CTCOLO~1

Prévia do material em texto

CT Colonography Versus Optical Colonoscopy
for Screening Asymptomatic Patients for
Colorectal Cancer:
A Patient, Intervention, Comparison, Outcome (PICO) Analysis1
Robert H. El-Maraghi, MD, Ania Z. Kielar, MD
Rationale and Objectives. The American College of Radiology has recently endorsed the use of computed tomographic colo-
nography (CTC) for colon cancer screening. With advances in technology and postprocessing software, the quality of computed
tomographic colonographic studies has improved, and new techniques are being developed to reduce radiation exposure and
increase patient acceptance of the procedure. The aim of colorectal cancer screening is to reduce the incidence of malignancy by
identifying and removing presymptomatic lesions. The aim of this study was to answer the clinical question: In an asymptomatic
patient at average risk for colon cancer, is CTC equivalent to optical colonoscopy (OC) for detecting clinically significant polyps?
Materials and Methods. A systematic literature review was conducted to evaluate CTC compared to OC, using the patient,
intervention, comparison intervention, outcome (PICO) search strategy. The PubMed search used Medical Subject Headings,
including the terms ‘‘computed tomography colonography,’’ ‘‘colonoscopy,’’ ‘‘screening,’’ and ‘‘polyp.’’ Each of the retrieved
articles was assigned a level of evidence using the Centre for Evidence-Based Medicine’s hierarchy of validity for diagnostic studies.
Results. PICO search criteria and review of abstracts identified 16 relevant studies. Using the Centre for Evidence-Based
Medicine’s hierarchy of validity, there were three level 1c studies, two level 2a studies, three level 2b studies, four level 3b
studies, two level 4 studies, and two level 5 studies. All relevant studies demonstrated that CTC had high or moderately high per
patient and per polyp sensitivity and specificity compared to OC for clinically relevant polyps (>5 mm).
Conclusions. The majority of evidence suggests that CTC is an acceptable alternative to OC, particularly in the group of patients
who are either unwilling or unable to undergo OC. The results of the large, multicenter American College of Radiology Imaging
Network study are pending. This trial presented preliminary results in 2007 suggesting that the sensitivity and specificity of CTC
are high and comparable to those of OC.
Key Words. Evidence-based medicine; methods standards; radiology; CT colonography; optical colonoscopy; polyp; colon cancer.
ª AUR, 2009
Currently, colorectal cancer (CRC) is the third most common
type of cancer to affect adults in North America, and the
approximate lifetime risk for developing this cancer in men is
Acad Radiol 2009; 16:564–571
1 From the Division of Medical Oncology, Royal Victoria Hospital, Barrie, ON,
Canada (R.H.E.); and the Ottawa Hospital, University of Ottawa, Diagnostic
Imaging Department, Ottawa Hospital-Civic Campus, 1053 Carling Ave,
Ottawa, ON K1Y 4E9, Canada (A.Z.K.). Received September 22, 2008;
accepted January 6, 2009. Address correspondence to: A.Z.K. e-mail:
aniakielar@gmail.com
ª AUR, 2009
doi:10.1016/j.acra.2009.01.008
564
6.7%, compared to 6.1% in women (1). In the majority of
cases, the precursor to developing this type of cancer is
a polyp, so screening has been advocated since the 1990s. The
aim of CRC screening is to reduce the incidence of malig-
nancy by identifying and removing presymptomatic lesions,
thereby reducing CRC morbidity and mortality (2). However,
only 42% of Americans aged >50 years have undergone any
type of screening, including fecal occult blood testing, sig-
moidoscopy, or optical colonoscopy (OC) (3,4). Numerous
factors have been postulated to be related to the low screening
rates, including a lack of awareness, patient discomfort,
socioeconomic causes, and a lack of availability (3,5,6).
mailto:aniakielar@gmail.com
Academic Radiology, Vol 16, No 5, May 2009 CTC: A PICO ANALYSIS
Recently, the American College of Radiology endorsed
the use of computed tomographic colonography (CTC) for
screening patients for colon cancer. CTC was first described
in 1994 by Vining et al (7) as a method for evaluating the
colonic lumen. This technique has been significantly refined
over the past decade, initially being used as a problem-solv-
ing tool for patients with incomplete colonoscopy or equiv-
ocal barium enema results. More recently, CTC has been
investigated as a screening modality for patients at average
risk for developing CRC. With advances in technology and
postprocessing software, the quality of the resulting studies
has improved, and new techniques are being developed to
reduce patient radiation exposure and increase patient ac-
ceptance of the procedure. With CTC as a minimally invasive
alternative to OC, it is possible that more people may be
willing to undergo CRC screening, possibly resulting in the
earlier detection of abnormalities, increasing the likelihood of
successful intervention, and ultimately reducing the number
of CRC deaths.
Vast amounts of data are published in the medical litera-
ture, and it is often daunting for busy clinicians to identify
what is the ‘‘best evidence’’ regarding a specific patient
scenario (8). Thus, the concept of evidence-based medicine
(EBM) was developed at the University of McMaster in the
1990s, with the goal of efficiently using current medical
evidence about available diagnostic, treatment, and preven-
tion options related to a defined patient-centered medical
question.
The patient, intervention, comparison intervention, out-
come (PICO) method applies five main steps as an evidence-
based approach and has been described in detail in a series of
review articles (9,10). The goal of the PICO method is to
reformat a clinical question, permitting a focused literature
search to retrieve relevant articles required to answer the
question effectively using the best available medical evi-
dence. These articles are evaluated and appraised so that in-
formation gleaned from the literature review may be applied
to clinical practice. Important elements include identifying
a specific patient or patient group of interest (P), as well as the
‘‘new’’ intervention as it compares to the reference standard
(I) particular to the clinical scenario (C), and then defining the
outcome of interest (O).
CLINICAL SCENARIO
A 55-year-old white man presented to his family physi-
cian for a routine examination. He had not previously un-
dergone any form of CRC screening and had no family
history of colon cancer. After talking to some of his friends
about their experiences, he was skeptical regarding the ac-
curacy of fecal occult blood testing and was not very enthu-
siastic about OC, because of perceived discomfort and his
dislike of the use of sedation. However, he had recently heard
about CTC as a potential screening tool that might be offered
in his area and wanted to know if it was a reasonable alter-
native to OC. Unable to provide a definitive answer, the
physician agreed to research the question and report back to
the patient. To this end, a PICO question was formulated to
perform an adequate literature search and to provide an evi-
dence-based medical summary of the retrieved articles: In an
asymptomatic patient at average risk for colon cancer, is CTC
equivalent to OC for detecting clinically significant polyps?
An average-risk individual is defined as one aged >50 years,
with no symptoms attributable to the bowel (eg, hemato-
chezia, iron deficiency anemia), and with no personal or
family history of CRC in a first-degree relative. In addition,
small polyps (#5 mm) have been deemed to be of no clinical
significance and are called diminutive polyps (11). Interme-
diate-sized polyps are those measuring 6 to 9 mm, and large
polyps are $10 mm in diameter. Polyps >5 mm are consid-
ered clinically significant.
METHODS
First, a search of secondary literature was performed,
includingevidence-based reviews and information systems
(eg, the Cochrane Collaboration) using the ‘‘evidence pyra-
mid’’ of Haynes (8). Next, a PubMed search of the primary
literature was undertaken using the Preview/Index tab and the
relevant Medical Subject Headings, including the terms
‘‘computed tomography colonography,’’ ‘‘colonoscopy,’’
‘‘screening,’’ and ‘‘polyp.’’ These terms were entered and
linked with the Boolean operators AND and OR, as described
by Staunton (Table 1) (10). Limits were then applied to ar-
ticles written in the English language, referring to humans,
and published in the past 5 years (to account for any recent
significant changes in the technology of CTC). Each of the
retrieved articles was assigned a level of evidence using the
National Health Service Centre for Evidence-Based
Medicine (at Oxford University) hierarchy of validity for
diagnostic studies (12) (Table 2). Both authors performed
searches independently and then adjusted their search
parameters to be most inclusive.
RESULTS
No useful information relative to the current clinical
question was identified from the search of the secondary lit-
erature and information systems. The search of the primary
literature retrieved 116 articles. Next, the abstracts were re-
viewed by both authors to determine which studies appeared
relevant to the clinical question, and the resulting 43 articles
were retrieved. Many studies that were retrieved from the
PICO search were not graded for level of evidence, because
565
EL-MARAGHI AND KIELAR Academic Radiology, Vol 16, No 5, May 2009
the study designs did not answer the specific PICO question.
The major reason for a study’s being rejected was the inclu-
sion of a high-risk population. Other discarded publications
evaluated the accuracy of low-dose techniques, patient pref-
erences regarding types of colon preparations, radiologists’
accuracy depending on software and reconstruction algo-
rithms used, and the impact of radiologists’ experience with
CTC on the accuracy of reports. After examining the papers,
it was determined that 16 dealt with screening populations (at
least in part), and these were evaluated in depth (11,13–27).
Using the Centre for Evidence-Based Medicine’s hierar-
chy of validity, there were three level 1c studies, two level 2a
studies, three level 2b studies, four level 3b studies, two level
4 studies, and two level 5 studies (Table 3).
From the retrieved articles, there were two meta-analyses
related to the use of CTC for CRC evaluation. The first was
a paper published in 2005 by Mulhall et al (13), which in-
cluded data from 33 prospective studies involving 6393 pa-
tients who had undergone CTC, with OC as the gold
standard. This publication was considered level 2a evidence.
Studies varied by reader experience, software platform (two-
dimensional vs three-dimensional vs both) and the type of
computed tomographic scanner (single slice vs multislice)
used. Only 26% of the total population represented average-
risk patients; as a result, one significant potential source of
bias was the prevalence of disease among the various study
populations. In patients at high risk, the expected prevalence
of disease would be higher. On the basis of the meta-analysis,
CTC was found to be very specific, particularly for polyps >9
mm, but sensitivity varied widely among studies. Reported
sensitivity for polyps 6 to 9 mm varied from 55% to 84%,
whereas for polyps >10 mm, sensitivity ranged from 79% to
91%. The authors stated that they were unable to completely
account for the wide variability in sensitivity among the
Table 1
PubMed Search Strategy Using PICO-focused Keywords
Search
Step Search Criteria
Number of
Retrieved Citations
1 Search ‘‘colonography, computed
tomographic’’ (all fields)
876
2 Search ‘‘colonoscopy’’ (all fields) 17,823
3 Search ‘‘colonic polyp’’ (all fields) 9354
4 Search ‘‘colorectal neoplasms/
diagnosis’’ (all fields)
4575
5 Search ‘‘colonic neoplasms/diagnosis’’
(all fields)
5164
6 Search [(#3) OR (#4)] OR (#5) 17,454
7 Search [(#1) AND (#2)] AND (#6) 360
8 Search [(#1) AND (#2)] AND (#6) 116
PICO, patient, intervention, comparison intervention, outcome.
Results were limited to studies published in the past 5 years,
involving humans, and in English.
566
reviewed studies, and as a result, they suggested that CTC
should be investigated further before being used in the gen-
eral population for screening. One suggested explanation for
the different sensitivities related to the slice thickness of the
acquired images; on the basis of a metaregression of data
from 19 of the included studies, the authors found that
for every 1 mm increase in collimation width, sensitivity
decreased by 4.9%. Studies also dated back to 1997, when
hardware and software technology were inferior to current
technology.
The second systematic review, authored by Halligan et al
(14) in 2005, included 24 studies published between 1994
and 2003 and included 2610 patients. This meta-analysis
demonstrated high sensitivity and specificity for detecting
polyps >10 mm. Per patient average sensitivity was 93%, and
specificity was 97% for CTC. When all clinically relevant
polyps (ie, >5 mm) were included in the analysis, the sensi-
tivity and specificity both decreased to 86%. Halligan et al
also reported a sensitivity of 96% for detecting the 150 can-
cers diagnosed in the patients included in the meta-anaslysis.
As with the meta-analysis by Mulhall et al (13), most of the
studies in this analysis, except the paper by Pickhardt et al
(15) in 2003, included a majority of higher risk patients and
as such may have increased the apparent sensitivity because
of a higher prevalence of polyps. Despite this, the authors felt
that their analysis suggested that CTC has high average
sensitivity and specificity for clinically relevant polyps but
that more trials are required in asymptomatic, low-risk,
screening populations. This information was also considered
level 2a evidence.
The recently published American College of Radiology
Imaging Network National CT Colonography (ACRIN) trial
by Johnson et al (16) is the largest trial undertaken, involving
a screening population of 2600 patients recruited at 15 cen-
ters in the United States. All patients were aged >50 years and
believed to be at average risk for colon cancer. All radiolo-
gists had prior experience reading CTC and were selected for
participation on the basis of competitive performance with
a testing set of CTC studies. The per patient sensitivity of
CTC for detecting polyps $6 mm was 78%, and it was 90%
for polyps $10 mm. Interestingly, there was no difference in
accuracy of CTC with the use of either primary two-dimen-
sional software (with three-dimensional reconstructions for
problem solving) or three-dimensional endoluminal fly-
though software. Given the strength of the study, this was
classified as level 1c evidence.
The publication in 2003 by Pickhardt et al (15) is often
quoted as a landmark study. In this prospective, multicenter
trial, there were 1233 patients, of whom >97% were deemed
at average risk. Segmental unblinding was used, which
means that during the retraction phase of the colonoscopy,
the findings from CTC were revealed to the gastroenterolo-
gist. All radiologists had read >25 scans prior to beginning
Academic Radiology, Vol 16, No 5, May 2009 CTC: A PICO ANALYSIS
Table 2
Levels of Evidence as Adapted from the Oxford Centre for Evidence-Based Medicine
Levels of Evidence Diagnostic Tests
1a Systematic review with homogeneity of level 1 diagnostic studies or clinical decision rule with 1b studies from different
clinical institutions
1b Cohort study with good reference standards and validation of clinical decision rule tested within a single clinical
institution
1c Studies with results that have very high sensitivity and/or specificity, such that a positive result rules in the diagnosis
anda negative result rules out the diagnosis
2a Systematic review of level 2 diagnostic studies with homogeneity
2b Exploratory cohort study with good reference standards looking for significant factors or clinical decision rule after
derivation or validated only on split samples or databases
3a Systematic review including 3b or studies of higher levels of evidence
3b Non-consecutive study or study without a consistently applied reference standard
4 Case-control study with either poor or non-independent reference standards
5 Expert opinion
Table 3
Summary of CT Colonography Compared to Colonoscopy Results with Levels of Evidence for Articles retrieved from PubMed Search
Using PICO-focused Keywords
Reference Year N
Sensitivity
(%)
Specificity
(%)
PPV
(%)
NPV
(%)
CT
Scanner
Level of
Evidence
Johnson et al. (16)y 2008 2531 78–90 86–88 23–40 98–99 16 slice 1c
Pickhardt et al. (15)yz 2003 1233 89–94 80–96 — — 4,8 slice 1c
Iannoccone et al. (17)y 2003 158 96 97 94 98 4 slice 1c
Mulhall et al. (13)y 2005 — 48–85 92–97 — — — 2a
Halligan et al. (14) 2005 — — — — — — 2a
Pickhardt et al. (11)* 2004 1233 86–92 — — — 4,8 slice 2b
Iannoccone et al. (18)* 2005 88 86–100 82–100 70–100 91–100 4 slice 2b
Vogt et al. (19)* 2004 115 91–100 82–83 — — 4 slice 2b
Copel et al. (20)y 2007 546 — — 33–65 — 4,8 slice 3b
Yun et al. (21)* 2007 113 89–91 — 76–87 — 16 slice 3b
Wessling et al. (22) 2005 78 81–100 86 — — 4 slice 3b
Macari et al. (23)* 2004 68 53–100 90D — — 4 slice 3b
Edwards et al. (24) 2004 93 — — 73 — 1 slice 4
Kim et al. (25) 2007 246 — — — — — 4
Rozen (26) 2006 — — — — — — 5
Rex (27) 2005 — — — — — — 5
CT, computed tomographic; FN, false negatives; FP, false positives; NPV, negative predictive value; PPV, positive predictive value; TN, true
negatives; TP, true positives.
* Per-polyp results.
y Per-patient results
z Studies using segmental unblinding.
the study. The sensitivity of CTC in detecting polyps >10 mm
was higher than that of OC, with results of 94% for CTC
compared to 89% for OC. The k value for the detection of
polyps >8 mm was 0.79, indicating good interobserver
agreement. On the basis of the size of the study and the
strength of the results, this was classified as level 1c evidence.
A study by Iannaccone et al (17) in 2003 evaluated a low-
dose technique to perform CTC. The authors examined 158
patients (although only 31 were referred for routine screen-
ing) and reported per patient sensitivity of identifying polyps
>5 mm of 96% and specificity of 97%. The simulated ef-
fective doses for CTC using their technique were 1.8 mSv in
men and 2.4 mSv in women. Although only a portion of the
study participants were a screening population, on the basis
of the strength of the results, the study was classified as level
1c evidence.
In a study by Pickhardt et al (11) in 2004, 1233 asymp-
tomatic adults underwent CTC followed by same-day OC as
567
EL-MARAGHI AND KIELAR Academic Radiology, Vol 16, No 5, May 2009
the gold standard comparator test. The per polyp sensitivity
of CTC was described as 86% for polyps >6 mm and 92%
for those >10 mm. The paper was classified as level 2b
evidence.
Iannaccone et al (18) performed a prospective trial in 2005
with a mixed population of low- and high-risk patients using
a low-dose computed tomographic colonographic protocol
and two sequential OC procedures. For all polyps >6 mm, per
polyp sensitivity was 86% for CTC and 84% for initial OC.
For polyps >10 mm, sensitivity was 100% for CTC and only
91% for initial OC. Initial OC failed to detect 16 polyps, six
of which were correctly identified on CTC. The findings in
this study suggested that OC may have a substantial miss rate
for identifying polyps. This was classified as level 2b evi-
dence.
A study published in 2004 by Vogt et al (19) reported use
of an ultra-low-dose technique with a total radiation dose of
only 1.14 mSv. This patient population was referred for
nonspecific abdominal symptoms such as chronic abdominal
pain and constipation. Sensitivity for polyps <5 mm was
76%, for polyps 6 to 9 mm was 91%, and for polyps >10 mm
was 100%. For all lesions that were ultimately demonstrated
to represent adenomas and were >5 mm, the sensitivity was
94% and the specificity 92%. This was classified as level 2b
evidence.
A single-institution, retrospective study by Copel et al
(20) in 2007 involved patients who were referred for CTC
because of incomplete OC. Causes of incomplete colono-
scopy included a redundant colon, colonic spasm, divertic-
ulosis, obstructing masses, and sharp angles due to previous
surgery with strictures. Of 546 patients, only 54 (9.9%) were
considered at average risk. The per patient positive predictive
value was 33% for lesions 6 to 9 mm and 65% for lesions 10
to 20 mm. This was considered level 3b evidence.
Yun et al (21) retrospectively analyzed findings CTC
compared to OC in 113 patients who had previously under-
gone both procedures from a total pool of 399 patients. The
authors also assessed the technique of fecal tagging for the
evaluation of CTC’s diagnostic performance. Interestingly,
Yun et al (21) found that the sensitivity for fecal tagging was
lower than that with no fecal tagging (82% vs 96% for polyps
>10 mm). Reasons for false-negative results included inter-
pretation errors, insufficient bowel-cleansing preparation
methods, a lack of optimal bowel distention, and morpho-
logic characteristics of the lesions, such as flat lesions. The
authors suggested that intravenous contrast enhancement
may be more helpful for detecting polyps than fecal tagging,
although they did not do an in-depth analysis comparing
these two techniques. This was considered level 3b evidence.
In 2005, Wessling et al (22) evaluated a small group of 26
patients, 83% of whom were asymptomatic. The authors
found 49 polyps using OC. CTC detected all three cancers, all
seven polyps >10 mm, 13 of 16 polyps (81%) that were 6 to 9
568
mm, and 19 of 26 polyps (73%) that were <5 mm. There were
14 false-positive findings on CTC: 10 polyps were <5 mm,
and the remaining four were later proved to be true-positive
results and had actually been missed at the time of the initial
OC. Per patient specificity for detecting all polyps was 86%.
This was classified as level 3b evidence.
Macari et al (23) published a study in 68 men from
a Veterans Affairs hospital. Each average-risk patient un-
derwent CTC and OC in a blinded fashion. Segmental un-
blinding was not used. During segmental unblinding, the
results of the CTC are provided to the optical colonoscopist
so that the areas of interest detected on CTC may be more
carefully evaluated during withdrawal of the scope. Per pa-
tient sensitivity of CTC was 90%. Per polyp sensitivity for
polyps 6 to 9 mm was only 53%, but it was 100% for the three
polyps >10 mm in size detected in this study population. This
was classified as level 3b evidence.
Edwards et al (24) evaluated a group of 1452 asymp-
tomatic patients using CTC. Only participants with positive
findings went on to have OC. Given that only this selected
group of patients underwent both tests (including OC, which
is considered the gold standard), this was classified as level 4
evidence. Part of the goals of the study was to determine
factors that might improve the acceptability of CTC as
a screening method for patients. One particular method in-
cluded looking at fecal tagging rather than traditional colon
preparations with computer software that digitally subtracts
the identified stool to allow for better visualization of the
colon.
In 2007, Kim et al (25) compared two cohorts of >3000
patients each, 98% of whom were asymptomatic; approxi-
mately 5% to 8% in each group had family histories of CRC.
From the original 3120 patients in the CTC group, 13% (404
patients) had positive findings on CTC. Of this subgroup, 246
were secondarily referred for same-day OC because of find-
ings of polyps>6 mm, whereas the other 158 patients chose
to be followed with serial imaging. A total of 100 patients
originally in CTC group had advanced neoplasia confirmed,
whereas there were 107 patients with similar outcomes in the
primary OC group (from 3163 patients initially). The authors
suggested that primary CTC with selective OC may be
a reasonable screening strategy, because it achieves similar
detection and outcomes but with less use of OC resources. In
this study, the majority of patients in the CTC group did not
undergo confirmatory OC, and therefore, the study was
considered level 4 evidence.
The work published by Rozen (26) in 2005 was a sum-
mary of expert opinions that had been presented at an inter-
national CRC screening meeting. This was classified as level
5 evidence.
Rex (27) published an opinion regarding the use of re-
ferring all patients with any size polyp detected on CTC to
OC. His proffered opinion was that polypectomy might be
Academic Radiology, Vol 16, No 5, May 2009 CTC: A PICO ANALYSIS
a reasonable option in healthy patients with intermediate-size
polyps and that monitoring patients with repeated CTC could
potentially be more costly and expose them to radiation
exposure risks. This was considered level 5 evidence.
DISCUSSION
Over the past 20 years, there has been a move toward
using evidence-based clinical decision making. However, to
more easily identify pertinent information, more efficient
ways of searching the medical literature are evolving to help
physicians locate and review data relevant to their specific
queries. The Centre for Evidence-Based Medicine at the
University of Oxford has developed a strategy for assigning
levels of evidence to the medical literature that hierarchically
classifies articles according to the strength of their study
designs (Table 2) (12). As a result, more emphasis may be
placed on reviewing data with higher levels of evidence.
On the basis of the PICO scenario, studies were chosen
and reviewed if they reflected a screening population of
average-risk individuals. All articles were then assessed for
‘‘validity’’ by determining if there was an independent
blinded comparison to a reference standard. In this case, CTC
was compared to OC, which was considered to be the gold-
standard test. In addition, the studies were evaluated to de-
termine if they had been described in enough detail that they
could be independently reproduced. This point is important,
because there have been many changes in computed tomo-
graphic scanning technology and reader software platforms
over the past several years.
CRC remains an important health concern in North
America. Data have been published since the 1990s demon-
strating efficacy of screening programs in reducing mortality
from colon cancer (28,29). Costs per life-year saved of other
types of population-based screening tests are similar to CRC
screening, varying between $9000 and $25,000 for screening
mammography, cervical cancer, and hypertension screening
(30,31).
A few economic evaluations of CTC have been per-
formed. A study by Ladabaum et al (32) in 2004 suggested
that OC would be preferred over CTC if sensitivities were
determined to be similar, unless CTC cost significantly less.
Another, somewhat controversial study by Heitman (33) in
2005 suggested that CTC would be very costly as a screening
test. Although CTC could potentially avoid some complica-
tions of OC, such as colonic perforation, overall, it was be-
lieved that it would likely lead to a small increase in mortality
from missed cancers. Some letters to the editor, written as
rebuttals to this argument, indicated that with improving
technology, the sensitivity of CTC approaches that of OC
(34,35).
A study published by Pickhardt et al (36) in 2007 indi-
cated that the cost-effectiveness of CTC is most obvious
when diminutive lesions, which are not believed to be clin-
ically relevant, are neither reported nor acted on, thereby
reducing the number of referrals for OC. In addition, removal
of these small lesions (<5 mm) was believed to be associated
with an unreasonably high cost and increased potential for
patient complications. The stated transformation rate of
polyps <5 mm to intermediate size (6–9 mm) and from this
second category to >10 mm is approximately 2% annually.
The transition from a polyp >10 mm to a carcinoma is ap-
proximately 3% per year (37). Given this, it may reasonable
to consider short-term follow-up CTC as an alternative, but
cumulative radiation doses need to be taken into account. In
an earlier study by Pickhardt et al (38), CTC was found to be
less sensitive than OC for detecting lesions of similar size but
that have no malignant potential compared to their
adenomatous counterparts. This may actually be a positive
attribute of CTC, because it may reduce the number of
false-positive assessments that would necessitate follow-up
OC.
Although OC is considered the gold standard for detecting
polyps, there is reported variability regarding completion
rates of OC, with up to 4% to 25% of studies being incom-
plete (20,39). Furthermore, between 10% and 20% of colonic
polyps and 5% of cancers may be missed by traditional OC
(40,41). Thus, when using OC as the gold standard for
evaluating the effectiveness of CTC, it is conceivable that the
true specificity and positive predictive value of CTC may be
higher than those reported.
CTC is theoretically superior to OC for evaluation of the
bowel proximal to areas of stenosis or obstruction. Addi-
tionally, it may be more easily performed in elderly and
frail patients, in whom mobility or sedation might be
problematic, and in patients who have allergies to medi-
cations often used for conscious sedation (23). Moreover,
multidetector computed tomography also has an advantage
over OC in that it has the ability to better localize lesions in
the specific segments of the colon, because the area in
question can more easily be correlated to surrounding
structures (42).
Test characteristics of CTC that make it desirable for ap-
plication to a screening population include its relative safety,
patient acceptance, avoidance of the need for sedation, and
ability to concurrently assess other intra-abdominal and pel-
vic organs (43–45). Beebe et al (46) surveyed patient pref-
erences to bowel preparations required for CRC screening.
The conclusion was that bowel preparation in general is
a major disincentive to complying with screening programs.
It is hypothesized that alternative procedures, such as fecal
tagging prior to CTC, which would obviate the need for
bowel cleansing, may be better accepted by patients in
a screening population.
569
EL-MARAGHI AND KIELAR Academic Radiology, Vol 16, No 5, May 2009
Related to this topic, a study by Iannaccone et al (47)
showed that CTC performed without the use of cathartics had
sensitivity of up to 96% for polyps >8 mm. However, Van
Gelder et al (48) published a study in 2004 in a high-risk
population analyzing patient preference and found that im-
mediately after the procedure, patients indicated that they
preferred CTC but that over time, the difference in preference
became less pronounced, whereas outcome measures (such
as finding and removing polyps) gradually replaced the per-
ceived inconveniences of OC.
A major concern with CTC is lifetime radiation exposure,
and thus the development of low-dose techniques is an im-
portant consideration. The effective dose for standard CTC
varies between 5 and 25 mSv. In comparison, a double-
contrast barium enema is typically associated with a radiation
dose of between 3 and 7 mSv (49). At these radiation doses,
for a patient aged >50 years, the added risk for inducing
cancer from iatrogenic causes is approximately 0.02%. As an
example, if one assumes an average radiation dose of 10
mSv/scan and if used for screening patients every 3 years, the
risk for exposure-related deathcould be as high as 0.3% for
men and 0.4% for women if screening begins at age 50 years
(50). In principle, decreasing radiation dose increases noise
and in turn reduces the signal-to-noise ratio and degrades
image quality. However, newer software is better able to
handle the input signal and create sharper, clearer images.
CTC continues to rapidly evolve both in image acquisition
and software rendering, including three-dimensional display
techniques. These factors, along with technological ad-
vancements in hardware, greatly enhance image quality and
reader accuracy for CTC. Johnson et al (51) performed
a retrospective analysis of readers with variable degrees of
experience and different software platforms to identify fac-
tors that may affect the accuracy of CTC. Previous experi-
ence reading CTC was found to have a positive association
with accuracy of reports, whereas familiarity with the soft-
ware platform did not lead to a statistically significant dif-
ference. In an earlier study by the same lead author in 2003
(43), reader experience and familiarity with software were
again the main factors evaluated. In this cohort, patients un-
derwent both CTC and OC. The result was that both reader
experience and familiarity with the software being used were
important factors rather than the specific type of software
used.
CONCLUSIONS
In the first ever joint consensus guidelines for CRC
screening released jointly by the American Cancer Society,
the American College of Radiology, and the Multi-Society
Task Force on Colorectal Cancer (a group that comprises
representatives from the American College of Gastroenter-
570
ology, the American Gastroenterological Association, and
the American Society for Gastrointestinal Endoscopy), CTC
was included as one of several possible CRC screening op-
tions in average-risk adults aged $50 years and was recom-
mended to be performed once every 5 years. The majority of
evidence suggests that CTC is an acceptable alternative to
OC, particularly in the group of patients who are either un-
willing or unable to undergo OC. A reasonable strategy,
therefore, may be to consider CTC with selective OC in pa-
tients found to have clinically meaningful polyps, because
this would appear to satisfy the goals of screening and pre-
vention while minimizing the use of OC resources.
REFERENCES
1. Rabeneck L. Population-based colorectal cancer screening; an overview
of Ontario’s program. Oncol Exchange 2007; 6:25–27.
2. Quintero E, Parra-Blanco A. Noninvasive diagnostic tools in colorectal
cancer mass screening. Curr Colorect Cancer Rep 2007; 3:29–34.
3. Swan J, Breen N, Coates RJ, et al. Progress in cancer screening prac-
tices in the United States: results from the 2000 National Health Interview
Survey. Cancer 2003; 97:1528–1540.
4. Umar S, Richmond E, Griebel DJ. Colorectal cancer prevention: diet,
drugs or nothing. Curr Colorect Cancer Rep 2007; 3:16–23.
5. McAlearney AS, Reeves KW, Kickinson SL, et al. Racial differences in
colorectal cancer screening practices and knowledge within a low-in-
come population. Cancer 2008; 112:391–398.
6. Shih YC, Zhao L, Elting LS. Does Medicare coverage of colonoscopy
reduce racial/ethnic disparities in cancer screening among the elderly?
Health Aff (Millwood) 2006; 25:1153–1162.
7. Vining DJ, Gelfand DW, Bechtold RE, Scharling ES, Grishaw EK,
Shifrin GY. Technical feasibility of colon imaging with helical CT and
virtual reality [abstract]. AJR Am J Roentgenol 1994; 162(suppl):S104.
8. Haynes B. Of studies, syntheses, synopses, summaries, and systems:
the ‘‘5S’’ evolution of information services for evidence-based healthcare
decisions. Evid Based Nurs 2007; 10:6–7.
9. Dodd GD. Evidence-based practice in radiology: steps 3 and 4—
appraise and apply diagnostic radiology literature. Radiology 2007; 242:
342–353.
10. Staunton M. Evidence-based practice in radiology: steps 1 and 2—ask-
ing answerable questions and searching for evidence. Radiology 2007;
242:23–31.
11. Pickhardt PJ, Choi JR, Nugent PA, Schindler WR. The effect of diagnostic
confidence on the probability of optical colonoscopic confirmation of
potential polyps detected on CT colonography: prospective assessment
in 1,339 asymptomatic adults. AJR Am J Roentgenol 2004; 183:
1661–1665.
12. National Health Service Centre for Evidence-Based Medicine. Levels of
evidence and grades of recommendations. Available at: http://www.
cebm.net/level_of_evidence.asp.
13. Mulhall BP, Veerappan GR, Jackson JL. Meta-analysis: computed
tomographic colonography. Ann Intern Med 2005; 142:635–650.
14. Halligan S, Altman DG, Taylor SA, et al. CT colonography in the detection
of colorectal polyps and cancer: systematic review, meta-analysis, and
proposed minimum data set for study level reporting. Radiology 2005;
237:893–904.
15. Pickhardt PJ, Choi JR, Hwang I, Butler JA, et al. Computed tomographic
virtual colonoscopy to screen for colorectal neoplasia in asymptomatic
adults. N Engl J Med 2003; 349:2191–2199.
16. Johnson CD, Chen M-H, Tolendano AY, et al. Accuracy of CT colonog-
raphy for detection of large adenomas and cancers. N Engl J Med 2008;
359:1207–1217.
17. Iannaccone R, Laghi A, Catalano C, et al. Detection of colerectal lesions:
lower-dose multi-detector row helical CT colonography compared with
conventional colonoscopy. Radiology 2003; 229:775–781.
http://www.cebm.net/level_of_evidence.asp
http://www.cebm.net/level_of_evidence.asp
Academic Radiology, Vol 16, No 5, May 2009 CTC: A PICO ANALYSIS
18. Iannaccone R, Catalano C, Mangiapane F, et al. Colorectal polyps:
detection with low-dose multi-detector row helical CT colonography
versus two sequential colonoscopies. Radiology 2005; 237:927–937.
19. Vogt C, Cohnene M, Beck A, et al. Detection of colorectal polyps by
multislice CT colonography with ultra-low-dose technique: comparison
with high-resolution videocolonoscopy. Gastrointest Endosc 2004; 60:
201–209.
20. Copel L, Sosna J, Kruskal JB, Raptopoulos V, Farrel RJ, Morrin MM. CT
colonography in 546 patients with incomplete colonoscopy. Radiology
2007; 244:471–478.
21. Yun J, Ro HJ, Park JB, et al. Diagnostic performance of CT colonography
for the detection of colorectal polyps. Korean J Radiol 2007; 8:484–491.
22. Wessling J, Domagk K, Lugenring N, et al. Virtual colonography: identi-
fication and differentiation of colorectal lesions using multi-detector
computed tomography. Scand J Gastroenterol 2005; 40:468–476.
23. Macari M, Bini E, Jacobs S, et al. Colorectal polyp and cancers in
asymptomatic average-risk patients: evaluation with CT colonography.
Radiology 2004; 230:629–636.
24. Edwards JT, Mendelson RM, Fritschi L, et al. Colorectal neoplasia
screening with CT colonography in average-risk asymptomatic subjects:
community-based study. Radiology 2004; 230:459–464.
25. Kim DH, Pickhardt PJ, Hoff G, Kay CL. Computed tomographic colo-
nography for colorectal screening. Endoscopy 2007; 39:545–549.
26. Rozen P. Report of the World Organization for Digestive Endoscopy
Colorectal Cancer Screening Committee Meeting, Chicago, 2005. Eur
J Cancer Prevent 2006; 15:95–102.
27. Rex DK. Colonoscopy is justified for any polyp discovered during
computed tomographic colonography. Am J Gastroenterol 2005; 100:
1903–1908.
28. Mandel JS, Bond JH, Church TR. Reducing mortality from colorectal
cancer by screening for fecal occult blood. N Engl J Med 1993; 328:
1365–1371.
29. Kronborg O, Fenger C, Olsen J, Jorgensen OD, Sondergaard O. Rand-
omised study of screening for colorectal cancer with fecal occult-blood
test. Lancet 1996; 348:1647–1671.
30. Garber AM, Phelps CE. Economic foundation of cost effectiveness
analysis. J Health Econ 1997; 16:1–31.
31. Brown ML, Fintor L. Cost-effectiveness of breast cancer screening:
preliminary results of a systematic review of the literature. Breast Cancer
Res Treat 1993; 25:113–118.
32. Ladabaum U, Song K, Fendrick AM. Colorectal neoplasia screeningwith
virtual colonoscopy: when, at what cost, and with what national impact?
Clin Gastroenterol Hepatol 2004; 2:554–563.
33. Heitman SJ, Manns BJ, Hilsden RJ, et al. Cost-effectiveness of com-
puterized tomographic colonography versus colonoscopy for colorectal
cancer screening. CMAJ 2005; 173:877–881.
34. Kiberd B. Colon cancer screening. CMAJ 2006; 174:975.
35. Peltekian K. Colon cancer screening. CMAJ 2006; 174:975.
36. Pickhardt PJ, Hassan C, Laghi A, Zullo A, Kim DH, Iafrate F, Morini S.
Cost-effectiveness of colorectal cancer screening with computed
tomography colonography: the impact of not reporting diminutive
lesions. Cancer 2007; 109:2213–2221.
37. Pickhardt PJ, Hassan C, Laghi A, et al. Small and diminutive polyps
detected at screening CT colonography: a decision analysis for referral to
colonoscopy. AJR Am J Roentgenol 2008; 190:136–144.
38. Pickhardt PJ. By-patient performance characteristics of CT colonogra-
phy: importance of polyp size threshold data. Radiology 2003; 229:
291–293.
39. Marshall JB, Barthel JS. The frequency of total colonoscopy and terminal
ileum intubation in the 1990s. Gastrointest Endosc 1993; 39:518–520.
40. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by
colonoscopic polypectomy. N Engl J Med 1993; 329:1977–1981.
41. Hixson LJ, Fennerty MB, Sampliner RE, Garewal HS. Prospective blinded
trial of the colonoscopic miss-rate of large colorectal polyps. Gastroint-
est Endosc 1991; 37:125–127.
42. Chung DJ, Huh KC, Choi WJ, Kim JK. CT colonography using 16-MDCT in
the evaluation of colorectal cancer. AJR Am J Roentgenol 2005;184:98–103.
43. Johnson CD, Harmsen WS, Wilson LA, et al. Prospective blinded evalu-
ation of computed tomographic colonography for screen detection of
colorectal polyps. Gastroenterology 2003; 125:311–319.
44. Gluecker TM, Johnson CD, Harmsen WS, et al. Colorectal cancer
screening with CT colonography, colonoscopy, and double-contrast
barium enema examination: prospective assessment of patient percep-
tions and preferences. Radiology 2003; 227:378–384.
45. Hara AK, Johnson CD, MacCarty RL, Welch TJ. Incidental extracolonic
findings at CT colonography. Radiology 2000; 215:353–357.
46. Beebe TJ, Johnson CD, Stoner SM, Anderson KJ, Limburg PJ. Assessing
attitudes toward laxative preparation in colorectal cancer screening and
effects on future testing: potential receptivity to computed tomographic
colonography. Mayo Clin Proc 2007; 82:666–671.
47. Iannaccone R, Laghi A, Catalano C, et al. CT colonography without
cathartic preparation for the detection of colorectal polyps. Gastroen-
terology 2004; 127:1300–1311.
48. Van Gelder RE, Nio CY, Florie J, et al. Computed tomographic colo-
nography compared with colonoscopy in patients at increased risk for
colorectal cancer. Gastroenterology 2004; 127:41–48.
49. Huda W. Radiation dosimetry in diagnostic radiology. AJR Am J Roent-
genol 1997; 169:1487–1488.
50. Wise KN. Solid cancer risks from radiation exposure for the Australian
population. Australas Phys Eng Sci Med 2003; 26:53–62.
51. Johnson CD, Toledano AY, Herman BA, et al. American College of
Radiology Imaging Network A6656. Computerized tomographic colo-
nography: performance evaluation in a retrospective multicenter setting.
Gastroenterology 2003; 12:688–695.
571
	CT Colonography Versus Optical Colonoscopy fornbspScreening Asymptomatic Patients for Colorectal Cancer
	Clinical scenario
	Methods
	Results
	Discussion
	Conclusions
	References