4

CHAPTER

Multidetector
Computed Tomography

Ashley S. Shaw and Adrian K. Dixon

ingly complex studies of small structures, which require the

Computed tomography: a brief history
focal spot to be kept to an absolute minimum even when the
Technical overview of MDCT
mA is increased. For example, one manufacturer offers a focal
Image reconstruction and analysis
spot measuring 0.60.9 mm, depending on the mA used.
Clinical applications of MDCT
The X-rays are detected by various forms of detector. A
Whole body MDCT in asymptomatic adults ceramic scintillator, for example, absorbs the X-rays and pro-
Radiation dose considerations duces visible light. This stimulates a photodiode to produce an
Radiotherapy electrical signal from which the image can be constructed. In
Future directions order to reach the spatial resolution, detectors have become
smaller (typically around 0.6 mm), more sensitive (allowing
some dose reduction) and react faster (allowing increased gan-
COMPUTED TOMOGRAPHY: A BRIEF HISTORY try rotation speeds). In short, the smaller each detector is, the
smaller the possible collimation, whereas the wider the bank of
The first CT machine used for clinical purposes, developed detectors, the greater the potential coverage per rotation. Some
by the late Sir Godfrey Hounsfield, was installed at the Atkin- manufacturers arrange detectors of varying width such that
son Morley Hospital, London, in the early 1970s1. Each axial the smallest are positioned centrally and the largest positioned
image of the head took several minutes to acquire and days to peripherally, whereas others use identically sized detectors
reconstruct at the EMI laboratories. Over the next decade, CT across the array.The majority of machines in use are characterized
machines became faster as the processing power of comput- by the number of detectors (e.g. 16-, 40- or 64-slice) and use a
ers improved. In the mid-1980s, the development of slip-ring single focal spot on the tube. By contrast, one 64-slice machine
technology enabled continuous revolution of the X-ray unit, has a 32-detector array with two rapidly alternating focal spots
which reduced the acquisition time and allowed helical data to in the tube. In effect, this produces overlapping projections that
be acquired2. Since the late 1990s, the focus of development has are acquired and processed simultaneously.
been on reducing the size of the detectors in order that they The end result of these advances is that the latest technology
could be arranged in multiple rows along the z-axis.This permits from all manufacturers is able to produce isotropic images with
the acquisition of multiple slices simultaneously, thus reducing sub-mm resolution in all planes with ever-increasing rapidity.
acquisition time yet further. At the time of writing, state of the
art CT systems have up to 64 rows of detectors, allowing up to
4 cm to be imaged per revolution, a revolution time around 0.4 s, IMAGE RECONSTRUCTION AND ANALYSIS
and with a resolution of approximately 0.4 mm.
The rapid acquisition of such large datasets has necessitated
a change in radiologists approach to CT reporting. As a bare
TECHNICAL OVERVIEW OF MDCT minimum, both thick and thin section axial images should be
produced and stored. Images need to be viewedwon a worksta-
Before multidetector CT, imaging protocols and patient tion capable of producing images in all planes, with the ability
throughput were frequently limited by the time taken for the to change window level and width.
tube to cool between exposures. Refinements in tube design When viewing images, our experience has shown that, in
now ensure that multiple phases of imaging can be performed addition to standard lung or soft tissue windows, review of the
without overheating, as the tube cools faster than the through- data on bone windows (preferably in the sagittal plane) often
put of patients. These design improvements also allow increas- demonstrates unsuspected abnormalities (Fig. 4.1).The optimal
80 SECTION 1 IMAGING TECHNIQUES AND GENERAL ISSUES

slice thickness will depend on the clinical question posed.

Cardiac imaging, for example, will require sub-mm images,
whereas renal calculi will be reliably visualized on images
up to 3 mm thick, and the detection of lung nodules can be
enhanced by scrolling through overlapping 10 mm images
(allowing differentiation from branching blood vessels).
For 3D viewing of datasets, there are four principal post-
processing techniques: multiplanar reconstructions (MPR),
maximum intensity projection (MIP), surface shaded display
(SSD) and volume rendered (VR) images (Fig. 4.2). MPR
and associated curved planar reconstruction (CPR)simply
allows the image to be viewed in any plane chosen, including
a curved plane with CPR (Fig. 4.3). MIP displays the highest
attenuation voxels in an image only, thus allowing visualization
of structures (usually vessels) not in a single plane. However,
visualization of arterial wall calcification can be difficult and
vessels near to bony structures may be obscured. SSD gives 3D
representations of the anatomy in grey scale, but is prone to
noise and artefacts. VR allows visualization of the vessels dis-
tinct from the surrounding anatomy and is of particular value
Figure 4.1 Sagittal MPR imaged on bone windows, demonstrating in cerebral angiography, where the vessels pass through and lie
multiple sclerotic lesions throughout the vertebrae in a patient presenting in close proximity to the skull base.
with lymphoma. Computer aided detection (CAD) of lung nodules is
currently under development by a number of groups. Prelimi-
nary results indicate that, while more nodules can be detected,

Figure 4.2 (A) Coronal MPR demonstrating renal cell carcinoma at the lower pole of the left kidney extending to the hilum. (B) Coronal MIP
demonstrating renal cell carcinoma at the lower pole of the left kidney extending to the hilum. (C) Coronal SSD demonstrating renal cell carcinoma at
the lower pole of the left kidney extending to the hilum. (D) Coronal VR demonstrating renal cell carcinoma at the lower pole of the left kidney extending
to the hilum.
 CHAPTER 4 MULTIDETECTOR COMPUTED TOMOGRAPHY 81

Dynamic perfusion CT (PCT)5 evaluates regional blood

flow within the brain by serially imaging a limited area follow-
ing the infusion of intravenous contrast medium.The resulting
images give a computerized map of relative cerebral blood
flow, mean transit time and cerebral blood volumes, albeit
with limited coverage of the brain (Fig. 4.4)6. In the context
of patients presenting with acute stroke, PCT is significantly
more accurate than non-enhanced CT in detecting stroke

Figure 4.3 CPR demonstrating a normal calibre right coronary

artery.

there is a significant rise in the number of false-positive

nodules. Although not current routine clinical practice, CAD
may become a useful prompt in the evaluation of resectable
metastases and on lung cancer screening programmes (should
they prove to be of clinical value).

CLINICAL APPLICATIONS OF MDCT

Neuroradiology
In many areas, CT has been superseded by magnetic resonance
imaging (MRI) for the evaluation of neurological disease as
it gives superior differentiation of the soft tissues. However,
CT remains the initial imaging investigation for many patients,
particularly those presenting following trauma or with acute
neurological syndromes.
Digital subtraction angiography (DSA) has long been
regarded as the gold standard for detection of an intracranial
aneurysm. Single row detector helical CT could not acquire
data sufficiently quickly to obtain reliable diagnostic angio-
graphic studies. MDCT has been shown to be highly accurate
in the detection of aneurysms, with a sensitivity of over
96% and specificity of 100% on a 16-detector row ma-
chine, identical to the DSA results3. Even for small aneu-
rysms (< 3 mm), sensitivity remained over 90%. This has led
to a significant change in neuroradiological and neurosurgical
practice, enabling treatment to be planned from 3D images at
a workstation, with a consequent reduction in the number of
diagnostic angiograms performed. Furthermore, it is now pos-
sible to perform multiple phases of vascular imaging within
Figure 4.4 (A) Unenhanced CT of the head in a patient presenting
the brain, obtaining both an arterial and a venous phase4. with acute aphasia. (B) Cerebral perfusion map demonstrating delayed
This may provide additional information in patients with an perfusion in the left middle cerebral artery territory. Images courtesy of
arteriovenous malformation, but needs further evaluation. Rosie Farmer, Addenbrookes Hospital, Cambridge, UK.
82 SECTION 1 IMAGING TECHNIQUES AND GENERAL ISSUES

(7586% versus 66%), the extent of stroke (94% versus 43%),

and has a high interobserver agreement7. This may prove to be
of clinical use in those patients in whom thrombolysis or other
novel therapies are being considered. Potentially, the technique
can differentiate the ischaemic penumbra from infarcted tissue,
identifying patients who would benefit from revascularization
and those in whom unnecessary treatment could be avoided.
The data obtained from this technique are very similar to that
obtained with perfusion and diffusion-weighted MRI but CT
is considerably easier in the acute situation. It is likely that
routine clinical use will be guided by developments in therapy
for acute stroke. Similarly, the use of perfusion CT follow-
ing severe head trauma has been reported to yield indepen-
dent prognostic information regarding functional outcome8,
although again this is not part of current routine clinical prac-
tice.

Cardiac
The rapid acquisition times and increased spatial resolution of
MDCT probably holds the greatest potential for cardiac imag-
ing. Imaging of the coronary arteries with CT has two major Figure 4.5 VR image of the heart depicting the coronary arteries.
components: unenhanced coronary artery calcium scoring
and contrast-enhanced coronary artery angiography. Elec-
tron-beam CT (EBCT) has been used for coronary calcium
scoring for approximately two decades and is a highly sensitive Assessment of perfusion can be made on delayed imaging.The
method of detecting coronary artery atherosclerosis. A number potential impact of noninvasive coronary imaging on clinical
of authors have shown there to be a high correlation between practice is clearly enormous, and will depend on the results
EBCT and MDCT. However, there is a well-recognized high achieved with newer MDCT technology.
interscan, interobserver and intraobserver variability and the
consensus of the American Heart Association/American Col- Thoracic
lege of Cardiology is that the incremental value of calcium The imaging of patients with suspected pulmonary venous
scores over traditional risk assessments has not been proven thromboembolism (PE) had, for many years, relied on either
and that its use could only be justified in selected groups at ventilationperfusion (VQ) studies or pulmonary angiogra-
intermediate risk9. phy. However, once CT became sufficiently fast to image the
CT coronary angiography is typically performed in a single thorax with thin collimation ( 2 mm) in a single breath-hold,
breath-hold, with ECG gating, during a rapid ( 5 ml/s) infu- CT pulmonary angiography (CTPA) rapidly became the test
sion of iodinated contrast material (Fig. 4.5). The sensitivity of choice for many clinicians. Indeed, the number of patients
and specificity of 16-detector row CT for detection of 50% referred for CT with suspected PE has increased dramatically
luminal stenosis is around 90%10,11, although a significant pro- since its introduction. This is likely to be due to increased
portion of the arteries (12%)11 were not demonstrated suffi- awareness of the diagnosis in combination with easy access to
ciently well to be evaluated. However, the negative predictive an accurate noninvasive study.
value (NPV) of a normal study is reported to be 95%12 and VQ studies expose the patient to a significantly lower
thus could be potentially used to exclude cardiac disease in radiation dose than CTPA, but have a number of drawbacks.
patients with chest pain.There are, of course, limitations: prob- There is a significant proportion in which the result will be
lems occur in patients with arrhythmias (e.g. atrial fibrillation), indeterminate due to cardiorespiratory disease or an abnormal
the heart rate should ideally be 60 beats per minute or below13, radiograph14. Once these are taken into account, the overall
and beta-blockers may be necessary for optimal imaging. radiation burden to the population is only slightly (approxi-
In order to detect a 1020% change in stenosis, imaging will mately 4%) higher using CTPA for all15. Moreover, the greater
require an isotropic spatial resolution of 0.3 mm and an acqui- specificity of a positive CTPA, compared with a high prob-
sition time of approximately 0.1 stechnical specifications ability of PE, gives the clinician greater confidence in the
that, as detailed above, are becoming possible with double tube diagnosis.
64-slice systems CT. It is anticipated that this development The negative predictive value of a CTPA for PE is 99.4% in
will further enhance the accuracy and reproducibility of coro- patients at 3-month follow-up, with a mortality attributable to
nary artery angiography and, importantly, give an overview PE within this time period of 0.01%16. The current generation
of cardiac morphology and function. By using all the data of MDCT machines produces consistently high quality images
during the various phases of the cardiac cycle, cine loops can of the pulmonary arteries down to the subsegmental level. How-
provide ejection fractions and other objective measurements. ever, the clinical significance of tiny peripheral emboli is unclear
 CHAPTER 4 MULTIDETECTOR COMPUTED TOMOGRAPHY 83

as they may not affect outcome17. Consequently, CTPA is now the presence of an endoleak will be seen as contrast medium
widely considered to be the equal of pulmonary angiography in within the aorta tracking into the thrombus and implies failure
terms of accuracy without the associated risks. of the stent graft. Enlargement of the aneurysmal sac with-
Imaging of the venous system 2 min postinjection from the out an apparent endoleak is termed endotension. Visceral
iliac crests to the popliteal fossa has been advocated by several infarction, particularly involving the kidneys, should also be
groups for all patients undergoing CTPA. The evidence sug- apparent on the arterial study.
gests that 20% of patients with a negative CTPA will have When using MDCT to investigate patients with peripheral
venous thrombosis in the legs18. However, it is difficult to vascular disease, it is important to appreciate that the flow rate
justify replacing a highly accurate, noninvasive and radiation- down both normal and atherosclerotic peripheral vessels can
free test (ultrasound) with CT, particularly as the gonads are be highly variable. Consequently, it is our practice to perform
unavoidably included in the imaging. two test bolus studies: one in the lower aorta and a second in
the popliteal fossa. We then subtract the time delay between
Vascular the two and set the table speed accordingly. Post-processing of
The high resolution and speed afforded by MDCT has a num- the data can be time consuming and the images in the distal
ber of potential applications in the assessment of the vascular vessels can be variable, as with all forms of peripheral vascular
system elsewhere. The increased speed allows a pure arterial imaging. In our experience, the results from 16-detector row
phase study to be obtained; however, the faster the machine, and above are comparable with magnetic resonance angiogra-
the less forgiving it is. It is imperative that all such studies are phy (MRA) and have the advantage of providing bony land-
either triggered by the arrival of contrast medium at a defined marks to help treatment planning. Both of these techniques
site or are based on a test bolus. In our practice, unenhanced yield a 3D dataset, unlike conventional arteriography, and have
and arterial-phase images are acquired. This aids detection of the advantage of being noninvasive. It would seem likely that
intramural haematoma or haemorrhage in acute patients and the role of MDCT in this field will grow in coming years,
differentiates calcification from endoleak in treated patients. with arteriography increasingly reserved for therapeutic cases.
The commonest vascular application clinically is the assess-
ment of aortic aneurysms, both before and after treatment.The Trauma
enhanced resolution, particularly in the z-axis, together with Accurate assessment of the cervical spine is of critical impor-
multiplanar reconstruction (Fig. 4.6), allows accurate planning tance in trauma cases: the consequences of missing a significant
of endovascular or surgical repair. Following endovascular injury are potentially devastating. The reported accuracy of
aneurysm repair (EVAR), MDCT is regarded by the Society plain radiographic studies is highly variable, but approximately
of Interventional Radiology (SIR) as the gold standard for 1020% of significant cervical spine injuries will be missed
monitoring the aorta. Particular attention should be paid to through a combination of suboptimal radiographs, interpreta-
the presence of any fractures or kinking of the stent on the tion error and absence of radiographic signs19. In the obtunded
unenhanced study. Following intravenous contrast medium, patient, these problems are compounded by the absence of any
clinical information from physical examination. Single detec-
tor row CT is superior to plain radiography, but the relatively
poor resolution in the z-axis could prove problematic in iden-
tifying fractures in the axial plane. MDCT has overcome this
with increased speed and resolution, such that the entire cervi-
cal spine can now be imaged and viewed in any plane within
a matter of seconds. The sensitivity, specificity and NPV of
MDCT are reported to be 98.1%, 98.8% and 99.7%, respec-
tively, with many major trauma centres now using MDCT as
the primary imaging method for evaluating the cervical spine
in the obtunded trauma patient20,21.
Imaging of the thorax and abdomen following trauma is
usually performed with MDCT; ultrasound may be of use in
confirming the presence of a haemoperitoneum in the unsta-
ble patient before laparotomy, but cannot reliably assess the full
extent of solid organ or GI tract injuries. In addition to the
non-enhanced imaging, the speed of image acquisition with
MDCT allows the radiologist to perform each phase of the
study at an appropriate delay following intravenous enhance-
ment. The thorax should be imaged in the arterial phase, as
deceleration injuries may lead to shearing injuries of the aorta
(most commonly occurring at the isthmus, just distal to the
Figure 4.6 Colour volume rendered image of the aorta origin of the left subclavian artery). The abdomen should ide-
demonstrating an infrarenal aneurysm extending down to the bifurcation. ally be imaged in the arterial and portal venous phases, in order
84 SECTION 1 IMAGING TECHNIQUES AND GENERAL ISSUES

that both vascular injuries (dissection, pseudoaneurysm, avul- identifying the arterial and venous anatomy of the donor for
sion) and parenchymal injuries may be assessed. The pelvis is surgical planning. Following transplantation, we have found
usually imaged in the portal venous phase. The sensitivity and MDCT to be particularly useful in evaluating the vascular
specificity of CT for detecting bowel and mesenteric injury supply to the graft in cases of suspected ischaemia where US is
are approximately 80%22, while for solid organ injuries these nondiagnostic (Fig. 4.7).
approach 100%.
Genitourinary
Gastrointestinal The evaluation of a patient presenting with haematuria may
CT colonography (CTC) has been extensively investigated involve intravenous urography, ultrasound, CT or MRI in com-
over the past decade, with the ultimate goal of providing an bination with cystoscopy. Increasingly, MDCT is being used as
accurate, minimally invasive screening test for detecting colonic the primary imaging investigation as it has a high sensitivity for
polyps before they become frankly malignant. The evolution detecting malignant lesions, calculi and traumatic injuries. The
of MDCT has enabled the development of CTC. Current other major causes (infection, coagulopathy, instrumentation)
best practice involves thin collimation ( 3 mm) and low dose may be diagnosed on clinical history and blood and urine test-
imaging of the insufflated colon in both prone and supine ing.The technique employed should encompass an unenhanced
positions following either purgative or faecal tagging with study (to detect calculi), a nephrographic phase (approximately
multiple doses of oral contrast media for 4872 h23. The use of 100 s, to assess the renal parenchyma) and an excretory phase
intravenous hyoscine is recommended by some authors. (810 min, to assess the ureters). The latter two stages may be
Reporting is usually, though not always, performed from combined by giving the intravenous contrast medium in two
the axial and MPR images with 3D endoluminal fly-through parts, 810 min apart, and then imaging 100 s after the second
views used for problem solving. However, the precise tech- dose. This will significantly reduce the radiation dose to the
niques employed are as variable as the results achieved, which patient. As MDCT may not depict flat tumours of the bladder
range from as low as 59% sensitivity for polyps > 10 mm24 to wall, cystoscopy must not be omitted.
93% (outperforming colonoscopy) in the detection of polyps The increased anatomical resolution and multiplanar
> 8 mm25. CTC outperforms double contrast barium enema reconstructions possible with MDCT have enabled its use for
(DCBE) in all studies. The clinical significance of missed and surgical planning. In particular, urologists are able to accurately
diminutive polyps has also been questioned, given the slow assess whether resection of a malignant lesion requires partial
progression of polyps over time26. Moreover, there is wide or complete nephrectomy, whilst potential organ donors may
variation in the reporting of polyps even among expert read- be noninvasively evaluated with an accurate depiction of the
ers27, illustrating that this truly is a technique in evolution from renal vasculature.
a technical perspective, with a steep learning curve for the
radiologist. Paediatric
Multicentre trials are currently underway in the USA Imaging children with CT poses a number of problems. It
(ACRIN) and the UK (SIGGAR 1) to evaluate the use of requires cooperation with instructions and the need to remain
CTC in the context of a screening programme, but at present still, requires a high degree of spatial resolution to depict
the majority view is that colonoscopy should be viewed as the smaller organs and their vessels and exposes patients to a rela-
gold standard, with CTC likely to replace the DCBE in the tively high dose of radiation. Some hold that around 1 per
future. Again CAD techniques are under evaluation.

Hepatobiliary
The use of multiple phases of contrast-enhanced imaging has
probably had the greatest impact in liver imaging, which is
reflected in the extensive US, CT and MR literature on the
subject. The detection and characterization of focal liver lesions
is largely based on patterns of vascular enhancement, with the
hypervascular lesionshepatocellular carcinoma (HCC), regen-
erative nodules, focal nodular hyperplasia and adenomain
particular providing diagnostic dilemmas. In order to optimize
contrast enhancement, our practice is to trigger the CT study
from the arrival of contrast in the abdominal aorta. CT-based
surveillance programmes for early detection of HCC with cir-
rhosis consist of an arterial and portal venous phase study at the
very least, with many centres also obtaining unenhanced images.
Both early and late arterial phase imaging can be performed, but
no additional benefit for this has yet been shown. Figure 4.7 Coronal MIP image demonstrating hepatic artery
In living related liver transplantation, MDCT may be used thrombosis with collateral arteries in a patient following liver
to determine liver volumes and is increasingly utilized for transplantation.
 CHAPTER 4 MULTIDETECTOR COMPUTED TOMOGRAPHY 85

12002000 patients undergoing CT might develop cancer population. In 20032004, it is estimated that this has increased
because of the effects of the CT radiation, and these risks are to 9%, with the dose from CT contributing around half of all
greater in children28. radiation from medical exposures33. Since 1997, CT has been
Advances in MDCT have led to a significant reduction in designated a high-dose procedure by the European Union,
acquisition time, which has been shown to reduce the need for along with interventional radiology and radiotherapy.
sedation of paediatric patients29. Simultaneously, it has become The reasons for this expansion in practice have been out-
possible to reduce the collimation below 1 mm, dramatically lined above, with MDCT able to perform increasingly complex
improving the resolution, and thereby not only aiding the diagnostic procedures noninvasively and, in some instances,
diagnosis but also enabling accurate assessment of congenital requiring multiple phases of imaging. Furthermore, unlike
abnormalities and surgical planning, particularly in patients conventional plain radiography, where the radiation dose from
with malignancy. Unlike in adult patients, imaging in differ- each exposure is to some extent regulated by automatic detec-
ent vascular phases is discouraged because of radiation issues. tors (increasing the dose will result in an overexposed, dark
Therefore it is important to select a single optimal imaging radiograph of little diagnostic value), increasing the exposure
sequence whenever possible (e.g. in a Wilms tumour, the late factors (and patient dose) for a CT study will provide the user
arterial phase will also opacify the renal veins and upper IVC). with higher quality images. More difficult to assess is the effect
Optimizing an MDCT study for a paediatric patient should that an increasingly litigious society has had on the practice of
also include reducing the radiation dose as much as is pos- medicine and in particular the number of requests for studies
sible while maintaining diagnostic quality. Reducing the kV to rule out underlying tumour, pulmonary embolism, etc.,
and/or mAs will significantly reduce the dose to the patient30, even when these are clinically unlikely.
and this can now be modulated automatically during imaging The responsibility for reducing patient dose should be
on newer machines, giving a constant image signal-to-noise shouldered by all parties. The referring clinician should ensure
ratio throughout the study. Additionally, increasing the pitch that the radiologist is given full clinical information to ensure
reduces the radiation dose significantly without loss of diag- that CT is indeed the most appropriate test. The radiologist
nostic quality30. should ensure that each study is justified, that the imaging pro-
tocols are optimized to answer the clinical question and that
the dose to the patient follows the ALARA (as low as reason-
WHOLE BODY MDCT IN ASYMPTOMATIC ably achievable) principle. CT manufacturers also play a key
ADULTS role in this area through the continued development of dose
modulation and the installation of low-dose preset protocols.
Whole body CT imaging of asymptomatic patients is con-
In order to maximize patient safety, it is essential that all these
troversial, yet is becoming commonplace in several countries,
issues are addressed in each case.
particularly the United States. Clearly, the potential benefits
of identifying early stage malignant lesions or coronary heart
disease with one noninvasive test could be enormous, but as RADIOTHERAPY
yet the case for whole body CT screening is far from proven.
Over the past decade, significant advances have been made in
When considering these studies, it should be borne in mind
the planning and delivery of radiotherapy. The introduction of
that a single CT examination cannot be optimized for looking
MDCT enabled the oncologist to map the extent of tumour in
at all organs at once. Consequently, the sensitivity and speci-
three planes, allowing accurate dose planning of irregular shapes
ficity of a whole-body study is likely to be significantly lower
while minimizing the dose to the surrounding normal tissues.
than with the dedicated organ-specific studies currently being
This 3D conformal radiotherapy (CRT) results in a significant
evaluated. Inevitably, further follow-up investigations will be
reduction in side-effects. However, there were limitations in
required for incidental findings, resulting in anxiety and distress
corrections that could be made to the delivered dose. More
for the patient and occasionally lead to morbidity and mortality.
recently, intensity modulated radiotherapy (IMRT) has been
Moreover, the radiation dose is nontrivial. A recent publica-
developed, whereby each radiation beam is divided into 1-cm2
tion exploring the hypothetical situation of a 45-year-old man
beamlets, each delivering a different prescribed dose. This has
undergoing annual screening CT until the age of 75 calculated
reduced even further the dose delivered to the surrounding
the overall estimated lifetime attributable risk of cancer mor-
normal tissues through increasing accuracy of delivery.
tality to be approximately 1.9% (1 in 50)31. This combination
Current research in radiotherapy includes the development
of high cost, low specificity and high risk indicates that it is
of techniques to account for movement during radiotherapy,
unlikely that whole body CT screening would be of benefit
particularly respiration, with 4D CRT. Moreover, techniques
to the population. In the UK, such CT examinations are not
are being developed which take account of the changes to the
recommended by the National Screening Committee32.
patient and tumour (weight loss and tumour shrinkage) that
occur during a course of radiotherapy, so-called adaptive radio-
RADIATION DOSE CONSIDERATIONS therapy. It is anticipated that both of these will again further the
accuracy of radiotherapy with more precise dose delivery and
In the late 1980s, CT represented approximately 2% of radio- reduced side-effects. The increasing use of functional imaging
logical investigations and 20% of the collective dose to the of tumours with 18FDG-PET is also beginning to impact on
86 SECTION 1 IMAGING TECHNIQUES AND GENERAL ISSUES

radiotherapy planning. The co-registration of anatomical and on electron-beam computed tomography for the diagnosis and
physiological images using CT-PET has resulted in significant prognosis of coronary artery disease. Circulation 102: 126140
10. Nieman K, Cademartiri F, Lemos P A, Raaijmakers R, Pattynama P M, de
advances in lung tumour radiotherapy in particular, as it is Feyter P J 2002 Reliable noninvasive coronary angiography with fast
now possible to differentiate between the central obstructing submillimeter multislice spiral computed tomography. Circulation 106:
tumour and distal collapse. 20512054
11. Ropers D, Baum U, Pohle K et al 2003 Detection of coronary artery
stenoses with thin-slice multi-detector row spiral computed
FUTURE DIRECTIONS tomography and multiplanar reconstruction. Circulation 107: 664666
12. Schuijf J D, Bax J J, Salm L P et al 2005 Noninvasive coronary imaging
and assessment of left ventricular function using 16-slice computed
The gantry speeds and coverage currently achieved in MDCT tomography. Am J Cardiol 95: 571574
are more than adequate for almost all clinical applications. The 13. Cademartiri F, Mollet N R, Runza G et al 2006 Diagnostic accuracy of
most notable exception, cardiac imaging, will undoubtedly drive multislice computed tomography coronary angiography is improved at
manufacturers to produce yet faster machines with greater cover- low heart rates. Int J Cardiovasc Imaging 22: 101105; discussion 107109
14. Cross J J L, Kemp P M, Walsh C G, Flower C D R, Dixon A K 1998 A
age until high-quality,real-time images are achieved, which could
randomized trial of spiral CT and ventilation perfusion scintigraphy for
potentially directly replace all diagnostic coronary angiography. the diagnosis of pulmonary embolism. Clin Radiol 53: 177182
A significant improvement in resolution will be more difficult 15. ONeill J, Murchison J T, Wright L, Williams J 2005 Effect of the
to achieve using the current technology with up to 64 banks of introduction of helical CT on radiation dose in the investigation of
detectors.This would require significant improvements in detec- pulmonary embolism. Br J Radiol 78: 4650
16. Quiroz R, Kucher N, Zou K H et al 2005 Clinical validity of a negative
tor efficiency and size reduction. However, the next generation computed tomography scan in patients with suspected pulmonary
of CT machines may instead employ flat panel detectors. Early embolism: a systematic review. JAMA 293: 20122017
reports of prototype machines indicate that spatial resolution is 17. Prologo J D, Gilkeson R C, Diaz M, Cummings M 2005 The effect of
greatly improved, with isotropic voxels of the order of 0.25 mm, single-detector CT versus MDCT on clinical outcomes in patients with
suspected acute pulmonary embolism and negative results on CT
with coverage of approximately 20 cm in the z-axis per revolu-
pulmonary angiography. Am J Roentgenol 184: 12311235
tion. However, contrast resolution is relatively poor and data 18. Cham M D, Yankelevitz D F, Henschke C I 2005 Thromboembolic disease
acquisition is still slow on these machines at present. One of the detection at indirect CT venography versus CT pulmonary angiography.
most important potential applications for such high-resolution Radiology 234: 591594
studies, in addition to those described above, would be imaging 19. Tins B J, Cassar-Pullicino V N 2004 Imaging of acute cervical spine
injuries: review and outlook. Clin Radiol 59: 865880
of the breast. Early reports indicate that the technique is fea- 20. Brohi K, Healy M, Fotheringham T et al 2005 Helical computed
sible at a dose comparable to that from 2-view mammography, tomographic scanning for the evaluation of the cervical spine in the
with the added advantages of better tumour localization and unconscious, intubated trauma patient. J Trauma 58: 897901
potentially better detection of tumours. The tumour detection 21. Griffen M M, Frykberg E R, Kerwin A J et al 2003 Radiographic clearance
of blunt cervical spine injury: plain radiograph or computed tomography
and dose characteristics will clearly be of greater importance
scan? J Trauma 55: 222226
than the speed of data acquisition for breast work, but increased 22. Elton C, Riaz A A, Young N, Schamschula R, Papadopoulos B, Malka
speed will be crucial for cardiac applications. V 2005 Accuracy of computed tomography in the detection of blunt
bowel and mesenteric injuries. Br J Surg 92: 10241028
23. Taylor S A, Halligan S, Bartram C I et al 2003 Multi-detector row CT
REFERENCES colonography: effect of collimation, pitch, and orientation on polyp
detection in a human colectomy specimen. Radiology 229: 12
1. Ambrose J, Hounseld G N 1973 Computerized transverse axial 24. Rockey D C, Paulson E, Niedzwiecki D et al 2005 Analysis of air contrast
tomography. Br J Radiol 46: 148149 barium enema, computed tomographic colonography, and colonoscopy:
2. Kalender W A, Seissler W, Klotz E, Vock P 1990 Spiral volumetric CT with prospective comparison. Lancet 365: 305311
single-breath-hold technique, continuous transport, and continuous 25. Pickhardt P J, Choi J R, Hwang I et al 2003 Computed tomographic
scanner rotation. Radiology 176: 181183 virtual colonoscopy to screen for colorectal neoplasia in asymptomatic
3. Tipper G, U-King-Im J M, Price S J et al 2005 Detection and evaluation adults. N Engl J Med 349: 21912200
of intracranial aneurysms with 16-row multislice CT angiography. Clin 26. Macari M, Bini E J, Jacobs S L et al 2004 Signicance of missed polyps at
Radiol 60: 565572 CT colonography. Am J Roentgenol 183: 127134
4. Matsumoto M, Kodama N, Sakuma J et al 2005 3D-CT arteriography 27. Johnson C D, Harmsen W S, Wilson L A et al 2003 Prospective blinded
and 3D-CT venography: the separate demonstration of arterial-phase evaluation of computed tomographic colonography for screen detection
and venous-phase on 3D-CT angiography in a single procedure. of colorectal polyps. Gastroenterology 125: 311319
Am J Neuroradiol 26: 635641 28. Brenner D J, Elliston C D, Hall E J, Berdon W E 2001 Estimated risks of
5. Miles K A, Hayball M, Dixon A K 1991 Colour perfusion imaging: a new radiation-induced fatal cancer from pediatric CT. Am J Roentgenol 176:
application of computed tomography. Lancet 337: 643645 289296
6. Gillard J H, Minhas P, Hayball M P et al 2000 Evaluation of quantitative 29. Pappas J N, Donnelly L F, Frush D P 2000 Reduced frequency of sedation
computed tomographic cerebral perfusion imaging with H215O positron of young children with multisection helical CT. Radiology 215: 897899
emission tomography. Neurol Res 22: 457464 30. Donnelly L F, Emery K H, Brody A S et al 2001 Minimizing radiation dose
7. Wintermark M, Fischbein N J, Smith W S, Ko N U, Quist M, Dillon for pediatric body applications of single-detector helical CT: strategies
W P 2005 Accuracy of dynamic perfusion CT with deconvolution in at a large childrens hospital. Am J Roentgenol 176: 303306
detecting acute hemispheric stroke. Am J Neuroradiol 26: 104112 31. Brenner D J, Elliston C D 2004 Estimated radiation risks potentially
8. Wintermark M, van Melle G, Schnyder P et al 2004 Admission perfusion associated with full-body CT screening. Radiology 232: 735738
CT: prognostic value in patients with severe head trauma. Radiology 32. Dixon A K 2004 Whole-body CT health screening. Br J Radiol 77: 370371
232: 211220 33. Shrimpton P C, Hillier M C, Lewis M A, Dunn M 2005 Doses from
9. ORourke R A, Brundage B H, Froelicher V F et al 2000 American College computed tomography (CT) examinations in the UK2003 review.
of Cardiology/American Heart Association Expert Consensus document National Radiation Protection Board, Oxon, UK