Harmonic Amplitude-Modulated Singular Value Decomposition for Ultrafast Ultrasound Imaging of Gas Vesicles

Abstract

Ultrafast nonlinear ultrasound imaging of gas vesicles (GV) contrast agents promises high-sensitivity biomolecular visualization with applications such as targeted molecular imaging of tumor markers or real-time tracking of gene expression. However, separating GV-specific signal from tissue remains challenging and requires the implementation of complex transmit schemes. In this work we introduce harmonic amplitude-modulated singular value decomposition (HAM-SVD), which synergizes pulse inversion (PI) with amplitude-modulated singular value decomposition (AM-SVD) to isolate GV-specific second-harmonic signals. In HAM-SVD, single-cycle plane waves at 9.6 MHz and five tilted angles (at a pulse repetition frequency of 2500 Hz) are transmitted under four duty cycles with alternating polarity. Beamformed IQ data are reshaped along a "space x pressure" matrix and decomposed via SVD; tissue background is cancelled by discarding the first and lowest singular modes, yielding an image comprised solely of pressure-dependent second harmonic signals. HAM-SVD sequence enables wide-field, ultrafast imaging without complex transmit sequences. Validation via simulations, in vitro phantoms, and in vivo rat lower limb experiments demonstrates HAM-SVDs outperformance compared to PI and AM-SVD. HAM-SVD is shown to achieve a 19.16 {+/-} 1.63 dB signal-to-background ratio (SBR) in vivo, surpassing PI (14.19 {+/-} 1.41 dB) and AM-SVD (15.79 {+/-} 1.38 dB). HAM-SVD overcomes limitations of conventional nonlinear techniques (e.g., depth restrictions, tissue clutter) by combining PIs harmonic sensitivity with AM-SVDs adaptive clutter filtering of tissue signals. This approach enhances molecular imaging specificity for GVs and holds potential for ultrasound localization microscopy of slow-flowing agents.