OBJECTIVE: Histotripsy is a non-invasive acoustic ablation technique that leverages cavitation to impart mechanical damage to a viscoelastic medium, such as tissue. Although histotripsy bubbles and lesions can be imaged with a variety of modalities, reliable methods to predict tissue disruption across different tissue-types remain to be determined. APPROACH: Several ex-vivo bovine tissues were ablated by intrinsic threshold histotripsy over a range of pulse-per-location acoustic doses. Acoustic Cavitation Emission (ACE) signals were captured following every other therapeutic pulse using transmit-receive capable histotripsy arrays. Final bubble lifespan, lifespan-slope, and percent-reduction were calculated and correlated against histologic necrosis score (0-5: 0=0% necrosis, 5=>
95% necrosis) and residual structure score (0-4: 0=none present, 4=intact) to evaluate the ability of features from ACE-signals to predict histotripsy-induced damage. Further, optimal ACE-feature thresholds were determined for binary evaluation of whether a necrosis score equal or greater than 4 had been reached. RESULTS: Measured lifespans increased and lifespan-slopes decreased with pulses per location (ppl) and eventually plateaued in all tissue types, in similar trends to those previously observed in tissue phantoms. Necrosis score increased and residual structure decreased with increasing acoustic dose. Bubble lifespan-slope and percent-reduction correlated well with necrosis score. Thresholds able to predict the necrosis score of 4 or greater in brain, liver, and kidney were calculated with high sensitivity/specificity (>
80%). The necrosis score of 4 and 5 is expected to correspond to near-complete/complete ablation by histological evaluation. CONCLUSION: Features measured from ACE-signals, particularly the lifespan-slope and percent reduction, were used to predict near-complete/complete ablation of large-volume histotripsy treatments in ex vivo bovine liver, kidney, and brain tissues with good accuracy. Tissue heterogeneities were observed to impact the histotripsy damage and corresponding ACE-signals, and thus the predication accuracy.