This is the simplest method for investigating flow patterns. The flow is seeded with a material which ideally is neutrally buoyant and can be visualised. Materials which have been used include injected dyes and inks, hydrogen bubbles, hollow glass spheres and a variety of solid particles. The seeded material will follow the flow-streamlines allowing the observer to visualise flow patterns. A video camera can be used to record the progress of the seeded material. In Fig. 1.​18 a simple straight tube phantom with dye injection illustrated the difference between laminar flow and turbulence. Phantoms with anatomical geometry provided early evidence of complex flow patterns in bifurcations, including flow recirculation and helical flow (Ku and Giddens 1983; Zarins et al. 1983) (Fig. 12.2). Flow visualisation as its name implies is only concerned with qualitative visualisation of flow patterns, not quantitative measurement of velocities.

Quantitative measurement of the velocity field may be undertaken using particle image velocimetry (PIV). In PIV the fluid is seeded with particles at low concentration of typically less than 1 % by volume. A pulsed laser is used to illuminate a single plane within the flow. A video camera is used to record sequential images of the illuminated particles. Between images the particles move a small distance. Image processing algorithms applied to groups of particles are used to estimate the distance the particles have moved; this distance is divided by the time between images to obtain the local velocity. Figure 12.3 shows an example of PIV data taken in a stenosis phantom. A related technique is particle tracking velocimetry (PTV) in which individual particles are tracked rather than groups of particles. Further details of PIV and PTV technology are provided by Adrian (1991), Westerweel (1997) and Prasad (2000).

PIV provides high accuracy high resolution data on the velocity field. Due to its simplicity, accuracy and robustness it is regarded as providing gold-standard data, which is useful in validation of flow-field data obtained from CFD or from medical imaging.

Quantitative measurement of velocity may also be undertaken using laser Doppler anemometry (LDA). Two laser beams of the same frequency overlap producing a set of interference fringes within the overlap region. Particles moving through the interference fringes scatter light. The scattered light is detected by a photodetector. The frequency of the scattered light is dependent on the velocity of the particle and also on the angle between the direction of motion and the laser interference wave direction. This is very similar to Doppler ultrasound in that the motion of a particle produces a Doppler frequency shift. Common positions for the detector are on the opposite side to the laser or on the same side. Use of a single detector enables measurement of one velocity component whereas use of multiple detectors allows measurement of two or three velocity components. Information is gathered at a single point, however 3D information may be obtained by scanning the interrogation region through the flow field. Information is gathered continuously so that LDA may be used to study high frequency changes in the flow field associated with turbulence. Further details of LDA technology are provided by Tropea (1995); Liepsch et al. (1998) reviews the use of LDA in carotid artery phantoms.

Materials which are optically transparent and which have been used for phantom manufacture include glass, acrylic, polyester and silicone elastomers. The use of glass is challenging requiring phantom manufacture using glass-blowing. Most optically transparent phantoms with complex vascular geometries are manufactured using polyester, acrylic or silicone elastomers using a casting technique involving pouring and setting of the liquid. A simple straight tube model may be manufactured using metal rods which are joined together. These rods are positioned in a box and the liquid is poured into the box, allowed to set and the rods removed (Fig. 12.4).

3D cardiovascular geometries such as bifurcations require a more complicated manufacturing approach. Phantom manufacture is increasingly dominated by recent developments in rapid-prototyping (3D printing). Figure 12.5 summarises the steps which have been taken in published studies and are described below.

Other methods have been used to acquire 3D geometries and create phantoms:

Once the phantom has been produced, the blood mimic must be designed to match the refractive index of the material of the phantom (Nguyen et al. 2004; Miller et al. 2006; Yousif et al. 2011) otherwise there will be visualisation artefacts (Fig. 12.8) and velocity measurement errors.

Flow phantoms for medical imaging should mimic key features of the cardiovascular system which, as discussed above will depend on the application. Features might include cardiovascular geometry, blood viscous behaviour, wall stiffness and wall motion. In addition flow phantoms should mimic key properties of tissue relevant to the imaging modality. This ensures that images produced from the phantom are a good representation of images from the cardiovascular system in vivo. Tissue-mimicking also enables conclusions (e.g. on quantification of measurement errors) to be applicable in vivo. Construction of flow phantoms for medical imaging therefore requires choice or formulation of materials with the correct physical properties.

Materials which mimic the relevant imaging properties of tissues are referred to as being ‘tissue equivalent’. The relevant imaging properties are summarised in Table 12.1.

The most commonly used phantoms consist of a length of tubing bought from a laboratory supplier. Stiff tubing includes acrylic, glass, PTFE and polypropylene, whilst softer tubing is mostly based on latex rubber. These simple vessel phantoms may be used to produce idealised flow to validate MRI velocity measurement methods; with the implication that the MR properties of the tube are unimportant as they have little effect on the measurement of velocity. Simple tube models have been used in ultrasound, however the tube has a large effect due to refraction and attenuation of the ultrasound beam within the vessel wall. C-flex tubing (Cole Parmer, Vernon Hills, IL, USA) has been used as its speed of sound (1556 m s⁻¹) is close to the standard speed of 1540 m s⁻¹ (Hoskins 2008).

Off-the-shelf materials can also be used to construct medical imaging phantoms with more complicated geometries. For CT and PET solid materials such as acrylic provide similar X-ray and gamma-ray attenuation characteristics to human tissue. This is convenient as phantoms can easily be constructed by milling or by casting. For MRI solid materials such as polyester, acrylic or silicone have been used to construct phantoms (e.g. Smith et al. 1999). Essentially these phantoms are identical to those used for optical imaging (Sect. 12.1). Figure 12.9 shows bifurcation phantoms and the associated MRI images of flow.

Solid materials such as polyester do not return an MRI signal which leads to unrealistic image appearance (Smith et al. 1999). This is an issue if the effect of surrounding tissue on velocity measurement is of interest. Solid materials such as acrylic have high speed of sound compared to soft tissue and are unsuitable for use in ultrasound phantoms. Ultrasound phantoms are available commercially based on urethane as the acoustic properties provide a reasonable match for soft tissue (Table 12.2).