Elizabeth B. Herbst

Ultra-Low-Dose Ultrasound Molecular Imaging for the Detection of Angiogenesis in a Mouse Murine Tumor Model: How Little Can We See?

ObjectivesThe objective of this study was to evaluate the minimum microbubble dose for ultrasound molecular imaging to achieve statistically significant detection of angiogenesis in a mouse model. Materials and MethodsThe preburst minus postburst method was implemented on a Verasonics ultrasound research scanner using a multiframe compounding pulse inversion imaging sequence. Biotinylated lipid (distearoyl phosphatidylcholine–based) microbubbles that were conjugated with antivascular endothelial growth factor 2 (VEGFR2) antibody (MBVEGFR2) or isotype control antibody (MBControl) were injected into mice carrying adenocarcinoma xenografts. Different injection doses ranging from 5 × 104 to 1 × 107 microbubbles per mouse were evaluated to determine the minimum diagnostically effective dose. ResultsThe proposed imaging sequence was able to achieve statistically significant detection (P < 0.05, n = 5) of VEGFR2 in tumors with a minimum MBVEGFR2 injection dose of only 5 × 104 microbubbles per mouse (distearoyl phosphatidylcholine at 0.053 ng/g mouse body mass). Nonspecific adhesion of MBControl at the same injection dose was negligible. In addition, the targeted contrast ultrasound signal of MBVEGFR2 decreased with lower microbubble doses, whereas nonspecific adhesion of MBControl increased with higher microbubble doses. ConclusionsThe dose of 5 × 104 microbubbles per animal is now the lowest injection dose on record for ultrasound molecular imaging to achieve statistically significant detection of molecular targets in vivo. Findings in this study provide us with further guidance for future developments of clinically translatable ultrasound molecular imaging applications using a lower dose of microbubbles.

The Use of Acoustic Radiation Force Decorrelation-Weighted Pulse Inversion for Enhanced Ultrasound Contrast Imaging.

Objectives The use of ultrasound imaging for cancer diagnosis and screening can be enhanced with the use of molecularly targeted microbubbles. Nonlinear imaging strategies such as pulse inversion (PI) and “contrast pulse sequences” (CPS) can be used to differentiate microbubble signal, but often fail to suppress highly echogenic tissue interfaces. This failure results in false-positive detection and potential misdiagnosis. In this study, a novel acoustic radiation force (ARF)–based approach was developed for superior microbubble signal detection. The feasibility of this technique, termed ARF decorrelation-weighted PI (ADW-PI), was demonstrated in vivo using a subcutaneous mouse tumor model. Materials and Methods Tumors were implanted in the hindlimb of C57BL/6 mice by subcutaneous injection of MC38 cells. Lipid-shelled microbubbles were conjugated to anti-VEGFR2 antibody and administered via bolus injection. An image sequence using ARF pulses to generate microbubble motion was combined with PI imaging on a Verasonics Vantage programmable scanner. ADW-PI images were generated by combining PI images with interframe signal decorrelation data. For comparison, CPS images of the same mouse tumor were acquired using a Siemens Sequoia clinical scanner. Results Microbubble-bound regions in the tumor interior exhibited significantly higher signal decorrelation than static tissue (n = 9, P < 0.001). The application of ARF significantly increased microbubble signal decorrelation (n = 9, P < 0.01). Using these decorrelation measurements, ADW-PI imaging demonstrated significantly improved microbubble contrast-to-tissue ratio when compared with corresponding CPS or PI images (n = 9, P < 0.001). Contrast-to-tissue ratio improved with ADW-PI by approximately 3 dB compared with PI images and 2 dB compared with CPS images. Conclusions Acoustic radiation force can be used to generate adherent microbubble signal decorrelation without microbubble bursting. When combined with PI, measurements of the resulting microbubble signal decorrelation can be used to reconstruct images that exhibit superior suppression of highly echogenic tissue interfaces when compared with PI or CPS alone.

Pipe Phantoms With Applications in Molecular Imaging and System Characterization

Pipe (vessel) phantoms mimicking human tissue and blood flow are widely used for cardiovascular related research in medical ultrasound. Pipe phantom studies require the development of materials and liquids that match the acoustic properties of soft tissue, blood vessel wall, and blood. Over recent years, pipe phantoms have been developed to mimic the molecular properties of the simulated blood vessels. In this paper, the design, construction, and functionalization of pipe phantoms are introduced and validated for applications in molecular imaging and ultrasound imaging system characterization. There are three major types of pipe phantoms introduced: 1) a gelatin-based pipe phantom; 2) a polydimethylsiloxane-based pipe phantom; and 3) the “Edinburgh pipe phantom.” These phantoms may be used in the validation and assessment of the dynamics of microbubble-based contrast agents and, in the case of a small diameter tube phantom, for assessing imaging system spatial resolution/contrast performance. The materials and procedures required to address each of the phantoms are described.

Ultrasound molecular imaging with modulated Acoustic Radiation Force-based beam sequence in mouse abdominal aorta: A feasibility study

Ultrasound-based molecular imaging has been implemented in pre-clinical studies of cancer and cardiovascular diseases. Unfortunately, existing methods face substantial challenges in large blood vessel environments. We hypothesized that a clinically translatable method, the modulated Acoustic Radiation Force (ARF)-based imaging, is capable of rapid detection of inflammation in the abdominal aorta of a murine model. Mice stimulated with tumor necrosis factor (TNF)-α were used as an inflammation model (MInflammation). Age-matched normal mice were used as controls (MNormal). P-selectin-targeted (MBP-selectin), and isotype control (MBControl) microbubbles were synthesized by conjugating anti-P-selectin, and isotype control antibodies to the shell of microbubbles, respectively. The abdominal aorta of mice were imaged for 180 s during a constant infusion of microbubbles. The parameter produced from the new imaging sequence, residual-to-saturation ratio (RSR), was used to assess P-selectin expression. For the inflammation model, RSR of the MInflammation + MBP-selectin group was 40.9%, significantly higher (p <; 0.0005) than other groups. Feasibility was demonstrated to achieve rapid and statistically significant assessment of P-selectin in a mouse abdominal aorta for the first time. The proposed technique closes the gap toward rapid targeted molecular imaging in large blood vessels, and thus has the potential for early diagnosis and treatment selection of atherosclerosis via ultrasound molecular imaging.

Ultrasound Molecular Imaging of Inflammation in Mouse Abdominal Aorta

Objectives The aim of this study was to demonstrate a new clinically translatable ultrasound molecular imaging approach, modulated acoustic radiation force-based imaging, which is capable of rapid and reliable detection of inflammation as validated in mouse abdominal aorta. Materials and Methods Animal studies were approved by the Institutional Animal Care and Use Committee at the University of Virginia. C57BL/6 mice stimulated with tumor necrosis factor &agr;, or fed with a high-fat diet, were used as inflammation (MInflammation) and diet-induced obesity (DIO) (MDIO) models, respectively. C57BL/6 mice, not exposed to tumor necrosis factor &agr; or DIO, were used as controls (MNormal). P-selectin–targeted (MBP-selectin), vascular cell adhesion molecule (VCAM)-1–targeted (MBVCAM-1), and isotype control (MBControl) microbubbles were synthesized by conjugating anti–P-selectin, anti–VCAM-1, and isotype control antibodies to microbubbles, respectively. The abdominal aortas were imaged for 180 seconds during a constant infusion of microbubbles. A parameter, residual-to-saturation ratio (RSR), was used to assess P-selectin and VCAM-1. Statistical analysis was performed with the Student t test. Results For the inflammation model, RSR of the MInflammation + MBP-selectin group was significantly higher (40.9%, P < 0.0005) than other groups. For the DIO model, RSR of the MDIO + MBVCAM-1 group was significantly higher (60.0%, P < 0.0005) than other groups. Immunohistochemistry staining of the abdominal aorta confirmed the expression of P-selectin and VCAM-1. Conclusions A statistically significant assessment of P-selectin and VCAM-1 in mouse abdominal aorta was achieved. This technique yields progress toward rapid targeted molecular imaging in large blood vessels and thus has the potential for early diagnosis, treatment selection, and risk stratification of atherosclerosis.

The use of acoustic radiation force decorrelation weighted pulse inversion (ADW-PI) in enhancing microbubble contrast

Nonlinear echo signals from microbubbles are commonly isolated and visualized using techniques such as pulse inversion (PI) or contrast pulse sequences (CPS). Although these methods function by suppressing the linear signal component, they are still susceptible to detection of a false-positive nonlinear signal arising from strong tissue interfaces. In this study, acoustic radiation force (ARF) was added to a PI imaging sequence to induce motion of specifically-targeted microbubbles in a mouse hind limb tumor. The interframe signal decorrelation from this induced motion was then used to isolate microbubble signal based on its decorrelation characteristics. Results demonstrate that this technique of decorrelation filtering significantly improves microbubble contrast-to-tissue ratio (CTR) by 1.5 dB, and thereby allows for higher levels of specificity. ARF was shown to be effective in further enhancing performance.

Microbubble signal classification using normalized singular spectrum area based filtering methods

Targeted ultrasound contrast agents, comprising shell stabilized gas filled microbubbles (MBs), can be used to detect molecular markers of disease present on the vascular endothelium. Current MB imaging techniques exploit the nonlinear echo response of microbubbles to provide signal contrast with respect to adjacent tissue signal. However, these methods are hampered by false positive artifacts arising from nonlinear signals from strongly reflecting tissue interfaces. In this study, we demonstrate an image processing method that utilizes normalized singular spectrum area (NSSA) to distinguish adherent and non-adherent MB signals from static tissue signals with high specificity.

Microbubble signal classification using NSSA-based filtering methods

Microbubbles (MBs) are capable of binding specifically to molecular markers on the vascular endothelium to enable sensitive detection of early stage disease. Traditional microbubble imaging techniques rely on the nonlinearity of microbubble signal to differentiate it from surrounding tissue. However, the presence of harmonic energy among echogenic tissue interfaces can limit the effectiveness of these methods in enhancing microbubble contrast. In many cases, it is challenging to achieve any quantitative separation between static tissue signal and bound MB signal (Fig. 1B). In this study, we use normalized singular spectrum area (NSSA) to differentiate between static tissue signal, bound MB signal, and free MB signal in a mouse hindlimb tumor.

Low-dose ultrasound molecular imaging in mouse tumor model

Targeted microbubbles are currently used as ultrasound contrast agents in pre-clinical studies to visualize disease markers present on the blood vessel endothelium. Rapid clinical translation of targeted microbubble imaging requires a minimization of the diagnostically effective microbubble dose. In this study, we demonstrate an imaging method that improves signal-to-noise ratio (SNR) for microbubble detection and allows for accurate prediction of VEGFR2 targeting in a murine tumor model at 2.5% of the previously established minimum microbubble dose.