Presupuesto

ultrasound system itself, such as the quality of the display monitor and its display settings (Sehgal et al., 2006). The system’s hardware include

transducer types that use variable frequencies, bandwidths, focal distances and aperture that can change the output display of the ultrasound image (Sehgal et al., 2006).

Ultrasonic Parameters for Optimal QUS Imaging

QUS data is based on the digitized radiofrequency signal from tissue backscatter. The practical challenges of optimal QUS imaging arises in terms of achieving a desirable resolution in both the lateral and axial direction of the ultrasound image and this is dependent on the ultrasound parameters used for imaging. In order to attain useful QUS data and information, QUS imaging parameters must be capable of resolving cellular and subcellular structures. For this, the optimal lateral and axial resolution are dependent on the following parameters: wavelength (𝜆), f number (ƒ₂₃₄₅₆₇), acoustic frequency bandwidth (B), and the speed of sound (𝑐). These parameters are described below in terms of its relationship to achieving the desired lateral and axial resolution in an ultrasound image for useful QUS analysis.

The lateral resolution is spatially perpendicular to the beam axis (O'Brien, 2007) and permits imaging objects that are positioned side-by-side. The lateral resolution is defined as the minimum resolvable distance to differentiate or contrast two adjacent reflectors or structures (O'Brien, 2007). The lateral resolution is high when the beam width is at its narrowest due to a small distance between acoustic scan lines. The lateral resolution is dependent on the frequency (and thus wavelength, l) and the geometric characteristic of the acoustic focus, such as the diameter of the transducer (D) and the focal depth (F). The optimal parameters for optimal imaging include the following

parameters:

i.Shorter wavelength (l); ii.High frequency (f) iii.Large diameter (D)

At distances beyond the focal zone (i.e. the far zone), the lateral resolution deteriorates as the beam width diverges (i.e. wider) beyond this point. Thus, optimal lateral-resolution is achieved within the near zone and up to the focal zone (Figure 1.20).

Figure 1.20: Lateral resolution is best when the beam width is narrow; allowing for distinguishing side-by-side structures. The lateral resolution is highest within the focal zone at the focal point (;₂); where the beam converges, and is

narrowest.

The lateral and axial resolution are important in imaging locally advanced breast cancer tumours due to the size of the tumour which can span several cm across the breast (>5 cm). The axial resolution is defined as the distance of one wavelength (l) along the axis of the ultrasound beam. Thus, an object can be resolved that is equal to the distance occupied from one cycle or pulse of ultrasound (O'Brien, 2007). In practical US imaging, single wavelengths are not used rather, pulses of ultrasound are employed that contain N wavelengths per pulse; this is termed the spatial pulse length (SPL). The SPL can be used to calculate the axial resolution based on the frequency bandwidth (Figure 1.21).

Figure 1.21: Axial resolution for ultrasound imaging. The axial resolution is determined by the wavelength of an ultrasound pulse (A). However, in practical ultrasound imaging, a pulse of ultrasound will contain several wavelengths as it propagates through the imaging medium (e.g. tissue). The spatial pulse length (SPL) is defined as the number of wavelengths in a repeated pulse waveform. For practical ultrasound imaging, the SPL is used to determine the optimal (best) axial resolution. In this example in (B), the SPL is equal to 2l.

For QUS quantification of biological tissue which can contain several scatterers, a Born approximation is assumed; that is, the backscattered field is the sum of the individual scatterers within that acoustic field (Chivers, 1977). Thus, QUS measurements in breast tumours represent the net change in scatterers (i.e. dying cells). It is also important to mention that in QUS modelling, the backscatter intensity calculations assume that the tumour is a low-density medium and that scatterers within the tumour microenvironment are randomly distributed (Oelze et al., 2002). Previous experiments have developed a framework to show the relationship between QUS parameters and scatterer properties (outlined below in Table 1.17) (Lizzi et al., 1997b, Insana and Hall, 1990, Kolios et al., 2002, Feleppa et al., 1986).

Scatterer Property QUS Parameters Findings Reference Size Spectral Intercept

A two-fold increase in the scatterer diameter showed an 18 dB increase in the Spectral Intercept. (Feleppa et al., 1986) Size Spectral slope

The size of the scatterer of less than ~20 µm has insignificant effects on the spectral slope (i.e. slope remains constant).

(Feleppa et al., 1986)

Size Mid-band Fit The mid-band fit increases from an increase in the scatterer diameter for effective diameters of up to ~60 µm (Feleppa et al., 1986). (Lizzi et al., 1997b) Number of scatterers Spectral Intercept

An increase in the number of scatterers increases the spectral intercept due to an increase in the number of scattering surfaces (Lizzi et al., 1997b) Concentration of scatterers Spectral Slope An increase in the concentration of the scatterers decreases the spectral slope

(Lizzi et al., 1997b)

Table 1.17: QUS dependence on scatterer properties. Scatterer properties such as the size, number and distribution have been shown previously to change the acoustic scattering in tissue

In general, the quality of US imaging is reliant on angle dependence, aliasing (indeterminate ultrasound signals) and a poor signal-to-noise ratio that can result from user error (Hamper et al., 1997). Other technical considerations for US include breast density and breast-tissue composition (connective tissue, lactiferous ducts) that can affect image quality since it can alter the speed-of- sound in tissue, cause speckle, and scattering; these factors may result in image artefacts (Sehgal et al., 2006). For elastography, tissue composition can affect the elasticity reading between patients that demonstrate higher fatty- tissue content and fibrosis, as these features can change the biomechanical properties of the breast and alter the strain measurements (Butcher et al., 2009, Wells and Liang, 2011). US imaging can also be limited by penetrance, also known as the acoustic impedance. This is because US imaging is dependent on the wavelength and frequency of the acoustic wave. Variations in the ultrasound frequency may limit the quality of images if tumours are situated deeper into the breast tissue (Athanasiou et al., 2009). Conventional breast sonography uses frequencies between 9-12 MHz, and this allows imaging axial distances of up to 5 cm (Athanasiou et al., 2009). As a rule, higher frequency (>20 MHz) ultrasound can improve resolution but due to high scattering in tissue, these frequencies are better suited for superficial lesions. Conversely, lower frequency ultrasound (2-5 MHz) can penetrate greater tissue depths, but resolution is compromised (Athanasiou et al., 2009). Previous US clinical studies have employed transducers that operated at between 5-13 MHz (Gangeh et al., 2016, Sadeghi-Naini et al., 2013b, Sannachi et al., 2015, Tadayyon et al., 2016, Tadayyon et al., 2017, Amioka et al., 2016, Shia et al., 2015). These frequencies are consistent with clinical breast sonography but the variations in acoustic parameters (i.e. frequency, time-gain compensation, pulse repetition frequency and focal depth) limit the comparisons that can be made between studies since the experimental conditions (i.e. technology, biological measurements, image processing techniques) were different.

Functional imaging techniques such as contrast-based ultrasound (CEUS) also have limitations. For CEUS, invasive injections are required and there are potential adverse reactions from using contrast agents (Stewart and Sidhu, 2006). Also, the contrast agents’ lifetime in blood vessels are short and therefore multiple injections are needed for optimal image acquisition (Heijblom et al., 2011).

1.13.8 DOS and QUS Measurements Represent the Tumour’s Spatial and