Abstract No. 240 Characterization of Shifts in the Histotripsy Treatment Zone When Treating Through Ribs in an Ex vivo Transcostal Model

Histotripsy is an emerging non-thermal, non-invasive and non-ionizing focused ultrasound (US) therapy. Cone-beam CT (CBCT) targeting has been developed to treat liver tumors not well visualized with US. Although the therapeutic US energy can be delivered through ribs, bone and intercostal tissue cause phase aberrations that can shift the focal point and treatment zone (TZ) location. Understanding the effect of transcostal tissue on the histotripsy TZ location is critical for accurate transcostal treatments. This work aims to characterize the effect of intervening ribs and intercostal tissue on the centroid location of the TZ. A section of a porcine rib-containing abdominal wall was encased in 1.5% agar with two pseudotumors (1-cm cube, made of porcine liver). The phantom was submerged in degassed water and targeted using CBCT with no aberration correction. Pseudotumor phantoms (n=2) without intervening tissue were used as controls (C). Acoustic cavitation was achieved using a 700kHz multielement therapy transducer (HistoSonics, Inc.). TZs over the pseudotumors were created (2.5 cm D) with transducer orientations specific to each target. Semi-automatic segmentations (3DSlicer) of treatment zones from post-treatment CBCTs were compared to the planned treatment zone location and size and the centroid offsets and percent overlap (Dice coefficient (DC), perfect overlap = 1) were calculated (MATLAB). Results were compared to in vivo measurements of focal point offsets from direct visualization of the bubble cloud with US (only z direction) at different depths. The differences between the centroid location of the segmented and planned TZ (mm offset) in X, Y, Z were -3.0, 1.6, 4.2 and -2.5,-0.6, 1.6 for TZ1 and TZ2 (Z = direction of US propagation). The DC was 0.7 and 0.8 for TZ1 and TZ2, respectively. Control treatments had an average offset of 0.5 mm and average DC of 0.9. In vivo Z offsets ranged from 3.7-5.5 mm at depths 24-42 mm below the skin (Table 240.1). Shifts in treatment zone location in the ex vivo model approximated the in vivo US measurements and were greater than ex vivo controls. All transcostal scenarios caused a shift in the TZ centroid compared to planned, suggesting that accounting for acoustic aberrations from ribs and intercostal tissue can improve targeting accuracy during transcostal treatment.