The authors have declared that no competing interests exist.
Conceived and designed the experiments: DVdS LSS ADA SB SSG. Performed the experiments: DVdS LSS ADA SB. Analyzed the data: DVdS. Wrote the paper: DVdS SSG.
The emerging field of photoacoustic tomography is rapidly evolving with many new system designs and reconstruction algorithms being published. Many systems use water as a coupling medium between the scanned object and the ultrasound transducers. Prior to a scan, the water is heated to body temperature to enable small animal imaging. During the scan, the water heating system of some systems is switched off to minimize the risk of bubble formation, which leads to a gradual decrease in water temperature and hence the speed of sound. In this work, we use a commercially available scanner that follows this procedure, and show that a failure to model intra-scan temperature decreases as small as 1.5°C leads to image artifacts that may be difficult to distinguish from true structures, particularly in complex scenes. We then improve image quality by continuously monitoring the water temperature during the scan and applying variable speed of sound corrections in the image reconstruction algorithm. While upgrading to an air bubble-free heating pump and keeping it running during the scan could also solve the changing temperature problem, we show that a software correction for the temperature changes provides a cost-effective alternative to a hardware upgrade. The efficacy of the software corrections was shown to be consistent across objects of widely varying appearances, namely physical phantoms, ex vivo tissue, and in vivo mouse imaging. To the best of our knowledge, this is the first study to demonstrate the efficacy of modeling temporal variations in the speed of sound during photoacoustic scans, as opposed to spatial variations as focused on by previous studies. Since air bubbles pose a common problem in ultrasonic and photoacoustic imaging systems, our results will be useful to future small animal imaging studies that use scanners with similarly limited heating units.
Photoacoustic tomography (PAT) is a hybrid optical and ultrasound imaging technique that combines the high spatial resolution of ultrasound with the high optical contrast of diffuse optical imaging. It is a highly promising modality that has received increasing attention from the biomedical imaging community because of its ability to image in three dimensions, deep into living tissue. PAT is based on the illumination of an object with brief pulses of laser light. As the light diffuses into the tissue, it is absorbed differentially by the various regions of the tissue. The light absorption causes brief localized heating and an increase in pressure proportional to the amount of light energy absorbed. This leads to the generation of pressure waves that travel out of the tissue and can be measured by ultrasonic transducers. The detected signals are then reconstructed into an image of optical absorption using a computer algorithm.
Photoacoustic tomography systems of many different geometries have been proposed and investigated in the literature, such as circular
Regarding the previous literature on speed of sound considerations in photoacoustic imaging, most studies so far have assumed a speed of sound distribution that does not change over the course of a scan, whether it be spatially homogeneous or heterogeneous
As for the remainder of this paper, the
Photoacoustic wave propagation is described by the general wave equation, given by
In this study, we acquired photoacoustic scans using the Nexus 128 scanner, manufactured by Endra Life Sciences (Ann Arbor, MI, USA).
(a–c) are reproduced with permission from
The system uses a tunable, near-infrared Nd:YAG laser that produces 7 ns pulses at up to 25 mJ per pulse. The wavelength can be tuned within a wavelength range of 680–980 nm. The repetition rate of the laser is 20 Hz. To reduce the effect of laser pulse power variations, and to increase the SNR of the recorded pressure signals, we average signals over a predetermined number of pulses (typically 100). At the laser repetition rate of 20 Hz, a standard scan using 180 bowl positions therefore takes 15 minutes per excitation wavelength.
A number of studies have determined the speed of sound in water
a |
|
f |
|
b |
|
g |
|
c |
|
h |
|
d |
|
i |
|
e |
|
j |
|
All animal work was conducted in accordance with the guidelines of and approved by the Administrative Panel on Laboratory Animal Care at Stanford University.
In this study, we have imaged two physical phantoms, an
The vessel phantom in (b) is a print-out of a digital phantom provided in the k-Wave Toolbox
(a) whole brain, (b–g) representative slices.
In this study, we made use of three different scanning protocols:
protocol 1: 72 angles, 5 pulse averaging (scan time: 20 seconds)
protocol 2: 180 angles, 100 pulse averaging (scan time: 15 minutes)
protocol 3: 180 angles, 400 pulse averaging (scan time: 60 minutes)
All scans were acquired at a wavelength of 750 nm. Protocol 1 represents a typical preliminary scan used to verify that the object is positioned correctly. Protocols 2 and 3 represent realistic scanning protocols as well, where the duration of protocol 2 is typical of a single wavelength scan, and that of protocol 3 is typical of a multi-wavelength scan. The main difference of interest between the protocols is in their duration, leading to differing amounts of cooling of the water during the scan. Note that the different angle numbers and pulse averaging numbers result in additional sources of difference between the scans. However, the chosen parameters suggest that image quality should improve from protocol 1 to 3, as the signal sampling improves. In the
As discussed in the ‘Photoacoustic scanner’ section, the standard protocol recommended by the manufacturer is to switch off the water heating pump during a scan to reduce the risk of air bubble formation. However, this causes the water temperature to drop slightly during a scan, and increasingly so for scans of increasing duration. As mentioned, we found this drop to be on the order of 1.5 degrees for a 15 minute scan, and 4.5 degrees for a 60 minute scan. While these temperature drops cause changes in the speed of sound in water of only 0.2% and 0.6%, respectively, we found that the impact on image quality can be significant. This is illustrated in
All reconstructions in a single row are displayed on the same grayscale (arbitrary units).
For the dot phantom, the temperature drop results in a swirl-like artifact of increasing strength for longer scans. For the vessel phantom, the temperature drop results in the formation of extra vessel branches (see red arrow in
Next, we investigated the efficacy of monitoring the water temperature during a scan, and applying a post-acquisition temperature (and hence speed of sound) correction for each bowl position separately. This was done by reading the water temperature at the end of the signal acquisition at each bowl position. By way of validation, we compared the resulting reconstructions to those obtained by scanning the phantoms with the water heating system switched on and assuming a constant water temperature. The phantom positions were identical for every scan. Extra care was taken to make sure that leaving the water heating system on during the scan did not produce air bubbles.
All reconstructions in a single row are displayed on the same grayscale (arbitrary units).
All reconstructions in a single row are displayed on the same grayscale (arbitrary units). During reconstruction, the water temperature was assumed to be the mean of the water temperature before and after the scan. Since the water pump was switched on, these temperatures were nearly identical and approximately equal to the target temperature of 38°C.
While viewing individual slices allows visualizing small local structures, they are not always useful for interpreting an entire 3D volume. For example, not all of the phantom structures are visible in the slices shown in
All reconstructions in a single row are displayed on the same grayscale:
All reconstructions in a single row are displayed on the same grayscale: 0 to 1 (arbitrary units).
Here, we examined the benefit of temperature monitoring and subsequent intra-scan speed of sound corrections for an
By applying the speed of sound correction, the vessel bifurcation in the lower left quadrant becomes visible (red arrow). The vessel coming in from the top left also becomes clearer.
Lastly, to demonstrate the benefit of variable speed of sound corrections on an image of yet a different appearance,
The correction for the changing speed of sound (d) greatly focuses the anatomical structures present in the image, compared to when reconstructing with the initial (a), average (b) or final (c) water temperature. These results illustrate that the image quality cannot be maximized by simply optimizing a single speed of sound value. Also note the close resemblance of (d) to the physical slice shown in
The emerging field of photoacoustic tomography for molecular imaging is rapidly evolving with many new system designs and reconstruction algorithms being published. However, for imaging systems that use water heated to body temperature as a coupling medium, relatively little attention has been paid to the importance of water temperature changes during a scan. The work presented in this paper addresses this gap. For intra-scan temperature drops as small as 1.5 to 4.5 degrees (resulting in changes in the speed of sound of only 0.2
In this study, we assumed that the speed of sound was uniform throughout the image domain (encompassing both the coupling water and the scanned object) at any point in time.
The authors would like to thank the anonymous reviewers for their comments, which helped to improve this paper. The authors would also like to thank Endra Life Sciences for providing access to their Nexus 128 photoacoustic scanner, and Dr. Cristina Zavaleta for help with preparing the printed phantoms.