Computed Tomography: An Overview

Godfrey Newbold Hounsfield: Godfrey Newbold Hounsfield (Fig. 1-6) was born in 1919 in Nottinghamshire, England. He studied electronics and electrical and mechanical engineering. In 1951, Hounsfield joined the staff at EMI Limited (Electric and Musical Industries [manufacturer of records and electronics, and recorded the Beatles under the EMI Label], now Thorn EMI) in Middlesex, where he began work on radar systems and later on computer technology. His research on computers led to the development of the EMIDEC 1100, the first solid-state business computer in Great Britain.

In 1967, Hounsfield was investigating pattern recognition and reconstruction techniques by using the computer. (In image processing, pattern recognition involves techniques for the observer to identify, describe, and classify various features represented in an image or a signal.) From this work, he deduced that, if an x-ray beam were passed through an object from all directions and measurements were made of all the x-ray transmission, information about the internal structures of that body could be obtained. This information would be presented to the radiologist in the form of pictures that would show 3D representations.

With encouragement from the British Department of Health and Social Security, an experimental apparatus was constructed to investigate the clinical feasibility of the technique (Fig. 1-7). The radiation used was from an americium gamma source coupled with a crystal detector. Because of the low radiation output, the apparatus took about 9 days to scan the object. The computer needed 2.5 hours to process the 28,000 measurements collected by the detector. Because this procedure was too long, various modifications were made and the gamma radiation source was replaced by a powerful x-ray tube. The results of these experiments were more accurate, but it took 1 day to produce a picture (Hounsfield, 1980).

To evaluate the usefulness of this machine, Dr. James Ambrose, a consultant radiologist at Atkinson-Morley’s Hospital, joined the study. Together, Hounsfield and Ambrose obtained readings from a specimen of human brain. The findings were encouraging in that tumor tissue was clearly differentiated from gray and white matter and controlled experiments with fresh brains from bullocks showed details such as the ventricles and pineal gland. Experiments were also done with kidney sections from pigs.

In 1971 the first clinical prototype CT brain scanner (EMI Mark 1) was installed at Atkinson-Morley’s Hospital and clinical studies were conducted under the direction of Dr. Ambrose. The processing time for the picture was reduced to about 20 minutes. Later, with the introduction of minicomputers, the processing time was reduced further to 4.5 minutes.

In 1972 the first patient was scanned by this machine. This patient was a woman with a suspected brain lesion, and the picture showed clearly in detail a dark circular cyst in the brain. From this moment on, and as more patients were scanned, the machine’s ability to distinguish the difference between normal and diseased tissue was evident (Hounsfield, 1980).

Dr. Hounsfield’s research resulted in the development of a clinically useful CT scanner for imaging the brain. For this work, Hounsfield received the McRobert Award (akin to a Nobel Prize in engineering) in 1972. During the same year, he became a Fellow of the Royal Society and subsequently was awarded the Lasker Prize in the United States and, in 1977, Dr. Hounsfield was appointed Commander of the British Empire (CBE). In 1979, Hounsfield shared the Nobel Prize in medicine and physiology with Allan MacLeod Cormack, a physics professor at Tufts University in Medford, Massachusetts, for their contributions to the development of CT. Figure 1-8 shows a note sent to the author from Dr. Hounsfield in response to several questions. After receiving this prestigious prize, he was knighted by her majesty Queen Elizabeth and became an Honorary Fellow of the Royal Academy of Engineering. Dr. Hounsfield died on August 12, 2004, at the age of 84 years (Isherwood, 2004). By developing the first practical CT scanner, Dr. Hounsfield opened up a new domain for technologists, radiologists, medical physicists, engineers, and other related scientists.

Allan MacLeod Cormack: Allan MacLeod Cormack (Fig. 1-9) was born in Johannesburg, South Africa, in 1924. He attended the University of Cape Town where he obtained a Bachelor of Science in Physics in 1944 and earned a Master of Science in Crystallography in 1945. He subsequently studied nuclear physics at Cambridge University before returning to the University of Cape Town as a physics lecturer. He later moved to the United States and was on sabbatical at Harvard University before joining the physics department at Tufts University in 1958.

Professor Cormack developed solutions to the mathematical problems in CT. Later, in 1963 and 1964, he published two papers in the Journal of Applied Physics on the subject, but they received little interest in the scientific community at that time. It was not until Hounsfield began work on the development of the first practical CT scanner that Dr. Cormack’s work was viewed as the solution to the mathematical problem in CT (Cormack, 1980). Cormack died at age 74 years, in Massachusetts on May 7, 1998. Additionally, back in South Africa, Dr. Cormack was granted the Order of Mapungubwe, South Africa’s highest honor, in December 2002, for his contribution to the invention of the CT scanner.

First 10 Years: Between 1973 and 1983 the number of CT units installed worldwide increased dramatically. Perhaps the first significant technical development came in 1974 when Dr. Robert Ledley (Fig. 1-10), a professor of radiology, physiology, and biophysics at Georgetown University, developed the first whole-body CT scanner. (Hounsfield’s EMI scanner scanned only the head.)

Dr. Ledley graduated with a doctorate in dental surgery from New York University in 1948, and in 1949 he earned a master’s degree in theoretical physics from Columbia University. He holds more than 60 patents on medical instrumentation and has written several books on the use of computers in biology and medicine.* In 1990 he was inducted into the National Inventors’ Hall of Fame for the invention of the automatic computed transverse axial CT scanner. In 1997, Dr. Ledley won the National Medal of Technology, an honor awarded by the President of the United States for outstanding contributions to science and technology (Ledley, 1999). Currently, he is the president of the National Biomedical Research Foundation at Georgetown University Medical Center.

These pioneering works were followed by the introduction of three generations (a term used to refer to the method of scanning) of CT scanners. In 1974 a fourth-generation CT system was developed (Fig. 1-11).

A computer capable of performing multiple functions is central to the CT system. The CT computer has undergone several changes over time (see Chapter 7).

Image quality is another significant development as a result of technologic changes. Although earlier images appeared “blocky,” images acquired later were remarkably improved (Fig. 1-12). Improvements in image quality included improved spatial resolution, decreased scan time, increased density resolution, and changes to the x-ray tube to facilitate the increased loadability required of whole-body scanners. For example, the matrix size in 1972 was 80 × 80; in 1993, it was 1024 × 1024. In addition, the spatial resolution and scan time in 1972 were reported to be three line pairs per centimeter (lp/cm) and 5 minutes, respectively, compared with 15 lp/cm and 1 second, respectively, in 1993 (Kalender, 1993). Increased loadability resulted in scanners capable of dynamic CT examinations that took a series of scans in rapid succession. Later-model CT scanners could operate in several modes such as the prescan localization mode, which produced a survey scan of the region of interest. Rapid reformatting of the axial scans into coronal, sagittal, and oblique sections also became possible.

High-Speed CT Scanners: In 1975 the dynamic spatial reconstructor (DSR) was installed in the biodynamics unit at the Mayo Clinic. The goal of the DSR was to carry out dynamic volume scanning to accommodate imaging of the dynamics of organ systems and the functional aspects of the cardiovascular and pulmonary systems with high temporal resolution as well as imaging anatomic details (Ritman et al, 1991; Robb and Morin, 1991). Research on this unit has since been discontinued.

In the mid 1980s, another high-speed CT scanner was introduced that used electron beam technology, a result of work by Dr. Douglas Boyd and colleagues during the late 1970s at the University of California at San Francisco. The scanner was invented to image the cardiovascular system without artifacts caused by motion. At that time the scanner was called the cardiovascular CT scanner. Later, this scanner was acquired and marketed by Siemens Medical Systems under the name Evolution and was subsequently referred to as the electron beam CT (EBCT) scanner. The most conspicuous difference between the EBCT scanner and conventional CT is the absence of moving parts. At that time, the EBCT scanner was capable of acquiring multislice images in as little as 50 and 100 milliseconds.

The U.S. Food and Drug Administration (FDA) cleared the EBCT scanner in 1983. As of 2007, the EBCT scanner is marketed by General Electric (GE) Healthcare under the name e-Speed and it now features proprietary technologies that play a significant role in imaging the heart. In October 2002, the e-Speed CT scanner received FDA clearance and it is now capable of 33-ms, 50-ms, and 100-ms True Temporal Resolution to produce images at up to 30 frames per second (GE Healthcare, 2007).

Because nearly 20 manufacturers made CT scanners between 1973 and 1983, a number of developments unique to particular manufacturers were also introduced (see Appendix B). The evolution of CT continued after 1983, with nearly 10 manufacturers competing for the CT market. (Box 1-1 highlights the developments from a major manufacturer actively involved in the research and development of the CT scanner).

Spiral/Helical CT Scanners: Volume Scanning: In conventional CT the patient is scanned one slice at a time. The x-ray tube and detectors rotate for 360 degrees or less to scan one slice while the table and patient remain stationary. This slice-by-slice scanning is time consuming, and therefore efforts were made to increase the scanning of larger volumes in less time. This notion led to the development of a technique in which a volume of tissue is scanned by moving the patient continuously through the gantry of the scanner while the x-ray tube and detectors rotate continuously for several rotations. As a result, the x-ray beam traces a path around the patient (Fig. 1-13). Although some manufacturers call this beam geometry spiral CT (the beam tracing a spiral path around the patient), others refer to it as helical CT (the beam tracing a helical path around the patient). This book uses both terms synonymously.

The idea of this approach to scanning can be traced to three sources (Kalender, 1995). In 1989, the first report of a practical spiral CT scanner was presented at the Radiological Society of North America (RSNA) meeting in Chicago by Dr. Willi Kalender (Fig. 1-14). Dr. Kalender has made significant contributions to the technical development and practical implementation of this approach to CT scanning. His main research interests are in the areas of diagnostic imaging, particularly the development and introduction of volumetric spiral CT.

Dr. Kalender was born in 1949 and studied medical physics in Germany. Subsequently, he continued his studies at the University of Wisconsin in the United States. He later worked with Siemens Medical Solutions in the area of CT, and in 1995 he became professor and chairman of the Institute of Medical Physics, which is associated with the University of Erlanger, Nürnberg, Germany.

The spiral/helical CT scanners developed after 1989 were referred to as single-slice spiral/helical or volume CT scanners. In 1992 a dual-slice spiral/helical CT scanner (volume CT scanner) was introduced to scan two slices per 360-degree rotation, thus increasing the volume coverage speed compared with single-slice volume CT scanners.

In 1998 a new generation of CT scanners was introduced at the RSNA meeting in Chicago. These scanners are called multislice CT (MSCT) scanners because they are based on the use of multidetector technology to scan four or more slices per revolution of the x-ray tube and detectors, thus increasing the volume coverage speed of single-slice and dual-slice volume CT scanners. The number of slices per revolution has been increasing at a steady pace, as outlined in Figure 1-15. In 2004, a 256-slice prototype CT scanner was undergoing clinical tests and more recently, a comparison of the doses from the 256-slice CT scanner and a 16-slice CT scanner has been reported (Mori et al, 2006). The front view of the 256-slice CT scanner is shown in Figure 1-16, (A, the wide area 2D detector is shown in Figure 1-16, B.)

In addition to the improvements in the performance characteristics such as, for example, minimum scan time, data per 360° scan, image matrix, slice thickness, spatial and contrast resolution, from 1972 to 2004, MSCT scanners are now based on new technologies. These include new x-ray tube and detector technologies to accommodate high-speed imaging, large-bore gantry apertures to facilitate ease of patient positioning and patient access, improved z-axis resolution, radiation dose optimization with dose modulation techniques, integrated software featuring multiple software platforms, and intuitive interfacing and software for not only data management such as workflow optimization but also for image processing. Two notable new concepts for MSCT, particularly for scanners that have 16 or greater detector rows, are the reconstruction algorithm and the definition of an important parameter called the pitch. These scanners now use cone beam image reconstruction algorithms to address the cone beam geometry as the detector width increases to accommodate 16 to 64 detector rows. The pitch for volume scanners has now been defined by the International Electrotechnical Commission (IEC), after much debate in the literature as to what the exact definition should be. The IEC states that the pitch for MSCT scanners is the ratio of the table feed per revolution of the x-ray tube to the total width of the collimated beam. Pitch is described in a subsequent chapter.

One of the central goals in the development of the CT scanner is to achieve isotropic resolution, that is the voxel is a perfect cube (equal length, width, and height) as opposed to nonisotropic resolution, where the voxel is not a perfect cube, as is clearly illustrated in Figure 1-17. The impact of isotropic resolution on CT images is improved image quality in all three dimensions; hence, 3D images are sharp in appearance and do not exhibit the stair-step artifacts seen in nonisotropic resolution 3D images. Scanners can now achieve isotropic resolution of less than 0.4 mm.

Dual Source CT Scanner: In 2005, Siemens Medical Solutions announced a new type of CT scanner at the RSNA, called the SOMATOM Definition. This is a dual source CT (DSCT) scanner that features two x-ray tubes and two detectors (Fig. 1-18) specifically intended for imaging cardiac patients in a very short time. As illustrated in Figure 1-19, the performance of the DSCT scanner compared with the single source CT (SSCT) scanner in imaging the beating heart at a low and stable heart rate (60 beats/min) is improved because of the higher temporal resolution of the DSCT scanner (Fig. 1-19, A). Furthermore, as the heart rate increases to 100 beats/min, the DSCT scanner offers much better image quality (sharpness) in both systolic and diastolic phases compared with the SSCT scanner, as illustrated in Figure 1-19, B.

Other Applications of Volume CT: Volume CT scanning has resulted in a wide range of applications such as CT fluoroscopy, CT angiography, 3D imaging, virtual reality imaging, and more recently improved cardiac CT imaging. CT has found applications in radiation therapy and nuclear medicine technologies as well. In radiation therapy, for example, the CT scanner is now used in CT simulation process (Fig. 1-20). As noted, the CT simulator includes the CT scanner that is specially equipped with a flat table top and lasers that facilitate accurate positioning of the patient and is coupled to the radiation treatment planning virtual simulator. As noted by Mutic et al (2003), “the scanner is accompanied by specialized software which allows treatment planning on volumetric patient CT scans in a manner consistent with conventional radiation therapy simulators.” This scanner can either be installed in the radiology department or in the radiation therapy department. In nuclear medicine, the CT scanner is combined with a positron emission tomography (PET) scanner (Ell, 2006) and single photon emission tomography (SPECT) to form single units referred to as a PET/CT scanner and a SPECT/CT scanner. PET/CT is described in detail in Chapter 16.

Mobile CT: A unique event in the evolution of CT technology was the development of a mobile CT scanner to image patients who are too ill or physically traumatized to be transported to a fixed CT scanner. Philips Medical Systems introduced one such scanner specifically for use in the operating room, intensive care unit, and emergency trauma unit. The portable CT scanner is compact and mounted on wheels to facilitate transportation of the unit to remote locations in the hospital by technologists.

Clinical Efficacy Studies: The term efficacy is synonymous with effectiveness, efficiency, and performance. A number of investigators designed efficacy studies to test the clinical usefulness of this new diagnostic technique. Studies reported the results of scanning the brain, spinal cord, neck, thorax, abdomen, retroperitoneum, pelvis, and extremities.

CT became well established in the diagnosis of diseases of the central nervous system, and in some cases, it eliminated the need for examinations such as pneumoencephalography and reduced the frequency of cerebral angiography. Disorders such as gliomas, metastases, intracranial lesions, aneurysms, infarctions, hemorrhage, and atrophy have been successfully detected by CT. Later, whole-body clinical applications (see Chapter 18) became effective. In addition, CT proved useful in radiation treatment planning to provide accurate isodose curves and in other areas, such as determination of the mineral salt content in bones, or quantitative CT. Studies using MSCT scanners clearly demonstrated their usefulness in a wide range of clinical applications, including playing a central role in cardiac imaging in applications such as coronary angiography, assessment of ventricular function and pulmonary veins, and calcium scoring (DeRoos, 2006).

Radiation Dose Studies: A fundamental goal of any new imaging technique is to provide maximum information content with the minimum radiation dose to the patient. Perry and Bridges (1973) measured both the cranial radiation dose and the gonadal radiation dose for a series of scans. Their results provided a foundation for future studies. Initially, the radiation dose to the patient was thought to be negligible because the beam in CT was tightly collimated, but the patient exposure for a series of CT scans usually exceeds that of film radiography of the same area (Seeram, 1982).

Radiation dose is an integral topic in CT technology because the CT beam geometry and method of acquiring images differ from those of conventional radiography. Several imaging parameters in CT affect dose, including slice thickness, noise, resolution detector efficiency, reconstruction algorithm, collimation, and filtration. Various dose studies have explored the ways in which these factors influence the dose.

These studies have also led to the development of special ways to measure and describe the dose (McNitt-Gray, 2002; Kalender, 2005; Seeram, 1999). Ionization chambers or thermoluminescent dosimeters are used to measure the dose. Dose descriptors include the single-scan dose profile, multiple-scan dose profile, CT dose index, multiple-scan average dose, and isodose curves.

Additionally, manufacturers have developed various schemes to reduce the dose in CT. One scheme uses three elements to keep the patient dose to a minimum during data acquisition: (1) combined applications to reduce exposure, a prepatient filtering technique that reduces the dose by about 15% compared with conventional CT, (2) new ultrafast ceramic detectors that reduce the dose by another 25%, and (3) on-line dose modulation, whereby the milliamperage (mA) is optimized to the patient characteristics (diameter and absorption) to reduce the dose. This particular development in CT dose reduction is referred to as automatic tube current modulation technique (Iball et al, 2006; Rizzo et al, 2006). Different manufacturers have different methods to do automatic exposure control (AEC) in CT. For example, although Siemens Medical Solutions and GE Healthcare refer to their AEC packages as “CARE Dose 4D” and “SmartmA,” respectively, Philips Medical Systems and Toshiba Medical Systems label their packages “DoseRight” and “SURE_Exposure,” respectively. CARE Dose 4D is illustrated in Figure 1-21.

Quality Control: As with any medical imaging system, CT scanners are subject to quality control procedures and tests. Testing system performance is vital to maintain optimal image quality and minimize the production of image artifacts. Because the CT system consists of several mechanical and electronic components, many quality control tests are currently available. These range from simple tests, which the technologist can perform by using various phantoms provided by the manufacturers, to more complex tests that may require the expertise of the radiologic physicist or biomedical engineer. More recently, the ACR has introduced a CT phantom referred to as the ACR phantom, as illustrated in Figure 1-22, to be used for any CT quality control program. The phantom basically consists of four sections which can be imaged simultaneously to generate several test results. Quality control for CT scanners will be described in detail in a later chapter.

Other Uses: CT is useful in areas other than medicine. For example, CT can be used to study internal log defects. Funt and Bryan (1987) investigated the use of CT technology in a sawmill. They developed and tested algorithms that automatically interpreted CT images of logs and stated, “The computer program uses the high density and elliptical shape of knots to distinguish them from good wood, and the low density and rough texture of rotten areas to separate rotten wood from sound wood.” Habermehl and Ridder (1997) detailed the use of a portable CT scanner to take images of trees to determine wood rots; locate knots, hollows, and other defects; and determine water distribution inside the tree trunk (Fig. 1-23). These portable tree CT scanners use a cesium 137 gamma radiation source with a fan beam falling on an array of about 30 sodium iodide detectors.

CT has also been used in paleoanthropology. Zonneveld et al (1989) found that CT can visualize internal anatomy of completely preserved Egyptian mummies (Fig. 1-24) (Yasuda et al, 1992). Additionally, Hoffman et al (2002) report the use of CT in what they refer to as paleoradiology. Recently, a CT system for dedicated imaging of mummies has become available and will be used in a major research project planned by the Egyptian Supreme Council of Antiquities in conjunction with Siemens Medical Solutions (who donated an MSCT scanner) and the National Geographic Society.

Several cases report the use of CT in paleo-ornithology, oil exploration, fat stock breeding, and other animal investigations (Fig. 1-25). In fat stock breeding, pigs are scanned to determine the meat quality defined as the best combination of water, protein, and fat, thus eliminating the need to kill the pigs for this determination (Fig. 1-26).

Sirr and Waddle (1999) used CT to evaluate bowed stringed instruments such as violins. The scans demonstrated varying degrees of internal damage (e.g., wormholes, air gaps, and plastic deformities of wood) or those resulting from repair (e.g., glue lines, filler substances, and wooden cleats and patches) not seen when the instruments are examined visually. The researchers also concluded that CT facilitated verification of authenticity and proof of ownership.

Another interesting and unique nonmedical use of CT is one by Meyer et al (2007). They used a 64-slice CT scanner to find out the combination of a bicycle lock belonging to the son of one of the authors who had forgotten the combination. Fortunately the lock was not on the bicycle. The lock was positioned on the scanner to show 0 on the right side of the lock. The lock was scanned with 350 mA, 120 kV, and a pitch of 0.33. The resulting images are shown in Fig. 1-27. As can be seen, the notches are clearly shown and reveal that the combination is 1789. This result must have made the little boy very happy.

More recently, CT scanners are now used at major airports for automated explosives detection. As baggage is scanned, the CT data (numbers) of the contents obtained as a result of the scanning are compared with CT numbers of known explosives or contraband. If the two sets of CT numbers correspond, an alarm is sounded.