J. H.

Bond Colon Polyps and Cancer

State of the Art Review

A large number of studies published last year in peer-reviewed by preventing cancer, is becoming a primary objective of screen-
medical journals help to better define the advantages and limita- ing programs. Several papers also show the potential of emerg-
tions of the different options for colorectal cancer screening. Di- ing new methods of screening for specific markers in stool and
rect colonoscopy screening appears to have the greatest poten- for imaging the colon with computed-tomographic colonogra-

Downloaded by: University of British Columbia. Copyrighted material.

tial to markedly reduce both the incidence and mortality of colo- phy (virtual colonoscopy). Other important publications high-
rectal cancer, but many obstacles limiting its widespread use in lighted in this review deal with the diagnosis of colorectal neo-
the general at-risk population still exist, and many questions re- plasia, familial colorectal cancer, colorectal polyps and the ade-
main incompletely answered. Recent studies stress the fact that noma–carcinoma sequence, and new and novel methods of im-
finding and resecting advanced adenomatous polyps, and there- proving the efficiency and safety of colonoscopic polypectomy.

Screening for Colorectal Cancer strated in a randomized controlled trial, but considerable indi-
rect evidence supports its potential for reducing the incidence
Current evidence-based guidelines in the USA recommend that and mortality of colorectal cancer. Although many experts and
all asymptomatic, average-risk individuals be offered screening professional groups are now recommending direct colonoscopy 27
for colorectal cancer, beginning at age 50 years [1, 2]. Rather as the “preferred” method of screening, several important ques-
than recommending a single method of screening, they provide tions have not yet been fully addressed. These include questions
a menu of five different, acceptable options: annual fecal occult of risk, cost, compliance, available resources, and capacity. The
blood tests (FOBTs), flexible sigmoidoscopy every 5 years, the US Veterans Affairs (VA) Cooperative Colonoscopy Screening
combination of FOBTs and flexible sigmoidoscopy, air-contrast Trial, the oldest and largest screening colonoscopy experience to
barium enema every 5 years, or colonoscopy every 10 years. date, has provided some answers [3]. Direct screening colonos-
Physicians and patients are urged to select one of these screening copy was performed in 3121 asymptomatic volunteers aged
strategies in a shared decision-making process, depending on 50 – 75 years who had not undergone examination of the colon
available resources, the make-up of each medical delivery sys- within the prior 10 years. These volunteers were recruited from
tem, and the patient’s wishes. The main objectives of screening the general medicine clinics of 13 VA medical centers, and colo-
are not only to detect early surgically curable cancer, but increas- noscopy was performed by study co-investigators, with all de-
ingly also to prevent cancer by the detection and resection of ad- tected polyps being measured, photographed, and removed. Pa-
vanced adenomatous polyps. tients were contacted after 24 h and 1 week to track all proce-
dure-related complications. Colonoscopy was complete to the
Of the five screening options, direct colonoscopy screening has cecum in 97.2 % of cases, and the mean insertion time to the ce-
the greatest potential for accomplishing both of these objectives. cum and total procedure times were 10.5 and 30.6 min, respec-
The efficacy of screening colonoscopy has not yet been demon- tively. No preprocedural patient characteristics were identified

Institution
Gastroenterology Section, Minneapolis Veterans Affairs Medical Center,
University of Minnesota, Minneapolis, Minnesota, USA

Corresponding Author
J. H. Bond, M.D. · University of Minnesota · Chief, Gastroenterology Section (111D) · VA Medical Center ·
One Veterans Drive · Minneapolis, MN 55417 · USA · Fax: + 1-612-725-2248 · E-mail: [email protected]

Bibliography
Endoscopy 2003; 35 (1): 27–35 H Georg Thieme Verlag Stuttgart · New York · ISSN 0013-726X
 that were predictive of an incomplete examination. Although at adenomas detected during colonoscopy in one endoscopy unit in
least one polyp was resected in 1672 patients (54 %), there were Perth [6]. Endoscopy reports of 2578 patients were reviewed. Of
no perforations and no deaths attributed to colonoscopy. Major all the adenomas detected, 44 % were left-sided only and 24.5 %
complications (mainly bleeding after polypectomy) occurred in were right-sided only. Carcinoma was observed in 7 % of cases,
nine cases (0.3 %). Only one complication (a major cerebrovascu- of which 37.5 % were left-sided only. There was an increased
lar event) occurred in subjects undergoing an examination that right-sided prevalence of both adenomas and carcinomas with
was only diagnostic. The authors concluded that screening colo- age. The author concluded that examination by flexible sigmoi-
noscopy can be performed in multiple centers with a high degree doscopy performed to the splenic flexure would miss 23 % of
of success and safely, in large numbers of asymptomatic average- colorectal neoplasms, and that examination of the proximal co-
risk people. lon becomes more important in older people. A colonoscopy se-
ries from Tokyo also reported that the proportion of patients
The accuracy and safety of colonoscopy depend on the expertise with right-sided colorectal cancer increased progressively with
of the endoscopists performing the examination. In order to as- age [7]. After the age of 70, over half of all cancers would be miss-
sess the rate of complications in an average community practice, ed if sigmoidoscopy alone were used for screening. Lastly, stud-
State of the Art Review

a study in Sweden retrospectively measured the complication ies indicate that there also is a right-sided shift of metachronous
rate of diagnostic and therapeutic colonoscopies performed by adenomas found in patients undergoing postpolypectomy sur-
community-based endoscopists in one county [4]. A total of veillance colonoscopy. In a study in Erlangen, 51 % of 556 patients
6066 colonoscopies were performed between 1979 and 1995. had metachronous adenomas detected during follow-up postpo-
The overall morbidity was 0.4 % (diagnostic 0.2 %, therapeutic lypectomy surveillance over an 18-year period [8]. In 37.9 % of
1.2 %). The most frequent major complications were bleeding these patients, there was a right-sided shift in the first genera-
(0.2 %) and perforation (0.1 %); there was no colonoscopy-related tion of metachronous adenomas in comparison with the location

Downloaded by: University of British Columbia. Copyrighted material.

mortality. These authors concluded that colonoscopy is a safe of the index adenoma. Proximally located first-generation meta-
procedure, and the rate of adverse events in this population- chronous adenomas would be missed by flexible sigmoidoscopy
based setting also was low. in 50.8 % of patients who had only distal adenomas at the base-
line colonoscopy. The authors concluded that total colonoscopy
In the VA Colonoscopy Trial, performed mainly in a male popula- is needed for postpolypectomy surveillance, regardless of the in-
tion, 10.6 % of patients were found to have one or more advanced itial adenoma site.
colorectal neoplasms (advanced adenomas 9.6 %, invasive cancer
1 %). Advanced adenomas were defined as those ‡ 1 cm or con- Compliance with the different screening options requires further
taining villous tissue or high-grade dysplasia. In order to esti- study before the overall success of a given screening program can
mate the relative sensitivity of different screening options, all be predicted. A group in California assessed patient satisfaction
subjects were asked to submit stool specimens on Hemoccult and acceptance after flexible sigmoidoscopy, air-contrast barium
28 (Beckman-Coulter, Inc., Palo Alto, California, USA) cards from enema (ACBE), and colonoscopy [9]. A total of 410 patients com-
three consecutive days for fecal occult blood testing prior to un- pleted questionnaires within 48 hours of an examination (80
dergoing screening colonoscopy [5]. The cards were rehydrated ACBE, 202 sigmoidoscopy, 128 colonoscopy). They reported that
prior to developing. The yield of sigmoidoscopy screening was sigmoidoscopy was more painful than other colonic imaging
defined as examination of the rectum and sigmoid colon during modalities. Although ACBE and colonoscopy caused similar dis-
colonoscopy. A total of 23.9 % of subjects found to have advanced comfort, patients were less willing to repeat ACBE. In aggregate,
neoplasia during colonoscopy had a positive test for fecal occult the data suggested that patients perceive colonoscopy as the
blood. As compared with subjects who had a negative FOBT, the most acceptable procedure. Although a number of recent surveys
relative risk (RR) of advanced neoplasia in those with a positive suggest that the level of colonoscopy screening in the commu-
FOBT was 3.47. Sigmoidoscopy identified 70.3 % of all subjects nity remains very low, most of these assessments do not take
with advanced neoplasia. Combined one-time screening with into account the fraction of the at-risk population that has al-
FOBT plus sigmoidoscopy identified 75.8 % of subjects with ad- ready undergone a total colonic examination for any reason. In-
vanced neoplasia. Thus, this study showed that one-time screen- vestigators from Pennsylvania performed a population-based
ing with both FOBT with rehydration and sigmoidoscopy fails to survey to determine the prevalence of total colonic examinations
detect advanced colonic neoplasia in 24 % of subjects with the in a sample of people aged 50 – 79 years [10]. Out of 496 sur-
condition. The authors concluded that clinicians should not be veyed, of those without a personal or family history of colorectal
confident that advanced neoplasia has been ruled out when the neoplasia, 50 %, 19.6 %, 39.8 %, and 17.5 % reported ever having had
results of one-time combined FOBT and flexible sigmoidoscopy an FOBT, flexible sigmoidoscopy, barium enema, or colonoscopy,
screening are negative. While screening programs that employ respectively. Thirty-one percent reported having FOBT within
repeated testing with FOBT and sigmoidoscopy are likely to be the previous year or flexible sigmoidoscopy within the previous
more effective, direct colonoscopy screening appears to be the 5 years. Including total colonic examination within the previous
most accurate method of detecting most advanced neoplasia. 5 years increased the measured compliance to 39.7 %. Thus, in-
cluding total colonic examination performed for any reason sub-
Large colonoscopy series also report that a sizable fraction of stantially increases the measured level of screening compliance.
adenomas and carcinomas exist in the right colon of patients in
the absence of a synchronous left-sided neoplasm. These would All of the major colorectal cancer screening guidelines have
therefore be missed unless full colonoscopy was performed. For looked at the age–incidence curves for this disease and conclud-
example, a study analyzed the age-matched anatomic location of ed that screening of average-risk individuals should begin at

Bond JH. Colon Polyps and Cancer · Endoscopy 2003; 35: 27 – 35

age 50, the time in life when the incidence begins to accelerate. flexible sigmoidoscopy. A contemporaneous report from the
Others have argued that screening should be offered earlier – at United Kingdom FOBT Screening Trial showed similar results
age 40 or 45 years – in order detect most developing advanced [14]. After 11 years of follow-up, there was a 13 % reduction in
adenomatous polyps and early-age cancers. In order to better es- colorectal cancer mortality as the result of biennial screening, in
timate the cost-effectiveness of screening at an earlier age, a ret- spite of a compliance rate for the first invitation to screen of only
rospective, cross-sectional analysis was performed on data col- 50 %. The mortality reduction for those accepting screening (i. e.,
lected from 906 consecutive asymptomatic adults aged 40 – 49 compliers) was 27 %. The revised US screening guidelines recom-
years who underwent colonoscopic screening for the first time mend annual rather than biennial FOBT screening if this option is
as part of an employee-based colonoscopy screening program chosen, because yearly testing is more effective than screening
[11]. Of those screened, 78.9 % had no detected lesions, 10 % had every 2 years. The Minnesota Trial has estimated that the colo-
hyperplastic polyps, 8.7 % had tubular adenomas, and only 3.5 % rectal cancer mortality reduction in those who complied with
had advanced neoplasms, none of which were cancerous. The au- all rounds of screening in the annually screened group exceeded
thors determined that at least 250 persons, and perhaps 1000 or 40 % [12]. Short of doing direct colonoscopy screening for the en-
more, would need to be screened with colonoscopy to detect one tire at-risk population, the FOBT currently appears to be the best

State of the Art Review

cancer in this age group. They concluded that the low yield of available method of identifying asymptomatic, average-risk peo-
screening colonoscopy in this age group is consistent with cur- ple most likely to benefit from colonoscopy.
rent guidelines that recommend that asymptomatic, average-
risk people should not be offered colorectal cancer screening un- The option of flexible sigmoidoscopy screening has some obvious
til age 50. inherent advantages. The procedure is safe and well tolerated,
and it can be completed in 5 – 10 min without sedation after a
Of all the strategies one can choose to screen for colorectal can- simple preparation procedure. Flexible sigmoidoscopy is very ac-

Downloaded by: University of British Columbia. Copyrighted material.

cer, only the FOBT has been shown in randomized, controlled curate in that part of the colon within its reach – there are few
trials to be effective in reducing both the mortality and incidence false-positives or false-negatives. It reliably examines that part
of colorectal cancer [12]. Five long-term, controlled trials have of the colon in which 50 – 70 % of cancers and polyps occur. Al-
been carried out, involving nearly 350 000 participants (two in though cohort and case–control studies suggest that flexible sig-
the USA and three in Europe). The Minnesota Trial was the first moidoscopy screening reduces the mortality from distal colorec-
of these to report definitive mortality end-point results. All tal cancers within reach of the examination by 60 – 85 %, no large,
screened participants with a positive FOBT were offered colonos- randomized and controlled trials showing the effect of screening
copy. After 13 years of screening and follow-up, the study report- flexible sigmoidoscopy on colorectal cancer mortality have yet
ed a clinically and statistically highly significant reduction in been completed. Two such trials are currently in progress, one
mortality from colorectal cancer of 33.4 % in an annually in the USA and one in the UK. A report from the UK trial pub-
screened group. After an additional 6 years of follow-up, an lished this year described the acceptability, safety, feasibility,
equal-sized biennially screened group had also experienced a and yield of flexible sigmoidoscopy [15]. Men and women aged 29
significant reduction in the colorectal cancer mortality, of 21 %. 55 – 64 years, in 14 UK centers, who responded to a mailed ques-
Later, this trial also reported a decrease in the incidence of colo- tionnaire that they would attend for flexible sigmoidoscopy
rectal cancer in the annually and biennially screened groups of screening if invited, were randomly assigned to screening or con-
20 % and 17 %, respectively, presumably as the result of detection trol groups. If multiple (three or more) or advanced adenoma-
and resection of premalignant adenomatous polyps. The Danish tous polyps (‡ 1 cm or with villous features, high-grade dyspla-
FOBT Screening Trial recently reported the effect of biennial sia, or invasive cancer) were found, follow-up colonoscopy was
FOBT screening after 13 years and seven rounds of screening performed. Of 354 262 people asked about their interest in hav-
[13]. In this trial, 30 967 people were selected from the Funan ing flexible sigmoidoscopy screening, 55 % responded positively,
population for screening, and only those who completed the first and 170 432 were randomized. Attendance for flexible sigmoid-
screening round were invited for further screening. Colonoscopy oscopy among those assigned screening was 71 %. Based on find-
was offered if a screening test was positive. Mortality from colo- ings, 5 % were referred for follow-up colonoscopy. Distal adeno-
rectal cancer was significantly less in the screening group com- mas and cancers were detected in 12.1 % and 0.3 %, respectively.
pared with an equal-sized unscreened control group (RR 0.82), Proximal adenomas and cancers were detected in those under-
and the reduction in mortality was most pronounced above the going colonoscopy in 18.8 % and 0.4 %, respectively. Of the can-
sigmoid colon. After seven rounds of screening, the relative risk cers, 62 % were Dukes stage A, or locally excised. There was one
was reduced to less than 0.70 compared with controls. Nonre- perforation after flexible sigmoidoscopy and four after colonos-
sponders had a significantly increased risk of death from colorec- copy. The authors conclude from this preliminary report that
tal cancer compared with those who accepted the full program. large-volume, population-based flexible sigmoidoscopy screen-
Thus, even when screening is done every other year, it reduces ing in the UK is acceptable, feasible, and safe. The prevalence of
mortality from colorectal cancer by about 30 % in those who neoplasia is high, and the colonoscopy referral rate of 5 % is
comply with the screening protocol. This is an important meas- acceptable and manageable.
urement, because it is what patients can expect if they comply
with recommendations to be screened with biennial FOBTs. The Investigators in the Norwegian Telemark Flexible Sigmoidoscopy
authors conclude that their findings support attempts to intro- Trial assessed the effect of flexible sigmoidoscopy screening on
duce large-scale FOBT screening programs. The smaller observed the prevalence of colorectal adenomas after 13 years of follow-
effect on mortality from distal colorectal cancer, also observed in up [16]. In this long-term trial, 799 men and women aged 50 –
other trials, supports combining FOBT screening with periodic 59 were drawn from the general population in 1983 and random-

Bond JH. Colon Polyps and Cancer · Endoscopy 2003; 35: 27 – 35

 ly assigned to equal screening and control groups. Screenees to flexible sigmoidoscopy screening. Although rudimentary
were offered a one-time flexible sigmoidoscopy, and those found knowledge about colorectal cancer screening appeared ade-
to have polyps were offered follow-up colonoscopy in 1985 and quate, a number of knowledge deficiencies were identified by
1989. All participants were offered colonoscopy in 1996 at the the study. The authors concluded that internal medicine resi-
end of the trial. As previously reported, the incidence of colorec- dents at this institution demonstrated poor compliance with
tal cancer was reduced by 80 % in the screened group compared colorectal cancer screening recommendations and that efforts
to controls. The yield of adenomatous polyps at the 1996 colo- to improve screening rates should include a focus on physicians
noscopy has now been reported for both groups. Of the 71 % in training.
who complied with the final colonoscopy, adenomas were found
in 37 % and 43 % of the screened and control groups, respectively,
and advanced adenomas (‡ 1 cm, or with villous components, or Emerging Screening Tests
high-grade dysplasia) were found in 8 % and 13 %, respectively.
The authors concluded that there was no significant difference Colorectal carcinogenesis is the result of a series of acquired ge-
in adenoma prevalence between the groups, but there was a netic alterations that occur in colonic epithelial cells as adeno-
State of the Art Review

trend toward more advanced adenomas in the control group un- mas develop, advance, and turn to cancer. It is now possible to re-
dergoing usual care. This suggests that there is little effect of cover analyzable DNA from stool and test for the presence of
one-time flexible sigmoidoscopy screening with follow-up colo- these genetic alterations. This may lead to a more specific stool-
noscopy surveillance on the overall prevalence of adenomas, but screening test indicating the likely presence of a cancer or an ad-
a preventive effect on the development of advanced adenomas, vanced adenoma and the need for colonoscopy.
consistent with the previously reported effect on cancer preven-
tion. An earlier preliminary study at the Mayo Clinic retrospectively

Downloaded by: University of British Columbia. Copyrighted material.

assessed the performance of a panel of selected genetic DNA al-
Current guidelines recommend a 5-year interval between nega- terations in stool archived from 33 patients with cancer or ad-
tive screening flexible sigmoidoscopies, even though some stud- vanced adenomas and from 28 controls without neoplasia [20].
ies indicate that the protective effect of a single examination may They reported a sensitivity of 91 % for cancer and of 82 % for ade-
persist up to 10 years. Investigators from Fukuoka studied 192 nomas ‡ 1 cm, with a specificity of 93 %. Another recent study of
asymptomatic, average-risk individuals who had undergone a 51 patients tested with a different panel of just three genetic
negative screening flexible sigmoidoscopy after age 50 and a sec- changes in stool DNA reported a sensitivity of 71 % for detecting
ond examination at least 3 years later [17]. The yield of detecting colorectal cancer [21]. Stool DNA tests are not yet available for
new adenomas was similar whether the second sigmoidoscopy general population-based screening, pending the results of on-
was performed after 3 – 5 years (11.5 %), 5 – 6 years (7.3 %), or 6 – going clinical trials in larger numbers of average-risk people.
8 years (12.2 %). All adenomas detected on the second examina-
30 tion were small (< 1 cm) tubular adenomas. The authors conclud- Another promising stool screening test detects fecal calprotec-
ed that the interval for screening flexible sigmoidoscopy may be tin, a marker found in neutrophils that apparently are shed
extended beyond 5 years in those with a negative screen. from ulcerated tumors. Investigators from London and Norway
compared the results of measuring fecal calprotectin with per-
Compliance with screening flexible sigmoidoscopy may be lim- forming FOBTs in 62 patients with colorectal cancer [22]. Elevat-
ited by patient concerns and anxiety about the discomfort ed stool calprotectin was found in 90 % of cancer patients, while
caused by the examination. A prospective study was conducted only 58 % had a positive FOBT. There was no significant differ-
at two sites in the USA in 764 patients presenting for endoscopy ence in fecal calprotectin levels when considering location or
(flexible sigmoidoscopy 175, colonoscopy 384, upper endoscopy stage of the cancers. Sensitivity for detecting adenomatous
205) [18]. Patients rated their anticipated preprocedure concerns polyps was 55 % using the calprotectin method and 10 % using
before each examination and then rated their actual level of dif- FOBTs. The overall sensitivity and specificity rates of calprotectin
ficulty after each endoscopy. Patients had fewer concerns about for colorectal cancer and adenomatous polyps were 79 % and
flexible sigmoidoscopy before undergoing the procedure, but 72 %, respectively, compared with sensitivity and specificity
afterward rated the actual level of difficulty with the examina- rates of FOBTs of 43 % and 92 %. The authors concluded that fecal
tion higher than did patients undergoing colonoscopy or upper calprotectin is a simple and sensitive noninvasive marker of
endoscopy. These findings suggest that it may be necessary to colorectal neoplasia, and that it is more sensitive than FOBTs
change the perception of flexible sigmoidoscopy as being the for detecting colorectal neoplasia, at the cost of a somewhat
best tolerated of the endoscopic procedures. In studies of barri- lower specificity.
ers to screening, one of the most common reasons people give
for not undergoing screening for colorectal cancer is because Virtual colonoscopy (VC) (computed-tomographic colonogra-
their primary physician did not recommend it. In another study, phy) is a very promising new radiographic method for imaging
investigators from the University of Nebraska assessed the level the large bowel that has undergone remarkable refinement since
of compliance with screening recommendations for patients it was first introduced in 1994 by Vining. It is likely that this
being cared for in medicine residents’ clinics [19]. Only 13 % and method will soon become an additional effective option for colo-
16 % of patients had been screened with FOBT and flexible sig- rectal screening. The technique combines rapid thin-section, he-
moidoscopy, respectively. The residents dramatically overesti- lical computed tomography with off-line computer processing
mated their perceived FOBT and sigmoidoscopy screening rates that can render high-resolution two-dimensional and three-di-
(88 % and 78 %, respectively). Most residents identified barriers mensional images of the colon and rectum. Advantages of VC

Bond JH. Colon Polyps and Cancer · Endoscopy 2003; 35: 27 – 35

over conventional colonoscopy include a shorter examination pain or discomfort with VC and that it was the preferred exami-
time, no need for intravenous sedation, and few, if any, serious nation for future screening.
complications. It allows scrutiny of both sides of the bowel wall
and bowel folds and very precise localization of abnormalities, In addition to examining the large bowel, VC may detect extraco-
and it can examine the proximal colon when an obstructing left lonic abdominal pathology. This may lead to unnecessary and ex-
colonic cancer prevents passage of a colonoscope. Current disad- pensive investigation of benign lesions, or may benefit the pa-
vantages of VC include the need for a very thorough bowel tient by identifying serious, treatable disease. A study in Austra-
cleansing preparation and somewhat uncomfortable gas disten- lia assessed the prevalence and characteristics of extracolonic
sion of the colon. Spasm of a bowel segment, and retained feces pathology found in 100 patients undergoing VC [29]. Fifteen pa-
or fluid, interfere with the interpretation of findings. Substantial tients (15 %) had extracolonic abnormalities detected. In four, the
expensive radiologist time is required to set up and read these pathology had been diagnosed earlier (umbilical hernia, gall-
scans, and, since VC is diagnostic only, a follow-up conventional bladder and renal calculi, a 3.5-cm abdominal aortic aneurysm,
colonoscopy is required whenever a suspicious mass or polyp is and an ovarian cyst). Eleven patients had new abnormalities de-
detected. tected: ovarian cysts in three, liver cysts in two, uterine fibroids

State of the Art Review

in two, and one each with gallstones, splenic calcifications, an
The accuracy of VC for detecting polyps has greatly improved aortic aneurysm, and a renal tumor. Two patients with ovarian
over the past few years and now exceeds that of double-contrast cysts underwent surgery, and histology showed benign cysts.
barium enema. One of the largest studies comparing VC with co- The authors concluded that extracolonic abnormalities are com-
lonoscopy was recently published from the San Francisco VA mon at VC. Most are benign, but may nevertheless lead to appre-
Medical Center [23]. Three hundred patients underwent VC fol- ciable investigative costs. Data such as these need to be carefully
lowed by standard colonoscopy. The overall sensitivity and spe- evaluated in feasibility and cost-effectiveness studies of VC

Downloaded by: University of British Columbia. Copyrighted material.

cificity of VC for cases in which at least one polyp was detected screening for colorectal cancer.
were 90.1 % and 72.0 %, respectively. The sensitivity for detection
of individual adenomas was 94 % for those ‡ 1 cm, 82 % for adeno-
mas 5.0 – 9.9 mm, and 66.9 % for adenomas smaller than 5 mm. Diagnosis of Colorectal Neoplasia
VC detected all eight cancers found in this series. The authors
concluded that VC has excellent sensitivity for the detection of Previous studies from the USA demonstrated that colonoscopy is
clinically important colorectal neoplasia. Smaller comparison substantially more accurate than air-contrast barium enema
studies published this year showed similarly impressive results. (ACBE) for the diagnosis of both colorectal cancers and adenoma-
In a series from Greece, 23 patients found to have colorectal tous polyps [30, 31]. A recent paper from Glasgow reported sim-
polyps at colonoscopy were referred for virtual colonoscopy ilar results [32]. A retrospective analysis was performed of pa-
[24]. Virtual colonoscopy visualized 17 of 19 (89 %) polyps 8 – tients undergoing ACBE (n = 1389) or colonoscopy (n = 1081) as
15 mm in diameter and 10 of 11 (91 %) larger than 15 mm. A study the primary investigation for large-bowel symptoms or for can- 31
in Italy of 66 consecutive symptomatic patients, reported a sen- cer or polyp surveillance. A computerized database search was
sitivity for detecting polyps ‡ 1 cm of 92.8 %, 84.6 % for polyps 6 – conducted 1 and 2 years following each procedure, to detect any
9 mm, and 24 % for polyps smaller than 5 mm [25]. missed polyps or cancers. This analysis indicated comparable
sensitivity rates for ACBE and colonoscopy for detecting cancer
Radiologists at Memorial Sloan–Kettering Cancer Center, New of 83 % and 97.5 %, respectively, and for detecting neoplastic
York, assessed the quality of VC in 61 patients who had under- polyps ‡ 1 cm of 21.7 % and 91.4 %, respectively. Nine patients
gone previous abdominopelvic radiation therapy or surgery, or (0.6 %) had a false-positive diagnosis of cancer in the ACBE group.
both [26]. These are situations in which conventional colonosco- The authors concluded that colonoscopy should be the investiga-
py may fail to examine the entire colon. Two blinded radiologists tion of choice for most patients with large-bowel symptoms sug-
determined that the average overall colonic distension and fluid gestive of neoplastic disease, or for polyp or cancer surveillance.
retention for this group did not differ from that of a control group
who had not undergone these treatments. They concluded that A second study in Perth assessed the accuracy of colonoscopy for
high-quality VC examinations are achieved in patients who diagnosing colorectal cancer by identifying cancers that were de-
have previously undergone radiation, surgery, or both, and that tected after a negative index colonoscopy [33]. Over a 5-year
VC is a good alternative method of imaging the colon when con- period, eight subsequent cancers were detected in 1543 patients
ventional colonoscopy fails in such cases. Two other studies as- (0.5 %), a rate not statistically different from that of the general
sessed patient preferences and acceptance of VC compared to Australian population, and significantly lower than that of all pa-
that of conventional colonoscopy for colorectal cancer screening. tients presenting for colonoscopy during the study period (5.2 %).
In the study at the San Francisco VA Medical Center discussed These authors also concluded that colonoscopy is highly accu-
above, 295 patients completed questionnaires after performance rate; very few patients harbor advanced neoplastic lesions after
of VC and after both VC and colonoscopy had been completed a negative examination. In a study in Lausanne, investigators as-
[27]. The patients reported that they had more pain, discomfort, sessed the effect of appropriateness criteria on the diagnostic
and embarrassment after VC and that they preferred convention- yield of colonoscopy [34]. Colonoscopy performed for appropri-
al colonoscopy and would wait longer in order to have that meth- ate criteria determined by the RAND panel method yielded sig-
od of screening. A similar study in Sweden, however, showed dif- nificantly more relevant lesions than did examinations judged
ferent results [28]. A survey of 111 patients undergoing VC fol- to be inappropriate. Of 51 colorectal cancers detected, all but
lowed immediately by conventional colonoscopy reported less one were found by colonoscopies judged to be appropriate or un-

Bond JH. Colon Polyps and Cancer · Endoscopy 2003; 35: 27 – 35

 certain. The authors concluded that use of such appropriateness trols undergoing screening flexible sigmoidoscopy or colonosco-
criteria could greatly improve patient selection for colonoscopy. py during the same time period. Seven (5 %) in the hernia group
were found to have colorectal cancer, compared to six (4 %) in the
A number of recent studies indicate that the most predictive sign control group (P = 0.8). The authors concluded that this study
or symptom of colorectal cancer requiring colonoscopy is rectal does not support previously published findings that patients
bleeding or unexplained iron-deficiency anemia [35]. A recent with inguinal hernias are more likely to have colorectal neopla-
study in Germany, however, presented compelling evidence that sia, and these patients should undergo screening for colorectal
not all patients with hematochezia require full colonoscopy as cancer at the same rate as the general population.
their initial evaluation [36]. This excellent study employed struc-
tured preprocedure interviews in 4265 patients referred for colo-
noscopy for rectal bleeding. Of these, 468 patients had scant Familial Colorectal Cancer
hematochezia, defined as the presence of small amounts of
bright red blood coating their stools and/or blood on the toilet At least 30 % of colorectal cancers are due at least in part to an in-
tissue. An additional 299 patients had a positive FOBT only, or herited familial risk. Identification and stratification of familial
State of the Art Review

dark blood, or blood mixed in with their stool. At colonoscopy, risk is an important element of screening and surveillance for
those with just scant hematochezia had a prevalence of neo- this malignancy. Investigators from Boston conducted a survey
plasms located above the reach of flexible sigmoidoscopy that of regional gastroenterologists and primary-care physicians to
was no different from that of an equal matched group of patients assess current knowledge and practice patterns regarding famil-
undergoing colonoscopy who had no rectal bleeding or other risk ial risk for colorectal cancer, and to estimate compliance with
factors for colorectal cancer (odds ratio 1.2). In contrast, the risk current screening guidelines with regard to familial risk [40].
of proximal neoplasia was significantly increased in patients Most gastroenterologists and primary-care physicians (85 % vs.

Downloaded by: University of British Columbia. Copyrighted material.

with either a positive FOBT (OR 3.1) or dark blood mixed with 72 %) correctly chose age 40 as the appropriate age to begin
their stool (OR 4.8). The authors concluded that flexible sigmo- screening for a family history of colorectal cancer, but relatively
idoscopy, rather than full colonoscopy, is an appropriate initial few (37 % vs. 36 %) recommended screening at age 40 for a family
evaluation for patients with the common complaint of scant history of adenomatous polyps. Gastroenterologists were signif-
hematochezia. For those with occult stool blood or other forms icantly more likely to recommend screening for familial polypo-
of hematochezia, colonoscopy is generally indicated in the initial sis (FAP) at puberty and for hereditary nonpolyposis colorectal
evaluation. cancer (HNPCC) at age 25 than were primary-care physicians.
Both groups recommended colonoscopy as the preferred method
Contrary to traditional teaching, diarrhea, constipation, or a of screening and surveillance for familial colorectal cancer. Gas-
change in bowel habits are unreliable predictors of serious neo- troenterologists were more likely to recommend genetic testing
plastic disease [37]. A study from Rome determined the yield of for FAP and HNPCC in appropriate clinical settings, and to recom-
32 open access colonoscopy according to the appropriateness of in- mend notification of first-degree relatives of colorectal cancer
dications for the procedure as recommended by the American patients about their need for special screening. Although gastro-
Society for Gastrointestinal Endoscopy (ASGE) [38]. Overall, enterologists were more likely than primary-care physicians to
1123 consecutive patients referred for open-access colonoscopy elicit a family history of colorectal neoplasia (both cancer and
were prospectively entered into the study. The number of colo- adenomatous polyps) and implement appropriate screening
noscopies performed for indications for which the ASGE guide- strategies, the authors of this study found that compliance with
line ruled were “generally not indicated” was 29 % (39 % for pri- guideline recommendations for familial risk was suboptimal.
mary-care physicians and 23 % for subspecialists). A clinically
significant endoscopic finding was detected in 35 % of the exam- A study from Japan investigated the results of FOBT screening in
inations. The yield was higher for those in whom colonoscopy ac- 44 821 subjects divided into two groups – those with and with-
cording to the guideline was indicated (43 %), compared with out a family history of colorectal adenomatous polyps [41]. The
those in whom it was determined not to be indicated (16 %). The FOBT was positive in 8.5 % and 4.8 % of subjects with and without
authors concluded that, although the rate of inappropriate colo- a family history, respectively. The positive predictive values for
noscopy was high, open-access colonoscopy was effective in de- colorectal cancer were 6.8 % and 2.4 % in subjects with and with-
tecting neoplastic lesions. Confining colonoscopy to the group out a family history of adenomatous polyps. Both of these com-
for which it was “generally indicated” would substantially in- parisons were statistically significant. These results support an
crease the cost-effectiveness of an open-access colonoscopy ser- earlier report from the US National Polyp Study that indicates
vice. Another “traditional“ but unproven belief is that there is an an increased risk for colorectal cancer in relatives of patients
increased prevalence of colorectal cancer in patients presenting with adenomatous polyps [42]. Those with a family history of
for inguinal herniorrhaphy. Apparently this idea arose from the colorectal adenomatous polyps as well as cancers should there-
view that an otherwise asymptomatic partially obstructing colo- fore be considered as a priority group for prevention of colorectal
rectal cancer might cause straining at stool that led to the symp- cancer.
tomatic inguinal hernia. Of 614 patients undergoing inguinal
herniorrhaphy during the last 10 years, investigators in Califor- Genetic testing for HNPCC is now successful in 50 – 70 % of fami-
nia analyzed 149 (24 %) who had no prior history of colorectal lies that meet the Amsterdam criteria for the disease. Once a mu-
neoplasia or rectal bleeding and who had had flexible sigmoidos- tation has been found in an index case, relatives can be tested
copy or colonoscopy performed during the perioperative period with near-100 % accuracy, and offspring who test negative do
[39]. They were matched in a case–control study with 149 con- not require the frequent colonoscopic surveillance recommend-

Bond JH. Colon Polyps and Cancer · Endoscopy 2003; 35: 27 – 35

ed for those who inherit the autosomal-dominant germline A total of 127281 patients were included. The authors found that
mismatch repair gene mutation responsible for the disease. Mi- small-intestinal cancer was significantly increased in men with
crosatellite instability (MSI) is a marker for the presence of an in- colorectal cancer after the age of 65 years. Colorectal cancer was
herited or acquired mismatch repair gene mutation and is found also significantly increased after a first diagnosis of cancer of the
in over 95 % of cancers from HNPCC patients, but only in about small intestine. Other cancer sites with a significant increase
15 % of sporadic colorectal cancers. New guidelines suggest that after colorectal cancer included cervical, uterine, and ovarian.
testing for MSI will help to identify those with HNPCC in families They concluded that some of these associations are consistent
that do not satisfy the Amsterdam criteria [43]. Investigators with the effects of known inherited cancer susceptibility genes.
from Denmark carried out studies to determine the frequency of
HNPCC among unselected Danish patients with colorectal cancer
and to determine the value of analyzing cancers for MSI as a pre- Colorectal Polyps
screening test prior to genetic testing [44]. When a positive fam-
ily history of malignancy was obtained, patients were screened Current evidence indicates that over 95 % of colorectal cancers in
for tumor MSI analysis and subjects underwent genetic testing Western countries arise in adenomatous polyps that grow slowly

State of the Art Review

for germline mutations in mismatch repair genes (hMLH1 and in the colon over many years before they turn malignant [47].
hMSH2). Patients who tested positive for germline mutations Small tubular adenomas, however, are very common in the pop-
and/or those with Amsterdam I or II criteria families were cate- ulation, and most never develop the additional acquired genetic
gorized as HNPCC patients. Among 1200 eligible colorectal can- alterations that make them advance and become cancerous. In-
cer patients, a total of 20 (1.7 %) were categorized as HNPCC pa- creasingly, clinicians are shifting their attention away from sim-
tients. The Amsterdam criteria were met in 18, and in another ply finding and harvesting all diminutive colorectal adenomas
two patients, pathologic mutations were detected in the absence toward strategies that allow the reliable detection of the much

Downloaded by: University of British Columbia. Copyrighted material.

of the Amsterdam criteria. The Amsterdam criteria were met in less common, but much more dangerous advanced adenoma. As
80 % of 10 gene carriers, all of whom had the MSI-high pheno- defined by many studies, including the US National Polyp Study,
type. The study concluded that the frequency of HNPCC in Den- an advanced adenoma is one that is ‡ 1 cm in size or contains
mark is approximately 1.7 % of all colorectal cancer cases and high-grade dysplasia or appreciable villous tissue. When screen-
14.3 % among patients younger than 50 years of age. Microsatel- ing colonoscopy is performed in average-risk, asymptomatic in-
lite instability analysis of tumor tissue is a reliable prescreening dividuals over age 50, the prevalence of advanced adenomas
method for the main mismatch repair gene germline mutations ranges from 6 % to 9 %. [48]. A study in Vienna analyzed 7590 ade-
in families suspected of having HNPCC. nomatous polyps from 4216 patients between 1978 and 1996 in
order to determine risk factors for adenomas with advanced
Colorectal cancers in HNPCC are thought to arise from adenomas pathological features (APF) (high-grade dysplasia or invasive car-
and mostly occur in the proximal colon. Recent studies have cinoma) [49]. Size was the strongest predictor. The percentages
helped define the adenoma–carcinoma sequence in HNPCC, the of adenomas with APF were 3.4 %, 13.5 %, and 38.5 % for adenomas 33
one condition in which this sequence appears to occur relatively < 0.5 cm, 0.5 – 1.0 cm, and > 1 cm, respectively. Villous change,
rapidly. left-sided location, and age ‡ 60 years were also associated with
APF. No invasive cancer was found in any polyp £ 0.5 cm, but
Investigators from the Netherlands compared 100 HNPCC adeno- since 3.4 % of these contained high-grade dysplasia, the authors
mas with 152 sporadic adenomas with regard to location, size, recommend their resection whenever possible. A molecular-ge-
and dysplasia [45]. Adenomas in HNPCC patients were more like- netic study at St. Mark’s Hospital, London, reported that colorec-
ly to be smaller than sporadic adenomas and located in the prox- tal adenomas harbor a spectrum of p53 mutations that is signifi-
imal colon (50 % vs. 26 %). All proximal HNPCC adenomas ‡ 5 mm cantly different from that found in colorectal cancers [50]. Their
in diameter were highly dysplastic compared, with only 17 % of findings suggest that some p53 mutations have a weaker effect
the larger proximal sporadic adenomas. Distal HNPCC polyps than others and are therefore more likely to be found in adeno-
were less likely to be highly dysplastic. Small (< 5 mm) HNPCC mas that have not progressed to carcinomas.
polyps, except for their more proximal location, were not differ-
ent from sporadic adenomas. The authors concluded that HNPCC In another study, researchers in Zurich reported that whether or
adenomas are located mainly in the proximal colon, and the pro- not a diminutive (£ 0.5 cm) polyp has clinically insignificant his-
gression to high-grade dysplasia is more common in proximal tology can be predicted from patient characteristics and endo-
than distal HNPCC adenomas, indicating a faster transformation scopic findings, obviating the need for expensive biopsy and his-
rate from early adenoma to cancer in the proximal colon. topathological examinations in many cases [51]. A total of 1681
polyps removed from 494 patients were grouped into adenomas
Patients with familial colorectal cancer, especially those in with advanced pathological features (appreciable villous compo-
HNPCC families, have an increased risk of noncolonic cancers, in- nent, high-grade dysplasia, or early invasive cancer) or insignifi-
cluding uterine, ovarian, urinary tract, small bowel, hepatobili- cant polyps (nonadenomas or simple tubular adenomas with
ary, and gastric. A study in south-east England attempted to low-grade dysplasia). They found that a small (£ 0.5 cm) right-
characterize the occurrence of multiple primary cancers in pa- sided polyp in a young patient (£ 60 years of age) has only a
tients diagnosed with colorectal cancer and explore the possibil- 3.8 % risk for containing advanced pathologic features. Polyps in
ity of a common etiology for different cancer sites [46]. The patients aged over 60, the presence of anemia, polyp size
Thames Cancer Registry database was used to identify patients > 1.0 cm, or left-sided location, as single or combined parameters,
with a first colorectal cancer diagnosed between 1961 and 1995. had a maximum predictive value of 75.4 % for advanced adeno-

Bond JH. Colon Polyps and Cancer · Endoscopy 2003; 35: 27 – 35

 mas. The authors concluded that costly histopathological exam- treatment with APC or no treatment. Routine follow-up colonos-
ination of small polyps in the low-risk group may not be requir- copy was performed at 3 months and 1 year. There were fewer
ed, since it does not provide any clinically relevant additional in- recurrences after APC (one of 10) compared with that of controls
formation. (seven of 11). The authors concluded that in patients with appar-
ent complete endoscopic snare resection of large, sessile adeno-
Recent colorectal cancer screening and surveillance guidelines in mas, postpolypectomy application of APC reduces the rate of
the USA recommend that clinicians should assess each patient’s adenoma recurrence. In a study in Taiwan, 75 sessile polyps in
risk for developing metachronous advanced adenomas and 68 patients were randomized to receive submucosal epinephrine
should tailor postpolypectomy surveillance strategies according- (1 : 10 000) injection, with an additional 76 polyps in 61 patients
ly [2]. Based on available clinical and pathological data, patients serving as controls [56]. There were a total of eight episodes of
with colorectal adenomas now can be stratified into high-risk immediate postpolypectomy hemorrhage, one in the epineph-
and low-risk groups. After the colon has been satisfactorily rine group and seven in the control group. The study concluded
cleared of all synchronous adenomas, repeat colonoscopy is re- that epinephrine injection prior to polypectomy is effective and
commended in 3 years for patients who are at high risk. These significantly prevents immediate bleeding.
State of the Art Review

include those who initially have multiple (three or more) adeno-

mas, a large (‡ 1 cm) adenoma, an adenoma with the advanced
pathological features of villous change, high-grade dysplasia, or References
invasive carcinoma, or a first-degree relative with colorectal can-
1
cer. Patients with a low risk of metachronous advanced adeno- Pigone M, Rich M, Teutsch SM et al. Screening for colorectal cancer
in adults at average risk: summary of the evidence for the U.S. Pre-
mas include those who initially have only one or two small
ventive Services Task Force. Ann Intern Med 2002; 137: 132 – 141
(< 1 cm) tubular adenomas without high-grade dysplasia or can-

Downloaded by: University of British Columbia. Copyrighted material.

2
Winawer SJ, Fletcher R, Rex DK et al. Colorectal cancer screening
cer, and no significant family history of colorectal cancer or ade- and surveillance: clinical guidelines and rational – update based
nomas. For these patients, the first postpolypectomy follow-up on new evidence. Gastroenterology, Feb. 2003 [in press]
3
Nelson DB, McQuaid KR, Bond JH et al. Procedural success and
colonoscopy can safely be delayed until at least 5 years – or, for
complications of large-scale screening colonoscopy. Gastrointest
selected older patients or those with substantial comorbidity, no Endosc 2002; 55: 307 – 314
follow-up may be indicated. 4
Dafnis G, Ekbom A, Pahlman L, Blomqvist P. Complications of diagnos-
tic and therapeutic colonoscopy within a defined population in Swe-
den. Gastrointest Endosc 2002; 54: 302 – 309
5
Lieberman DA, Weiss , DG , VA CooperativeStudy Group 380. One-
Colonoscopic Polypectomy time screening for colorectal cancer with combined fecal occult
blood testing and examination of the distal colon. N Engl J Med
In a multicenter study in the USA, 520 polyps < 10 mm in size 2001; 345: 555 – 560
6
Hoffman PK. The anatomical distribution of colorectal polyps at colo-
were sprayed with indigo carmine dye and examined with a
noscopy. J Clin Gastroenterol 2001; 33: 222 – 225
34 high-resolution endoscope [52]. Polyps with a surface pitted ap- 7
Ikamoto M, Shiratori Y, Yamaji Y et al. Relationship between age and
pearance were judged to be hyperplastic, and those with surface site of colorectal cancer based on colonoscopic findings. Gastrointest
grooves or sulci were classed as adenomatous. All polyps were Endosc 2002; 55: 548 – 551
8
Nusko G, Mansmann U, Wiest G et al. Right-sided shift found in meta-
resected and examined histopathologically. Forty polyps (7.7 %)
chronous colorectal adenomas. Endoscopy 2001; 33: 574 – 579
could not be classified. For the remainder, the sensitivity, specif- 9
Kim LS, Koch J, Yee J et al. Comparison of patients’ experiences during
icity, and negative predictive values for indigo carmine determi- imaging tests of the colon. Gastrointest Endosc 2001; 54: 67 – 74
10
nations were 82 %, 82 %, and 88 %, respectively. The authors con- Schoen RDE, Weissfeld JL, Trauth JM et al. A population-based, com-
munity estimate of total colon examination: the impact on compli-
cluded that this method of high-resolution chromoendoscopy
ance with screening for colorectal cancer. Am J Gastroenterol 2002;
reliably separates adenomatous from nonadenomatous polyps, 97: 446 – 451
obviating the need for expensive biopsy and pathological evalu- 11
Imperiale TF, Wagner DR, Lin CY et al. Results of screening colonos-
ation. A German study employed indigo carmine chromoendos- copy among persons 40 – 49 years of age. N Engl J Med 2002; 346:
1781 – 1785
copy in 100 patients and reported finding 178 lesions in 27 of 48 12
Bond JH. Fecal occult blood test screening for colorectal cancer. Gas-
patients whose colonic mucosa appeared normal without dye trointest Endosc Clin N Am 2002; 12: 11 – 21
spraying [53]. The mean size of these lesions was 3 mm, and 13
Jorgensen OD, Kronborg O, Fenger C. A randomized study of
most showed hyperplastic or inflammatory changes. Overall, screening for colorectal cancer using faecal occult blood testing:
results after 13 years and seven biennial screening rounds. Gut
adenomas were differentiated from nonadenomas with a sensi-
2002; 50: 29 – 32
tivity and specificity of 92 % and 93 %, respectively. A third study 14
Scholefield JH, Moss S, Sufi F et al. Effect of faecal occult blood
of magnification chromoendoscopy, in Taipei, reported an overall screening on mortality from colorectal cancer: results from a ran-
diagnostic accuracy in differentiating neoplastic from nonneo- domized controlled trial. Gut 2002; 50: 840 – 844
15
UK Flexible Sigmoidoscopy Screening Trial Investigators. Single
plastic lesions of 80.1 % [54].
flexible sigmoidoscopy screening to prevent colorectal cancer:
baseline findings of a UK multicentre randomized trial. Lancet,
New polypectomy techniques continue to improve the efficiency 2002; 359: 1291 – 1300
16
and safety of colonoscopic polypectomy. Endoscopists at St. Thiis-Evensen E, Hoff GS, Sauar J et al. The effect of attending a flex-
ible sigmoidoscopic screening program on the prevalence of colo-
Mark’s Hospital, London, evaluated the effectiveness of treating
rectal adenomas at 13-year follow-up. Am J Gastroenterol 2001; 96:
the base and margin of large (> 1.5 cm) resected sessile polyps 1901 – 1907
with argon plasma coagulation (APC) on preventing polyp recur- 17
Yoshinaga M, Watabe R, Yanagisawa J et al. The interval between flex-
rence [55]. A group of 21 patients with polyps believed by the en- ible sigmoidoscopy screening examinations can be expanded beyond
five years. Hepatogastroenterology 2001; 48: 437 – 439
doscopist to have been totally excised were randomly assigned to

Bond JH. Colon Polyps and Cancer · Endoscopy 2003; 35: 27 – 35

18 37
Ringel Y, Dalton CB, Brandt LJ et al. Flexible sigmoidoscopy: the pa- Lieberman DA, de Garmo PL, Fleischer DE et al. Colonic neoplasia in
tient’s perception. Gastrointest Endosc 2002; 55: 315 – 320 patients with nonspecific GI symptoms. Gastrointest Endosc 2000;
19
Zack DL, DiBaise JK, Quigley EM, Roy HK. Colorectal cancer screening 51: 647 – 651
38
compliance by medicine residents: perceived and actual. Am J Gastro- Hassan MS, Meucci G, Toldi A et al. Diagnostic yield of open access co-
enterol 2001; 96: 3004 – 3008 lonoscopy according to appropriateness. Gastrointest Endosc 2001;
20
Ahlquist DA, Skoletsky JE, Boynton et al. Colorectal cancer screening 54: 175 – 179
39
by detection of altered human DNA in stool: feasibility of a multitar- Gerson LB, Triadafilopoulos G. Is colorectal cancer screening necessary
get assay panel. Gastroenterology 2000; 119: 1219 – 1227 in the preoperative assessment of inguinal herniorrhaphy? A case–
21
Dong SM, Traverso G, Johnson C et al. Detecting colorectal cancer in control study. Am J Gastroenterol 2001; 96: 1914 – 1917
40
stool with the use of multiple genetic targets. J Natl Cancer Inst Schroy PC, Barrison AF, Ling BS et al. Family history and colorectal can-
2001; 93: 858 – 865 cer screening: a survey of physician knowledge and practice patterns.
22
Tibble J, Sigthorsson G, Foster R et al. Faecal calprotectin and faecal oc- Am J Gastroenterol 2002; 97: 1031 – 1036
41
cult blood tests in the diagnosis of colorectal carcinoma and adenoma. Nakama H, Fukazawa K. Colorectal cancer risk in first-degree relatives
Gut 2001; 49: 402 – 408 of patients with colorectal adenomatous polyp. Hepatogastroenterol-
23
Yee J, Akerkar GA, Hung RK et al. Colorectal neoplasia: perform- ogy 2002; 49: 157 – 159
42
ance characteristics of CT colonography for detection in 300 pa- Winawer SJ, Zauber AG, Gerdes H et al. Risk of colorectal cancer in the
tients. Radiology 2001; 219: 685 – 692 families of patients with adenomatous polyps. N Engl J Med 1996;

State of the Art Review

24
Xynopoulos D, Stasinopoulou M, Dimitroulopoulos D et al. Colorectal 334: 82 – 87
43
polyp detection with virtual colonoscopy (computed tomographic co- Rodriguez-Bigas MA, Boland CR, Hamilton SR et al. A National Cancer
lonography): the reliability of the method. Hepatogastroenterology Institute workshop on hereditary nonpolyposis colorectal cancer syn-
2002; 49: 124 – 127 drome: meeting highlights and Bethesda guidelines. J Natl Cancer Inst
25
Laghi A, Iannaccone R, Carbone I et al. Computed tomographic colo- 1997; 89: 1758 – 1762
44
nography virtual colonoscopy): blinded prospective comparison with Katballe N, Christensen M, Wikman FP et al. Frequency of heredi-
conventional colonoscopy for the detection of colorectal neoplasia. tary non-polyposis colorectal cancer in Danish colorectal cancer
Endoscopy 2002; 34: 441 – 446 patients. Gut 2002; 50: 43 – 51
26 45
Gollub MJ, Ginsberg MS, Cooper C, Thaler HT. Quality of virtual colo- Rijcken FE, Hollema H, Kleibeuker JH. Proximal adenomas in heredi-

Downloaded by: University of British Columbia. Copyrighted material.

noscopy in patients who have undergone radiation therapy or sur- tary non-polyposis colorectal cancer are prone to rapid malignant
gery: how successful are we? AJR Am J Roentgenol 2002; 178: 1109 – transformation. Gut 2002; 50: 382 – 386
46
1116 Evans HS, Moller H, Robinson D et al. The risk of subsequent primary
27
Akerkar GA, Yee J, Hung R, McQuaid K. Patient experience and prefer- cancers after colorectal cancer in southeast England. Gut 2002; 50:
ences toward colon cancer screening: a comparison of virtual colonos- 647 – 652
47
copy and conventional colonoscopy. Gastrointest Endosc 2001; 54: Bond JH. Clinical evidence for the adenoma–carcinoma sequence, and
310 – 315 the management of patients with colorectal adenomas. Semin Gastro-
28
Svensson MH, Svensson E, Lasson A, Hellstrom M. Patient acceptance intest Dis 2000; 11: 176 – 184
48
of CT colonography and conventional colonoscopy: prospective com- Lieberman DA, Weiss D, Bond JH et al. Use of colonoscopy to screen
parative study in patients with or suspected of having colorectal dis- asymptomatic adults for colorectal cancer. N Engl J Med 2000; 343:
ease. Radiology 2002; 222: 337 – 345 162 – 168
29 49
Edwards JT, Wood CJ, Mendelson RM, Forbes GM. Extracolonic find- Gschwantler M, Kriwanek S, Langner E et al. High-grade dysplasia and
ings at virtual colonoscopy: implications for screening programs. Am invasive carcinoma in colorectal adenomas: a multivariate analysis of
J Gastroenterol 2001; 96: 3009 – 3012 the impact of adenoma and patient characteristics. Eur J Gastroenterol
30
Winawer SJ, Steward ET, Zauber AG et al. A comparison of colonoscopy Hepatol 2002; 14: 183 – 188 35
50
and double-contrast barium enema for surveillance after polypecto- Hao XP, Frayling IM, Sgouros JG et al. The spectrum of p53 mutations
my. N Engl J Med 2000; 342: 1766 – 1777 in colorectal adenomas differs from that in colorectal cancers. Gut
31
Rex DK, Rahmani EY, Haseman JH et al. Relative sensitivity of colonos- 2002; 50: 834 – 839
51
copy and barium enema for detection of colorectal cancer in clinical Kulling D, Christ AD, Karaaslan N et al. Is histological investigation of
practice. Gastroenterology 1997; 112: 17 – 23 polys always necessary? Endoscopy 2001; 33: 454 – 457
32 52
Smith GA, O’Dwyer PJ. Sensitivity of double-contrast barium ene- Eisen GM, Kim CY, Fleischer DE et al. High-resolution chromoendosco-
ma and colonoscopy for the detection of colorectal neoplasms. py for classifying colonic polyps: a multicenter study. Gastrointest En-
Surg Endosc 2001; 15: 649 – 652 dosc 2002; 55: 687 – 694
33 53
Ee HC, Semmens JB, Hoffman NE, Perth Teaching Hospitals Endoscopy Kiesslich R, von Bergh M, Hahn M et al. Chromoendoscopy with indigo
Group. Complete colonoscopy rarely misses cancer. Gastrointest En- carmine improves the detection of adenomatous and nonadenoma-
dosc 2002; 55: 167 – 171 tous lesions in the colon. Endoscopy 2001; 33: 1001 – 1006
34 54
de Bosset V, Froelich F, Rey JP et al. Do explicit appropriateness criteria Tung SY, Wu CS, Su MY. Magnifying colonoscopy in differentiating
enhance the diagnostic yield of colonoscopy? Endoscopy 2002; 34: neoplastic from nonneoplastic colorectal lesions. Am J Gastroenterol
360 – 368 2001; 96: 2628 – 2632
35 55
Bond JH. Rectal bleeding: is it always an indication for colonoscopy? Brooker JC, Saunders BP, Shah SG et al. Treatment with argon plasma
Am J Gastroenterol 2002; 97: 223 – 225 coagulation reduces recurrences after piecemeal resection of large
36
Eckardt VF, Schmitt T, Kanzier G et al. Does scant hematochezia ne- sessile colonic polyps: a randomized trial and recommendations. Gas-
cessitate the performance of total colonoscopy? Endoscopy 2002; trointest Endosc 2002; 55: 371 – 375
56
34: 599 – 603 Hsieh YH, Lin HJ, Tseng GY et al. Is submucosal epinephrine injection
necessary before polypectomy? A prospective, comparative study. He-
patogastroenterology 2001; 48: 1379 – 1382

Bond JH. Colon Polyps and Cancer · Endoscopy 2003; 35: 27 – 35