Ultrasound contrast microbubbles in imaging and therapy: physical principles and engineering

Abstract

Microbubble contrast agents and the associated imaging systems have developed over the past 25 years, originating with manually-agitated fluids introduced for intra-coronary injection. Over this period, stabilizing shells and low diffusivity gas materials have been incorporated in microbubbles, extending stability in vitro and in vivo. Simultaneously, the interaction of these small gas bubbles with ultrasonic waves has been extensively studied, resulting in models for oscillation and increasingly sophisticated imaging strategies. Early studies recognized that echoes from microbubbles contained frequencies that are multiples of the microbubble resonance frequency. Although individual microbubble contrast agents cannot be resolved-given that their diameter is on the order of microns-nonlinear echoes from these agents are used to map regions of perfused tissue and to estimate the local microvascular flow rate. Such strategies overcome a fundamental limitation of previous ultrasound blood flow strategies; the previous Doppler-based strategies are insensitive to capillary flow. Further, the insonation of resonant bubbles results in interesting physical phenomena that have been widely studied for use in drug and gene delivery. Ultrasound pressure can enhance gas diffusion, rapidly fragment the agent into a set of smaller bubbles or displace the microbubble to a blood vessel wall. Insonation of a microbubble can also produce liquid jets and local shear stress that alter biological membranes and facilitate transport. In this review, we focus on the physical aspects of these agents, exploring microbubble imaging modes, models for microbubble oscillation and the interaction of the microbubble with the endothelium.

Figure 1

Observed and predicted microbubble oscillation. (a) and (b): experimental streak images of 1.4-μm lipid-shelled microbubbles insonified with a 2.25 MHz pulse with a PRP of ~100 (a) or 300 kPa. (b). The streak image is an optical image acquired with an ~15 ns shutter showing a single line through the center of the microbubble (shown as dashed line in cartoon) as it expands and contracts during the ultrasound pulse. Momentum is transferred to the microbubble from the ultrasound pulse resulting in a net displacement of the microbubble. At the higher ultrasound pressure, the microbubble fragments and the resulting small microbubbles can be visualized. (c) and (d): predicted microbubble expansion ratio versus time for a bubble with an initial size of 2 μm during insonation with 1-MHz 5-cyle ultrasonic pulses using the equation and parameters as in (Zhao et al., 2004). (c) For a low transmitted PRP, microbubble oscillation is harmonic at a frequency close to the incident ultrasound frequency and the bubble expansion ratio is nearly proportional to the transmitted PRP; (d) For a high transmitted PRP, the oscillation nonlinearity increases and the expansion ratio increases more rapidly than the transmitted PRP.

Figure 1

Observed and predicted microbubble oscillation. (a) and (b): experimental streak images of 1.4-μm lipid-shelled microbubbles insonified with a 2.25 MHz pulse with a PRP of ~100 (a) or 300 kPa. (b). The streak image is an optical image acquired with an ~15 ns shutter showing a single line through the center of the microbubble (shown as dashed line in cartoon) as it expands and contracts during the ultrasound pulse. Momentum is transferred to the microbubble from the ultrasound pulse resulting in a net displacement of the microbubble. At the higher ultrasound pressure, the microbubble fragments and the resulting small microbubbles can be visualized. (c) and (d): predicted microbubble expansion ratio versus time for a bubble with an initial size of 2 μm during insonation with 1-MHz 5-cyle ultrasonic pulses using the equation and parameters as in (Zhao et al., 2004). (c) For a low transmitted PRP, microbubble oscillation is harmonic at a frequency close to the incident ultrasound frequency and the bubble expansion ratio is nearly proportional to the transmitted PRP; (d) For a high transmitted PRP, the oscillation nonlinearity increases and the expansion ratio increases more rapidly than the transmitted PRP.

Figure 2

(a) Cartoon demonstrating the mechanisms of ultrasound contrast agent destruction, including diffusion, acoustically-driven diffusion and fragmentation. (b) Recordings of the echoes from single microbubbles during insonation with a train of 4.4-MHz pulses with a pulse duration of 1.5 cycles, pulse repetition frequency of 1 kHz and varied PRP.

Figure 3

(a) Relative expansion of microbubbles during insonation at 2.25 MHz with PRP ranging from 310 kPa to 1200 kPa. (b) Relative expansion of microbubbles for 1, 1.5, 2, and 3.5 MHz transmitted center frequencies.

Figure 4

In vitro and ex vivo optical images of microbubbles oscillating near boundaries. (a) A cavitating bubble collapsing and impinging on a gel surface, Reproduced with permission from (Kodama and Tomita, 2000). (b) Microbubble at boundary of cultured cell with feature in center that indicates jet formation. Reproduced with permission from (Prentice et al., 2005). (c) Microbubble oscillation in a microvessel within the rat cecum undergoing asymmetric oscillation with feature in center indicating toroidal microbubble shape. Reproduced with permission from (Caskey et al., 2007). (d) Microbubble coalescence shown in ex vivo microvessel with large bubble interacting with vessel wall. Reproduced with permission from (Caskey et al., 2007).

Figure 5

A snapshot of the predicted pressure field induced by a microbubble with an initial diameter of 3 μm as it oscillates in a microvessel with inner diameter of 8 μm in an ultrasound field with a PRP of 0.5MPa and center frequency of 1MHz.

Figure 6

Diagram of the generation and processing of contrast agent echoes for the creation of images, corresponding to the entries in Table 3. The detection of echoes from microbubble contrast agents may be accomplished by processing the returned echo from a single pulse or the returned echoes from a train of ultrasound pulses.

Figure 7

Examples of ultrasound contrast agent images of Met-1 murine tumor. (a) B-mode and (b) corresponding Contrast Pulse Sequence (CPS) (gold color) image overlaid on B-mode (grey). Region of low echogenicity in B-mode image (circled in red) is region of tumor, with vascular density shown by the gold contrast agent overlay.

Figure 8

Hypothesized mechanisms of drug transport across endothelium. Illustration of (a) local shear stress created on cell during microbubble oscillation, (b) fluid jet formation, and (c) intra-cellular transport that are hypothesized to result from the stresses induced by microbubble activity, including generation of gaps at tight junctions, expression of cell adhesion molecules due to inflammatory process and the creation of vesicles for trans-cellular transport.

Figure 9

(a) Glomerular capillary hemorrhage vs the ratio of PRP to center frequency for a range of transmitted ultrasound frequencies using a diagnostic ultrasound scanner to insonify the rat kidney. Reproduced with permission from (Miller et al., 2008) (b) Glomerular capillary hemorrhage vs PRP for a range of transmitted ultrasound frequencies using a diagnostic ultrasound scanner to insonify the rat kidney. Reproduced with permission from (Miller et al., 2008) (c) Survey of transmission frequencies and PRP observed in literature. Asterisk indicates multiple studies using same parameter. a: (Lawrie et al., 2000), b: (Stieger et al., 2007), c: (Endoh et al., 2002), d: (Miller and Gies 1998), e: (Chen et al., 2002), f*: (Shohet et al., 2000; Vannan et al., 2002; Bekeredjian et al., 2003; Bekeredjian et al., 2005; Guo et al., 2004), g*: (Frenkel et al., 2002; Chen et al., 2003), h: (Miller et al., 2005), i: (van Der Wouw et al., 2000), j: (Li et al., 2004), k: (Li et al., 2003), l: (Chapman et al., 2005), m: (Pislaru et al., 2003), n: (Kobayashi et al., 2003), o: (Christiansen et al., 2003), p*: (Kobayashi et al., 2003; Kobayashi et al., 2002), q: (Wible et al., 2002), r: (Ay et al., 2001), s: (Teupe et al., 2002), t: (Stieger et al., 2007), u: (Skyba et al., 1998), v: (Miller and Gies, 2000), w: (Miller and Quddus, 2000), x: (Wible et al., 2002)

Figure 10

Effect of burst length and contrast agent dose on threshold for biological effects. (a) Effect of burst length on the threshold for blood brain barrier (BBB) disruption. Reproduced with permission from (Mcdannold et al., 2008). (b) Occurrence of petechial hemorrhage as a function of Optison dosage in the mouse intestine and muscle. Reproduced with permission from (Miller and Quddus, 2000).