Ultrasound RF Time Series for Classification of Breast Lesions

Abstract

This work reports the use of ultrasound radio frequency (RF) time series analysis as a method for ultrasound-based classification of malignant breast lesions. The RF time series method is versatile and requires only a few seconds of raw ultrasound data with no need for additional instrumentation. Using the RF time series features, and a machine learning framework, we have generated malignancy maps, from the estimated cancer likelihood, for decision support in biopsy recommendation. These maps depict the likelihood of malignancy for regions of size 1 mm(2) within the suspicious lesions. We report an area under receiver operating characteristics curve of 0.86 (95% confidence interval [CI]: 0.84%-0.90%) using support vector machines and 0.81 (95% CI: 0.78-0.85) using Random Forests classification algorithms, on 22 subjects with leave-one-subject-out cross-validation. Changing the classification method yielded consistent results which indicates the robustness of this tissue typing method. The findings of this report suggest that ultrasound RF time series, along with the developed machine learning framework, can help in differentiating malignant from benign breast lesions, subsequently reducing the number of unnecessary biopsies after mammography screening.

A Fusion-Based Approach for Breast Ultrasound Image Classification Using Multiple-ROI Texture and Morphological Analyses

“…For example, in quantitative ultrasound imaging of cancer [42, 43], the raw ultrasound radio-frequency (RF) signals are divided into small ROIs, and each ROI is analyzed to extract spectral features for tissue characterization. Moreover, a recent study by Uniyal et al [44] has compared the classification performance of a combination of ultrasound-based texture, spectral, and RF time series features that are extracted from the entire breast tumor with the performance obtained by dividing the tumor into 1 × 1 mm 2 ROIs and extracting similar ultrasound-based features from each individual ROI. This study demonstrates that the classification performance obtained by classifying the individual 1 × 1 mm 2 ROIs outperforms the classification results achieved by classifying the entire tumor.…”

Breast Cancer Classification with Ultrasound Images Based on SLIC

“…Moon et al located the region of tumor, and then classified the benign and malignant tumor with interior echo and morphological characteristic [10]. Uniyal et al utilized RF time sequence characteristics and machine learning method to generalize the estimated probability graph of tumor and then realized classification [11]. And Nayeem et al proposed a method based on sparse representation to classify tumor [12].…”

Effectiveness of Existing CAD-Based Research Work towards Screening Breast Cancer

“…It was found that the dictionary significantly assists in classification process. Uniyal et al [44] have implemented a technique that uses time-series analysis to perform classification of the malignancy in breast cancer. The technique generates a map of malignancy using machine learning approach.…”