Page 24 - Spring 2019
P. 24

‘l'|'iarn=ni.n1ie lkvieafinn
ml K“; i
Babb/efsl al ‘ % A my M]: PM ‘ L
_ lwspnnse
spnenaei wavefront g
Estimate _,_,,_ caminne {if “MY "°"‘ K.
.m**§~¢‘ <1<“ _ $1 
L
’ Steerlo
 — Iocain-ins in C?
_ | ROI
t are + «»e
M
Figure 2.Ti'me darnain illustration nfpassive ai.-aastic mapping (PAM) processing. Bubble emissions are received on an army afsensms. Sig-
mils (black) have relative delays that are chavacterittir oftlie distance between the array and the source lacatian. After filtering the raw data
to isalate either bmndband ar mzrrowbtzrid izmusti: emissilms of interest, the first pmrusing step is to steer the army to a paint in the region nf
interest (ROI) by applying time shifts to each array element. Ifthe steered laeatian matches that afthe source (x = x_), the signak will be time
aligned (red); otherwise, the signals will be tzrnpomlly misaligned (blue; x e x). The second processing step combiner the timeshifted signals
to estimate signal pnwer; poorly correlated data will lead ta a low power estimate while well-ronelated data will identify the source. Repeating
this process over a grid in the ROI lends ta an image afsource (bubble) strength (battain left). The image has a dynamic range ofl00, and red
iridicutfl a maximum value.
domain, but frequency domain implementations offer equiva- in tissue sound speed and attenuation. Unlike MR or active
lent performance with potentially lower calculation times. ultrasonic methods, PAM data must be superimposed on tis-
The roots of PAM techniques are found in passive beam- 5“: mofphulogy Images Produced by _mher Lmagmg nfethflds
forming research Performed in the (omen of Seismology) to provide a context for treatment guidance and monitoring.
underwater acoustics, and radar. The utility of these tecl'i- Clinical PAM Example
niques comes from their specific benefits when applied to Over the last decade, PAM research has progressed from
noninvasive cavitation monitoring. small-rodent to large-primate in vivo models, and recently,
o Images are formed i.ri the near field of the receive array so a clinical cavitation mapping dataset was collected during a
that sources may be identified in at least two dimensions Phase 1 trial of ultrasound-mediated liver heating for drug re-
(e.g., distance and angle). lease from thermally sensitive liposomes (Lyon et al., 2.018).
. Received data may he filtered to identify bubble-specific Figure 3 shows an axial computed tomography (CT) image of
emissions (half-integer harmonics or broadband noise one trial participant, including the targeted liver tumor. The
elevation), thereby decluttering the image of nonlinear incident therapeutic-focused ultrasound beam (FUS; Figure
background scattering and identifying imaging regions 3, red arrow) was provided by a clinically approved system,
that have different cavitation behaviors. while the PAM data (Figure 3, lzlue arrow) was collected us-
o A single diagnostic ultrasound array system can be used ing a commercially available curvilinear diagnostic array. The
to provide both tissue imaging and cavitation mapping CT-overlaid PAM image (and movie showing six consecutive
capabilities that are naturally coaligned so that the moni- PAM frames, see acousticstodayorg/g;i'ay-media) was gener-
toring process can describe the tissue and bubble status ated using a patient-specific sound speed estimate and an
before, during, and after therapy. adaptive beamformer to minimize beamwidth. Although
. Real-time PAM may allow automated control of the ther- the monitoring was performed over an hour-long treat-
apy process to ensure procedural safety. ment, only a small handful of cavitation events was detected
For both passive and active ultrasonic methods, image qual- (<0.1% of exposures). This was as expected given that no ex-
ity and quantitative accuracy may be limited by uncertainties ogenous microbuhhles were used and the treatment settings
ea 1 AI:i'.iulI:II:- Tbdly 1 Spring znis





















































   22   23   24   25   26