information that can be used for detection of prostate cancer.