Abstract: A system and computer-implemented method for detecting flow of biological fluid, and facilitating the diagnosis of stenosis or other pathology of a hepatic artery or other non-carotid vasculature by using machine learning to analyze a spectral Doppler ultrasound waveform of blood flow or flow of other biological fluids. A region of interest is identified in a sonogram, a waveform envelope is generated for a waveform within the region of interest, an extracted waveform is generated based on the waveform envelope, a Fourier transform is performed on the extracted waveform, and a Fourier spectrum is generated. Machine learning is used to determine a likelihood of pathology based on waveform parameters, including the Fourier spectrum, and to determine a pathology prediction value. Additional patient information may be accessed and used in determining the likelihood of the pathology.

Abstract: A system and computer-implemented method for detecting flow of biological fluid, and facilitating the diagnosis of stenosis or other pathology of a hepatic artery or other non-carotid vasculature by using machine learning to analyze a spectral Doppler ultrasound waveform of blood flow or flow of other biological fluids. A region of interest is identified in a sonogram, a waveform envelope is generated for a waveform within the region of interest, an extracted waveform is generated based on the waveform envelope, a Fourier transform is performed on the extracted waveform, and a Fourier spectrum is generated. Machine learning is used to determine a likelihood of pathology based on waveform parameters, including the Fourier spectrum, and to determine a pathology prediction value. Additional patient information may be accessed and used in determining the likelihood of the pathology.