Genetically optimized waveforms for contrast harmonic imaging with tri-state pulsers

The efficacy of contrast-enhanced ultrasound imaging relies on the selective detection of non-linear acoustic responses from microbubbles while suppressing linear tissue signals. Achieving optimal signal cancellation is challenging on systems employing tri-state pulsers, where fixed voltage levels and hardware non-linearities can degrade the performance of conventional pulse sequences and limit contrast-enhancement. We present a genetic algorithm (GA) that co-optimizes complementary transmit waveforms that can be encoded to perform specific cancellation schemes. The algorithm iteratively refines the transmit pulse parameters, including state levels and durations, to minimize the residual cancellation signal. A multi-objective fitness function guides the evolution to simultaneously enhance cancellation while maintaining sufficient signal amplitude for individual pulses. To validate this framework, we applied it to a pulse inversion amplitude modulation sequence on a Verasonics Vantage system with a P4-2 transducer. After 107 generations, the GA-derived waveforms achieved a 24.6-dB reduction in residual signal. In a microbubble flow phantom, the optimized transmit parameters yielded a contrast-to-tissue ratio of 16.6 dB, a 1.4-dB improvement over a conventional pulse-inversion sequence. The results demonstrate that this optimization methodology can create hardware-specific transmit schemes that enhance contrast imaging performance.