Homemade Ultrasound Phantom - Essay Example

CHAPTER ONE: INTRODUCTION This assignment was to design, construct and test a homemade ultrasound (US) phantom. Its purpose essentially is to assist in enhancing the perception of some basic principles of diagnostic US. Additionally, it also permits greater understanding of their relevance and application to practice. Aims and Objectives The predominant aims were: Build a US phantom which will permit demonstration of some principle physics and thus allowing a greater understanding of their importance on clinical practice. Explore inter-observer variability when measuring an ‘bone artifact’ within the phantom. These aims will be investigated by Undertaking an extensive literature review which will assist in justification of the projects conclusion and rationale for the project itself. Designing, building, and testing a phantom. Conducting a small audit within a US department of caliper measurements between three operators (inter-observer variability). CHAPTER TWO: LITERATURE REVIEW INTRODUCTION The utilization of diagnostic US in medical imaging is well established and one which is continually developing (Colquhoun et al 2005). A plethora of authors throughout the literature have recognized that for diagnostic US to be used effectively the operator must possess (among other skills) an appreciation of the basic principles of US imaging (Hoskins et al 2003, Langer and Kofler 1997). There are numerous physical principles of US to consider however, there are some basic principles that US assumes. These include the assumption that sound waves at medical imaging take the longitudinal compression wave from (Dendy and Heaton 1999). These sound waves propagate along negligible beam widths at a constant speed (Langer and Kofler 1997). Returning echoes originate only from the objects located in the beam axis (Bushong 1999) and the amplitude of the received echo is governed by the nature of the reflector and its acoustic impedance (Hoskins et al 2003). Violations and imperfections can occur causing artifacts and limitations on the resultant US image. In order to minimize artifacts and distinguish them from the actual diagnostic information it is imperative that the operator gains hands on experience to improve clinical scanning techniques and interpretation (Langer and Kofler 1997). Phantom can be means of achieving this. Additionally basic physical principles and imperfections need to be considered when constructing the phantom itself. Phantoms Gent (1997) characterizes an US phantom as a device that permits qualification and testing of the various aspects of an US system performance and transducer. Furthermore phantoms are able to demonstrate limitations and artifacts of the equipment (Gent 1997). Test phantoms play a fundamental role in quality control and performance testing. Browne et al (2002) suggest phantoms should be tissue mimicking however, the literature provides a number of phantom types. These include fluid filled phantoms, thin film test objects with digital scatters (Phillips and Parker 1998) and digital image analysis converters (Zdero et al 2002). It is beyond the scope of this assignment to discuss the merits and drawbacks of each individual test object therefore expansion on the most commonly used phantom- tissue equivalent/tissue mimicking phantoms (Browne et al 2002) will be discussed. This is the type of phantom which will attempt to be constructed. Tissue mimicking phantoms should possess certain characteristics namely; 1) Similar scattering properties and echogenicity often compared to that of the liver. 2) Equivalent attenuation co-efficient to soft tissue (Gent, 1997) 3) Comparative acoustic velocity to the speed of sound in soft tissues, 1540m/s (Browne et al 2002). Numerous mediums can be utilized in order to achieve such properties which are often referred to as tissue mimicking materials. Tissue-Mimicking Materials Traditionally TMM’s of agar gel combined with graphite has been utilized as it possesses a speed of sound of 1540m/s and similar attenuation coefficients to soft tissues (Goldstein, 2000 and Dudley et al 2002). However graphite/agar TMM’s tend to have a limit to its productive life as the gel tends to deteriorate. To overcome this problem other materials have been tested namely evaporated milk and urethane rubber. The study by Goldstein (200) suggests that although urethane rubber phantoms have along useful life (in comparison to agar gel) their acoustic velocities are much lower than 1540m/s and concludes that such phantoms should not be used to check distance measurement accuracies of transducers or focusing performance in a clinical setting. On the other hand Dudely et al (2002) highlight the point that all TMM’s cannot be reliable when utilized to predict clinical focusing performance because real tissues are inhomogeneous and not truly represented. Hoskins et al (2003) supports this by reminding us that the soft tissue velocities vary around the mean value of 1540m/s approximately five percent which can lead to distortions in measurements of an image. A number of reasons for electing a TMM will come into play and may involve availability, durability, and expense. Whichever TMM is adopted the tissue mimicking phantom will still aim to test certain parameters of the systems performance. Some of these include sensitivity, dead space, resolution, and calibration (Gent 1997). This assignment will focus on predominantly on calibration aspects as this relates to the phantom under construction. Calibration Calibration is defined as the accuracy of the electronic calipers utilized to take measurements from the US display (Gent 1997). Measurements are frequently used within medical US especially in the obstetrics field where earliest US measurements were recorded (Hoskins et al 2003). Caliper use can range from simple linear measurements to circumferential or volumetric methods. Two dimensions of calibration can be assessed, vertical calibration, and horizontal calibration. Vertical calibration is measurement perpendicular to the beams axis (Goodsitt et al 1998). Vertical calibration error is less likely to occur than horizontal as it occurs as a result of internal timing circuits of the system. Horizontal errors however occur as a result of transducer flaws. Errors of greater than two percent are considered unacceptable (Goodsitt et al 1998) Accuracy measurements can be affected by a plethora of factors one to consider is that of image pixel size. As the smallest distance represented in an image is one pixel any measurement will essentially have an uncertainty value of plus or minus one pixel. This in turn will affect caliper increments as it can never be smaller than the size of one pixel, moving one pixel at a time. To aid in overcoming this problem objects to be measured should be magnified to reduce pixel size to its smallest thus improving accuracy (Hoskins et al 2003). Resolution Accuracy measurements are also dependent on image resolution. US spatial image resolution is finite and consists of axial and lateral resolution. Axial resolution is described as the smallest distance between two points which can be demonstrated separately in the perpendicular plane to the beam axis (Gent 1997). Lateral resolution is generally considered inferior to axial resolution which relates back to the fact that US systems assume that all received echoes for any beam position arise from its central axis. Therefore separate structures within the beam width at the same instant will be resolved as only one echo whereas structures further apart than the beam width will display as two separate echoes (Gent 1997). Errors are likely to be of the lateral resolution type (Hoskins et al 2003). Machine settings can aid in minimizing errors pertaining to lateral beam width, in particular focusing of the beam (Gent 1997). Beam Forecasting To appreciate focusing of the beam it should be recognized that an US beam is divided into two regions, the near field zone (the zone closest to the transducer face) and the far field zone (the zone furthest from the transducer face). It is at the near field/far field transition zone where spatial resolution is optimum. This is described as the focal zone (Bushong. 1999). To improve the spatial resolution a narrow beam width is necessary and is achieved by focusing (Gent 1997). Focusing on the focal zone can greatly enhance lateral resolution (Gent 1997) and can be employed in both transmission of the beam and when detecting the returning echoes (Chudleigh and Thilaganthan, 2004). Essentially there are two methods of achieving transmission focusing; a concave transmitting source (crystal) or via an acoustic lens (Hoskins et al 2003). As Chudleigh and Thilaganathan (2004) reminds us, the effect of narrowing the beam at a selected depth results in increased divergence of the beam at other depths resulting in degradation beyond the focus. This “trade off” can be reduced with the utilization of multiple focal zones (Gent 1997). Multiple pulses are transmitted which are focused at different depths and the echoes from the non-focused depths are “rejected” (Chudeligh and Thilaganathan 2004). Each pulse has data from its specific focused depth which is superimposed to compile overall improved resolution (Chudeligh and Thilaganathan 2004). Unfortunately this additionally comes with a trade off, which is that of reduced frame rate (Gent 1997). Reduced frame rate in certain situations such as patient movement for example, may not be acceptable (Gent, 1997). It is the operator who controls the transmission focus and this should be placed at the depth of the object or area of interest. Predominantly an arrowhead alongside the image indicates the point at which the transmission focus is set (Hoskins et al 2003). For example in a clinical setting an arrowhead (focus) would be placed at the level of the fetal femur when measuring femur length in obstetric scanning. Acoustic Impedance US propagation velocities can also affect the accuracy of measurements in US systems. Within medical imaging sound waves travel through any particular medium at a speed which is determined by the medium itself (Hoskins et al 2003). The speed of travel of an US wave is frequently referred to as the velocity (Gent, 1997). Each medium will have acoustic impedance which is a measure of how the particles in that medium respond to a wave of given pressure (Hoskins et al 2003). The acoustic impedance of a medium is determined by its compressibility and density, therefore the speed of the sound wave’s propagation will additionally be determined by the mediums compressibility and density (Gent 1997). As stated previously US systems are predominantly designed to assume an acoustic velocity of 1540m/s (Browne et al 2003). Human tissues vary around this mean that value by approximately five percent which can lead to slight distortions in the US image. Langer and Kofler (1997) provide an example of this by explaining that if the speed of sound travels at 1450m/s through fat compared to 1540mm/s through soft tissues. An object distal to a soft tissue/fat interface will be displayed at an exaggerated distance in a machine which is calibrated to 1450m/s. Attenuation As an US wave propagates through a medium (tissue) it is subject to various interactions which reduce its amplitude and intensity (Dendy and Heaton, 1999). This effect is known as attenuation. Bushong (1999), states that the degree of attenuation of any particular homogeneous tissue is expressed by the attenuation coefficient (α), expressed in dBcm-1. Attenuation co-efficient varies considerably between different tissues and is dependent upon frequency. Bushong (1997) provides a calculation demonstrating that as frequency increases attenuation co-efficient increases proportionally (See Appendix A). Several interactions are responsible for attenuation; predominant factors are reflection, scattering, absorption, and divergence (Gent 1997). As this project incorporates the use of bones, the process of reflection will be discussed in further details. When an US beam meets an interface where there is acoustic impedance mismatch a proportion of the incident energy will be reflected and return as an echo to the transducer and some will be transmitted into the second medium (Gent, 1997). Specular reflection is said to occur at smooth soft tissue interfaces (Bushong, 1999) and requires the US beam. As bone is a high level reflector strong echoes are produced and should appear bright on the resultant US image (Chudleigh and Thilaganathan, 2004). This is apparent in obstetrics with calcification of fetal limbs. Therefore the bones within the phantom to be constructed should appear as bright white echoes on the display. Artifacts Other distortions occur when the basic assumptions of US imaging are violated and are commonly referred to as artifacts. Langer and Kofler (1997) further define an US artifact as any feature which is not related to a real structure. There is a whole range of artifacts which can arise and incorporate mirroring, shadowing, enhancement, ring down, and refraction (Gent 1997). Refraction is when the US wave is obliquely incident on an interface where there is a difference in the speed of sound between tissues (Hoskins e al 2003). A beam that has been refracted will still be displayed on the image as if it has derived from an undeviated axis. The extent of this will depend on the degree of the refraction angle known as Snell’s Law (Appendix B) and the speed between different mediums (Gent 1997). Hoskins et al (2003) highlight that refraction should be considered particularly in the obstetric field as objects can appear distorted due to refraction around the maternal mid-line. Human Error Human error is further source of inaccuracy to be considered in US measurements. Factors which contribute to such errors might be inadequate training, lack of training, and unspecified standards and protocols. All of which may lead to inaccurate placing of calipers (Hoskins et al 2003). Inter and intra observer variability is a commonly studied area especially pertaining to the obstetric field. Walker et al (1993) also raise the issue that any visual tests such as caliper accuracy are performed by the eye and are therefore subjective. Although sensitive to subtle changes the eye is still limited by the iris itself which will be different in each individual. Conclusion US are a fundamental tool within the obstetric field and accounts for thirty five percent of reported scans (Duck, 2005). Furthermore Dudley and Chapman (2002) advice that almost all obstetric patient in the United Kingdom will have at least one routine scan. Accurate measurements are therefore needed in fetal biometry where accuracy and reproducibility are crucial (Dudley and Griffith 1996). Such measurements may be utilized in the interpretation of further investigation such as a baseline for subsequent growth assessment (Dudley and Chapman 2002). Implications of inaccuracy can consist of inappropriate and untimely intervention, perinatal compromise and most significantly perinatal death (Dudley and Chapman 2002). CHAPTER THREE: METHODOLOGY Ethical Approval The inter-observer variability section of the project was classified as a small departmental audit therefore, ethical approval was not necessary. Agreement however was sought from the departmental superintendent. Participants were voluntary and were advised they could pull out of the audit at any time. Confidentiality and anonymity were also ensured. METHODOLOGY INTRODUCTION Initially an extensive literature search was conducted that provides a foundation on which subsequent new knowledge can be based (Polit et al 2001). Search engines such as BNI, Cinhal, Amed, and Embase were utilized to achieve this. Numerous journals, books, and ultrasound physics, caliper measurements in US and femur length studies. The study approach incorporates both qualitative and quantitative aspects. The qualitative aspects will be utilized with the author’s scrutiny of the phantoms performance and artifacts which may be displayed. Quantitative aspects involve the inter-observer variability. As outlined within the literature review the phantom acoustic velocity should be as close to 1450ms-1 as possible (Browne et al 2003). A gelatin powder was utilized as a TMM as its use has discussed in various studies (Gane, 2000, Mokhtari-Dizaji, 2001). Furthermore it is readily available and low cost (Mokhtari-Dizaji, 2001). The literature additionally advises that phantoms should possess similar attenuation properties and echogenicity patterns to that of liver tissue (Dudley et al 2002). Gane (2002) describes the use of graphite powder to achieve this however; she highlights the issue of its inefficiency and difficulty in use. It was therefore decided to omit this property from the phantom and this will need consideration when discussing the final results. Chicken bones were selected for measuring purposes and were thought to be comparable to femur lengths measured in obstetric scanning. Shaw et al (1999) claim that different types of bones exhibit significant different properties which have not been universally agreed upon. A gap in the literature prevented further scrutiny of this subject. Constructing the Control Model Gane (2002) advises that normal usage of gelatin according to manufacturer’s instructions resulted in a mix which was not sufficiently durable. A mix using a 3:1 gelatin to water ratio was used (three sachets of gelatin to one print of hot water). This was mixed well until all particles dissolved. A solitary chicken bone was placed into the mix and stored in a refrigerator to set. Shaw et al (1999) points out that the temperature at which the gel sets can affect its attenuation co-efficient and suggests refrigeration of the mix assists in minimizing this variation. Once set the control mix was scanned using a Toshiba Apilio imaging system with a curvilinear 5MHz probe on the obstetric setting. Coupling gel was applied to provide an air free interface between the transducer and the phantom. The transducer was placed in direct contact with the exposed surface of the gelatin. Gain was set to mid-range and focus set to just below the level of the chicken bone. This ensures the transmission of the beam is as narrow as possible, concentrating the power into a focal zone which optimizes lateral resolution (Hoskins et al 2003). All equipment and settings were chosen to be as comparable as possible to those used in a clinical obstetric setting. To establish the acoustic velocity of the phantom a calculation provided by Goldstein (2000) was used. Crs= Dim Drs * Ccal Where Crs = Actual velocity of phantom Dim=distance of measured image Drs=actual distance Ccal=1540-1 When re-arranged the control mix velocity calculation is: Speed of sound in Phantom=70mm ×1540 70mm = 1540ms-1 It was observed that the mix set well and maintained durability when scanned. The chicken bone was visible as an echogenic object and was compared to obstetric femur images. It was considered moderately comparable although, longer than fetal femur lengths taken at a routine twenty week scan. Phantom- Main Study The phantom medium was constructed as per the control model using the same gelatin mix ratio and refrigerated. Three chicken bones were washed and measured and it was attempted to obtain lengths as similar as possible. The length of the container was divided into three equal portions, marked on the outside with a permanent marker pen. Vertical markers were also drawn onto the containers outer surface. Using a protractor three different angles was set. One bone at 900 to the beam axis, one at 450 and the other one at 150. Holes were pierced at the indicated markers using a sharp implement. Fishing wire was tied around the ends of each chicken bone which was then thread through the pierced holes using a sewing needle. This was secured using waterproof tape and then super glued to prevent any leakage of the mixture. The agar mix was then poured into the container to a depth of six centimeters and allowed to set. (See appendix C for images of the phantom) Scanning the Phantom The phantom was scanned using a curvilinear 5MHz probe on a Toshiba Aplio again with the use of a coupling gel. The gain was set to mid-range and the focus was set to the level of each bone being imaged. These were manipulated by the author only and not the operators as to decrease other factors of variability. Three operators scanned the phantom and were all blinded to the actual measurements and also the measured measurements displayed on the screen. This was to assist in limiting any bias into the audit. Each bone was measured a total of three times each to assist in a representative average (Dudley and Chapman 2002). Finally the phantom was scanned by the author to demonstrate incidental artifacts displayed within the phantom. CHAPTER FOUR RESULTS Qualitative Results The author observed the phantom as it was being scanned by the operators and additionally scanned the phantom themselves to demonstrate any artifacts present. The following observation were noted The phantom demonstrated an inhomogeneous echo texture with areas of hyperechogenicity. See below. As graphite powder was not used in the construction of the phantom a typical “liver like” echogenicity was not demonstrated. Chicken bones appeared different to scanning obstetric femur bones. They did not appear to demonstrate the bowed ends of a femur which represents the diaphysis. There was no shadowing effect of the bone. The bones did appear highly reflective and echogenic allowing measurements to be taken Artifacts were present. Reverberation of the plastic container was apparent; comet tail artifacts from the fishing wire and mirroring effects were also displayed. (See Appendix D) Quantitative Results Initially the acoustic velocity of the phantom was calculated. Speed of Sound in Phantom= 60mm×1540 57mm = 1621ms-1 As the speed of the sound in the phantom is greater than 1540ms-1 the echoes will arrive back at the transducer earlier than expected and therefore will display an object nearer to the transducer than it really is (Hoskins et al 2003). Secondly the measurements taken for each bone from each operator were taken and the mean average value was calculated independently. The mean average value and percentage error was then calculated and these results can be seen in Appendix E. It was envisaged that all measurements would then be subject to various statistical calculations to allow further discussion of inter/intra-observer variability. Unfortunately due to time constraints and limits of the word count these were not performed. Statistical analysis most likely appropriate for such evaluation may be intra class correlation co-efficient, interclass correlation co-efficient and the reliability co-efficient as utilized in studies of Perni et al (2004) and Rosati et al (2004). CHAPTER FIVE DISCUSSION AND LIMITATIONS Phantom Performance The inhomogeneous echo texture of the phantom was difficult to explain however it may be as a result of the phantoms attenuation co-efficient. As detailed within the methodology section the inclusion of graphite or similar substance was omitted from the construction. Materials such as graphite are said to mimic the attenuation properties and echogenicity patterns to that of the liver (Dudley et al 2002). This therefore may have affected both the measurements obtained and the echo texture pattern displayed on the image. Furthermore Shaw et al (1999) suggest that the temperature at which the gel sets can also affect its attenuation co-efficient. Although the phantom mix was left to set the refrigerator temperature fluctuations are still possible and therefore this is another factor to consider. In hindsight the depth of the phantom was not sufficient and a deeper model would be more appropriate. Some of the bones within the phantom were very superficial and although this can occasionally occur in a clinical obstetric setting, bones would have been more appropriately placed approximately five centimeters deep. This would provide a closer resemblance to a maternal abdomen. If time permitted the phantom would be re-designed to be much deeper. With regards to the chicken bones it is impossible to establish if they possess similar acoustic properties to fetal femur bones as there are no definite guidelines to compare these. In fact Shaw et al, (1999) claim that different bones exhibit significant different properties. It was observed however that they did not possess the acoustic shadowing properties possessed by fetal femur bones when scanned. On the other hand they did demonstrate reflection properties similar to fetal femurs and displayed on the US monitor as bright white echogenic objects (Chudleigh and Thilaganathan, 2004). (See Appendix F) Measurements The measurements of the bones did vary considerably not only from those taken by one operator but also amongst different operators. Both bones A and B achieved a percentage error of greator than three percent which is deemed beyond the normal limits (+/- 2% IPSM). Unexplainably bone C achieved a percentage error of zero which would be acceptable. Various reasons for inaccuracy of the measurements taken will now be considered. As the acoustic velocity of the phantom was higher than the 1540ms-1 assumed by US machines (Gane, 2002), it was thought that all measurements taken from the image would be shorter than the actual bone measurements. This in reality did not occur. Bone A and Bone C measured at an increased value whereas Bone B measured at a decreased value. This was unexplainable although one factor which may have affected this is the temperature of the phantom itself. The measurements were taken over the course of a few hours and the phantom temperature and room temperature may have fluctuated. Browne et al (2003) suggests that temperature affects the TMMs acoustic properties. They suggest phantom should be tested at room temperature and additionally suggest large changes in attenuation co-efficient due to temperature can dramatically affect quality control and performance test results of the US machines. As the acoustic velocity of the phantom was different to the 1540ms-1 assumed by US machines (1621ms-1), measurements values may have been skewed. Goldstein (2000) and Browne et al (2003) highlight the fact that when acoustic velocities of TMMs differ significantly from the calibrated acoustic velocity of US machines the effect on distance measurement and lateral resolution will occur. Finally US is a modality which is considered significantly operator dependent (Dendy and Heaton, 1999). Measurements will always be dependent upon the eye of the operators, the operators training and departmental protocols, for example (Hoskins et al 2003). Often measurements are different in real clinical settings and Chudleigh and Thilaganathan (2004) suggest measuring a femur three times from three different images, this should aid in achieving a more accurate result (Dudley and Chapman 2002). This was adopted within the study however measurements still varied considerably between operators. CHAPTER SIX CONNCLUSION This assignment permitted construction of an US phantom which successfully demonstrated some basic physical principles of US. The chicken bones utilized were moderately comparable to fetal femurs and permitted various operators to perform measurements of them. Unfortunately due to time constraints and other limiting factors full exploration of inter/intra observer variability was not achieved and this an area to further explore in the future. In constructing the phantom and performing the literature review some physical principles have become evident and as Langler and Kofler (1997) suggest, it is imperative that operators gain hands on experience to improve clinical scanning techniques and interpretation. Personally the author’s knowledge, understanding, and appreciation of US physics are significantly increased and this will no doubt benefit their clinical scanning abilities. References Browne et al 2003. Assessment of the Acoustic Properties of Common Tissue-Mimicking Test Phantoms [online]. Ultrasound in Medicine and Biology, 29 (7), pp 1053-10060. Available from http://www.sciencedirect.com. Accessed on 14th March 2006. Bushong, S., 1999. Diagnostic Ultrasound. USA. The McGraw-Hill Companies Inc. Chudleigh, T., and Thilaganathan, B. 2004. Obstetric Ultrasound – How, Why, and When. Third Ed. London. Elsevier. Colquhoun et al. 2005. (iv) Basic Science: ultrasound. Current Orthopaedics [online] 2005 (19) pp 27-33. Available from http://www.sciencedirect.com Accessed on 14th April 2006. Dendy, P., and Heaton, B., 1999. Physics for Diagnostic Radiology. Second ed. Croydon. Institute of Physics Publishing. Duck, F. 2005. Ultrasound exposure measurement: a hidden science. The British Journal of Radiology. 78 (2005), pp 289-291. Dudley, N., and Chapman, E. 2002. The importance of quality management in fetal measurement. Ultrasound in Obstetrics and Gynecology. 19, pp 190-196. Dudley, N and Griffith, K. 1996. The importance of rigorous testing of circumference measuring calipers. Ultrasound in Medicine and Biology. 22 (8), pp1117-1119. Gane, A. 2002. Constructing a Homemade Ultrasound phantom-the problems encountered. BMUS Bulletin August 2002. 10 (3), pp 33-36. Gent, R., 1997. Applied Physics and Technology of Diagnostic Ultrasound. Australia. Open Press Publishers. Goldstein, A. 2000. The effect of acoustic velocity on phantom measurements [online]. Ultrasound in Medicine and Biology. 26 (7), pp 1133-1143. Available from http://www.sciencedirect.com. Accessed on 22nd May 2006. Goodsitt et al 1998. Real-time B-mode ultrasound quality control test procedures. Report of AAPM Ultrasound Task Group No 1. Med Physics 25 (8), pp 1385-1401. Hoskins et al (2003). Diagnostic Ultrasound-Physics and Equipment. London Greenwich Medical Media Ltd. Langer, S., and Kofler, J. 1997. A series of Teaching Phantoms for Displaying Diagnostic Ultrasound Image Artifacts. Journal of Diagnostic Medicall Sonography. 13 (1), pp 22-27. Longo et al 2004. Femur and Humerus Length in trisomy 21 fetuses at 11-14 weeks of gestation [online]. Ultrasound in Obstetric and Gynaecology, 23 (2), pp 143-147. Available from http://www3.interscience.wiley.com/cgi-bin/fulltext/107567899/HTMLSTART. accessed on 11th March 2006. Manijhe, M-D. 2001. Tissue mimicking materials for teaching sonographers and evaluation of their specifications after three years [online]. Ultrasound in Medicine and Biology 27(12), pp1713-1716. Available from http://www.sciencedirect.com. Accessed on 14th April 2006. Perni et al 2004. Intra observer and inter observer reproducibility of fetal biometry [online]. Ultrasound in Obstetrics and Gynaecology, 24 (6), pp 654-658. Available from http://www3.interscience.wiley.com/cgi-bin/fulltext/109675879/HTMLSTART. Accessed on 11th March 2006. Philips, D. and Parker, K. 1998. A new imaging science test object for performance measurements of ultrasonic imaging systems [online]. Phys. Med Biology, 43 (1998) pp455-465. Available from http://www.sciencedirect.com. Accessed on 14th March 2006. Rosati et al 2004. Intra and onterobsever repeatability of femur length measurement in early pregnancy [online]. Ultrasound in obstetrics and Gynaecology. 23 (6), pp 599-601. Available from http://www3.interscience.wiley.com/cgi-bin/fulltext/10856319/HTMLSTART. Accessed on 11th March 2006. List of Appendices Appendix A: Attention Co-efficient Calculation As frequency increases attenuation co-efficient increases proportionally Atenuation (dB)= Frequency (MHz) ×attenuation co-efficient (dB/cm/MHz)×tissue path (cm). Appendix B: Snell’s Law Sin =c1 Sin c2 Appendix C: Appendix D: Laser images of the phantom artifacts Appendix E: Results of the mean average measurements for each bone Appendix F: Thermal images demonstrating bone artifacts Read More