Introduction
Diminutive polyps (≤ 5 mm) are the most common types found during colonoscopies [1 ]. Their histopathologic examination is costly while their potential for malignancy
is low [1 ]
[2 ]
[3 ]. Because diminutive polyps have such a low risk for harboring or progressing to
cancer it might be reasonable to forgo histopathology in favor of less costly strategies.
Image enhanced endoscopy can be used for real-time optical diagnosis of colorectal
polyps. Validated classification systems based on color, vascular pattern and other
criteria have been developed to determine if a polyp is neoplastic or not. The approach
of replacing histopathology with optical diagnosis has been named the resect-and-discard
strategy [4 ]
[5 ]
[6 ]
[7 ]. Studies have shown that in the hands of adequately trained endoscopists, real-time
optical diagnosis provides a good alternative to pathology while being more cost-effective
[8 ]
[9 ]. Consequently, multiple gastroenterology societies have recommended adopting resect-and-discard
as part of screening programs if certain quality standards can be met [10 ]
[11 ]
[12 ]. Clinical implementation of this approach, however, seems sparse.
We were therefore interested in evaluating the current clinical uptake and barriers
for implementation of the resect-and-discard strategy, understanding the perceived
cancer risk for diminutive polyps among endoscopists and potential concerns with leaving
diminutive polyps unresected or with using computed tomography-colonography in screening
and surveillance.
Methods
The Checklist for Reporting Results of Internet E-Surveys (CHERRIES) criteria was
used to present the results of our survey [13 ].
Study design
A cross-sectional international survey study was conducted. Target population were
endoscopists with active colonoscopy practice regardless of profession, case-load,
or experience. The survey was developed through discussion and consensus between 3
gastroenterologists (H.P., A.B., D.vR.). The study protocol and survey were approved
by the CRCHUM Institutional Review Board (CER-number: 17.063). No personal information
was collected or stored during the survey.
Recruitment process
Twenty-nine Gastroenterology, Endoscopy and Surgical associations were contacted between
July 2017 and May 2018 (Table S1 ). These associations were asked to send our online survey to their members using
their mailing list, newsletter or social media (Twitter and Facebook). Survey participation
was on a voluntary basis. The survey was advertised as taking a total of 3 minutes
to complete. No monetary or non-monetary incentives were provided for the completion
of our survey. Double clicking was allowed, results were only transmitted to the database
if the participant clicked on survey completed at the end of the questionnaire. The
survey advertisement text is available in Appendix 1 .
Survey content
We used “Google Forms” to administer an online questionnaire comprising 26 questions
distributed on five pages (Appendix 2 ). The questions were listed in the same order for all participants. Participants
were able to return to previous pages to modify their answers. The survey was only
accessible through the link provided to the contacted associations.
The primary outcome was endoscopist uptake of the resect-and-discard strategy in current
clinical practice. Secondary outcomes included perceived feasibility and barriers
to implementation for a resect-and-discard strategy; perceived cancer risk for resect-and-discard
implementation; perceived cancer risk for diminutive polyps; concerns with leaving
diminutive polyps unresected; and the perceived cancer risk of adopting computed tomography
(CT) colonography as primary screening and surveillance strategy.
The survey collected information on participant demographics and practice setting
data including the country of practice; private, academic or mixed practice setting;
training specialty; years in practice; number of yearly colonoscopies performed and
practice reimbursement. The survey further assessed participants’ knowledge of the
existence of the resect-and-discard strategy; current practice implementation and
barriers for implementation of a resect-and-discard approach; opinion on cancer risk
associated with diminutive polyps and the resect-and-discard strategy; as well as
usage of forceps/snares for different polyp types. The survey questions with possible
answer options can be found in Appendix 2 .
Statistical analysis
We used descriptive analysis with frequency and percentages to describe the participant’s
characteristic and their answers. To examine the association between participant characteristics
and survey responses, we used univariate using chi square test and multivariate analysis
using multilevel logistic regression model. A two-tailed P < 0.05 was considered significant. For statistical analysis, SPSS 25 (Chicago, Illinois,
United States) was used.
Results
The survey was distributed by nine associations (Table S2 ). We estimated that a total of 13,818 individuals were reached through emails based
on the feedback we received regarding societies’ mailing lists. An additional 8,991
individuals were reached through social media. This number was estimated by the number
of unique views from our posts on social media pages or the number of people following
the page at the time of the post if the number of unique views was not accessible.
This resulted in a distribution to 21,807 individuals, not considering individuals
that could be members of more than one association. Overall, 808 responders participated
in the survey. The response rate was 3.7 %. Most of the responses received were from
the ACG mailing list (459/808 responders). Response rate from ACG members was 3.8 %,
which was similar to the overall response rate of 3.7 % using newsletter, mailing
list and social medias (P = 0.96).
Endoscopist demographics
The majority of survey participants were from to the United States (56.2 %), Europe
(11.4 %) and Canada (9.9 %). Most participants were Gastroenterologists (84.4 %).
Endoscopist practice setting was private (38.1 %), academic (28.8 %), and community-based
(16.5 %). 45.5 % of endoscopists were reimbursed on a fee-for-service and 32.9 % on
a salary basis. More details provided in [Table 1 ].
Table 1
Characteristic of survey participants.
Characteristics
Responses (n = 808)
Country of practice
N (%)
454 (56.2)
92 (11.4)
80 (9.9)
61 (7.5)
55 (6.8)
29 (3.6)
25 (3.1)
12 (1.5)
Practice setting
308 (38.1)
233 (28.8)
133 (16.5)
126 (15.6)
8 (1.0)
Training and level
682 (84.4)
50 (6.2)
38 (4.7)
14 (1.7)
10 (1.2)
6 (0.7)
8 (1.0)
Years in practice
373 (46.2)
166 (20.5)
258 (31.9)
11 (1.4)
Colonoscopies performed each year
33 (4.1)
151 (18.7)
163 (20.2)
454 (56.2)
7 (0.9)
Practice reimbursement
368 (45.5)
266 (32.9)
165 (20.4)
9 (1.1)
Attitudes and practices with regards to a resect-and-discard strategy
84.2 % (95 % CI 81.6 %–86.7 %) of endoscopists were not using the resect-and-discard
strategy at the time of the survey, while 59.9 % (95 % CI 56.5 %–63.2 %) did not believe
such an approach was feasible for implementation ([Fig. 1 ]). However, 80.3 % (95 % CI 77.5–83.0 %) of endoscopists believed that using a resect-and-discard
strategy for diminutive polyps would not lead to an increased cancer risk for patients.
Practicing in Canada (OR = 0.25; 95 %CI 0.12–0.51) and in the United States (OR = 0.09;
95 %CI 0.05–0.14) were factors that were statistically significantly associated with
not practicing a resect-and-discard strategy in the multivariate analysis ([Table 2 ]).
Fig. 1 Endoscopist usage of resect-and-discard strategy and perceptions of feasibility.
Table 2
Comparison of demographic characteristics between survey participants who use resect-and-discard
in their current practice and those who don’t.
Use of the resect-and-discard
Yes[1 ] n, (%)
No n, (%)
univariate analysis
multivariate analysis (OR)
Country of practice
0.01
5 (17.2 %)
24 (82.7 %)
n.s.
11 (13.7 %)
69 (86.3 %)
0.25 (0.12; 0.51)
23 (5.1 %)
431 (94.9 %)
0.09 (0.05; 0.14)
27 (45.0 %)
33 (55.0 %)
n.s.
35 (38.5 %)
56 (61.5 %)
n.s.
13 (23.6 %)
42 (76.4 %)
n.s.
9 (44.0 %)
14 (56.0 %)
n.s.
Practice setting
0.40
39 (16.7 %)
194 (83.3 %)
n.s.
23 (17.4 %)
109 (82.6 %)
n.s.
41 (13.3 %)
267 (86.7 %)
n.s.
24 (18.4 %)
102 (81.6 %)
n.s.
Training and level
0.30
104 (15.3 %)
577 (84.7 %)
n.s.
4 (28.6 %)
10 (71.4 %)
n.s.
8 (21.6 %)
29 (78.4 %)
n.s.
3 (30.0 %)
7 (70.0 %)
n.s.
1 (16.7 %)
5 (83.3 %)
n.s.
7 (14.0 %)
43 (86.0 %)
n.s.
Years in practice
0.01
72 (19.4 %)
300 (80.6 %)
n.s.
23 (13.9 %)
142 (86.1 %)
n.s.
30 (11.6 %)
228 (88.4 %)
n.s.
Colonoscopies per year
< 0.01
12 (36.4 %)
21 (63.6 %)
n.s.
32 (21.2 %)
119 (78.8 %)
n.s.
26 (16.0 %)
136 (84.0 %)
n.s.
57 (12.6 %)
396 (87.4 %)
n.s.
Procedure reimbursement
0.01
45 (12.2 %)
323 (87.8 %)
n.s.
47 (17.7 %)
219 (82.3 %)
n.s.
34 (20.9 %)
129 (79.1 %)
n.s.
n.s., not statistically significant
1 Includes people who answered yes for polyps up to 5 mm, yes for polyps up to 10 mm
and yes for rectosigmoid polyps only
When stratified by region, the majority of endoscopists (range: 75.4–90.1 %) replied
that they had heard of the resect-and-discard strategy, however, most were not using
it in their current practice (range: 55–94.9 %). European (38.5 %) and Asian (45 %)
endoscopists had the highest rates of resect-and-discard implementation, while Canadian
(13.8 %) and American (5.1 %) endoscopists displayed lower rates. When asked if this
strategy would be feasible in general practice most endoscopists replied “no” with
the exception of European endoscopists, 54.3 % of whom replied “yes”. There was geographic
agreement that resect-and-discard did not increase patient cancer risk (Table S3 ).
When asked about barriers for implementation of a resect-and-discard strategy 44.6 %
of participants (95 % CI 41.1–48.0 %) were afraid of making a wrong diagnosis, 53.8 %
(95 % CI 50.4–57.3 %) were concerned of potential medico-legal issues and 58.3 % (95 %
CI 54.9–61.7 %) were afraid of assigning incorrect surveillance intervals to patients
([Fig. 2 ]). For American endoscopists, the most important issue preventing implementation
of a resect-and-discard strategy was fear of medico-legal issues (67.2 %), which was
significantly more than the rest of the world (37.7 %; P < 0.001). Fear of making a wrong diagnosis was the main barrier for the majority
of endoscopists from Australia/New Zealand (60 %) and Asia (37.7 %). For the rest
of the world, fear of incorrect surveillance interval assignment was the most important
issue (Table S4 ).
Fig. 2 Endoscopist perception of reasons for resect-and-discard non-feasibility. Multiple
answers were allowed.
Perception of cancer risk associated with diminutive polyps
Overall, 63.0 % of survey participants partly or completely agreed that diminutive
polyps can be left unresected until the next screening colonoscopy because of a low
associated cancer risk. Endoscopists were however evenly split on the effects of leaving
such polyps unresected with regards to patient cancer risk, with 48.4 % (95 % CI 45.0 %–51.9 %)
thinking that leaving diminutive polyps in place would increase the cancer risk of
patients ([Table 3 ]).
Table 3
Endoscopist perceptions on the cancer risk of diminutive polyps.
Questions
[1 ]
Responses; N (%)
Do you think that leaving diminutive polyps increases the risk of cancer of patients?
N = 797
No
411
(51.6)
Yes
386
(48.4)
Cancer risk in a diminutive polyp is so low that such polyp can be left unresected
until the next follow-up colonoscopy
N = 803
I agree
101
(12.6)
I partly agree
405
(50.4)
I partly disagree
129
(16.1)
I completely disagree
168
(20.9)
If you leave a diminutive polyp unresected, the next colonoscopy should be within
a maximum of
N = 790
1 year
92
(11.6)
3 years
245
(31.0)
5 years
383
(48.5)
10 years
70
(8.9)
Do you leave diminutive polyps (up to 5 mm) in place in your current practice?
N = 808
Sometimes
370
(45.8)
In the majority of cases
51
(6.3)
Always
4
(0.5)
If appearance of the polyp suggests it is non-adenomatous
420
(52.0)
If patient is on anticoagulation medication
170
(21.0)
If patient has severe comorbidities
161
(19.9)
If follow-up colonoscopy already scheduled
98
(12.1)
1 Multiple answers were allowed
When stratified by region, only 20 % of participants from Australia/New Zealand partly
or completely agreed that diminutive polyps can be left unresected until the next
screening colonoscopy, in contrast to participants from all other regions where between
58.2 and 68 % partly or completely agreed that this approach was safe. The majority
of endoscopists from the United States (51.3 %), Asia (59.0 %) and South/central America
(54.5 %) thought that leaving diminutive polyps in place could increase cancer risk.
In contrast, most European (58.2 %) and Canadian (63.3 %) endoscopists thought that
this was not so (Table S5 ).
There was no consensus on appropriate follow-up after leaving diminutive polyps in
place. 61.1 % of North American endoscopists chose a maximum of 5-year follow-up after
unresected diminutive polyps compared to other regions where 74.4 % chose a 3-year
or less maximum follow-up (P < 0.001). Endoscopists from the United States (12.5 %) and Canada (10.1 %) had much
higher rates of recommending the maximum option of 10-year follow-up after unresected
diminutive polyps. When stratified by profession (Gastroenterologist, surgeon, internist,
nurse endoscopist), 50.3 % of gastroenterologists chose a maximum of 5-year follow-up
and 73.5 % of other specialties chose a maximum of 3-years or less follow-up when
leaving diminutive polyps unresected (P = 0.003).
Perceptions of non-resection of diminutive polyps and use of CT-Colonography
Perceptions of non-resection of diminutive polyps and use of CT-Colonography
Overall, 52 % of endoscopists were leaving non-adenomatous appearing diminutive polyps
in place in their current practice. Geographical distribution was very variable: between
37.7 % (Asia) and 80.0 % (Australia/ New Zealand) of endoscopists responded that they
were leaving non-adenomatous appearing diminutive polyps in place in their current
practice (Table S5 ).
Of the surveyed endoscopists, 54.7 % (95 % CI 53.6–60.4 %) thought that current CT-colonography
guidelines were “probably” or “definitely” putting patients at higher cancer risk
([Table 4 ]). When stratified by region, the majority of American, European, Asian and South/Central
American endoscopists agreed with that statement versus a minority of Canadian (32.5 %)
and Australian/New Zealand (20 %) endoscopists.
Table 4
Endoscopist perceptions of CT-colonography, resect-and-discard and cancer risk.
Questions
Responses; N (%)
Do you think that current CT-colonography practice, which leaves polyps < 6 mm in
place until the next surveillance exam, leads to an increased risk of colon cancer
for the patient?
N = 804
No
59
(7.3)
Probably Not
305
(37.9)
Probably Yes
324
(40.3)
Yes
116
(14.4)
Do you think that using the resect-and-discard strategy for diminutive polyps increase
the risk of cancer of patients?
N = 808
No
639
(80.3)
Yes
157
(19.7)
CT, computed tomography
Discussion
To our knowledge, our study represents the largest international survey available
to date studying endoscopist opinion on resect-and-discard and diminutive polyps.
It is also the first providing data on geographic differences in endoscopist attitudes
towards resect-and-discard. Endoscopist characteristics were also well distributed
according to practice setting, practice reimbursement and years in practice which
lends external validity to our data. Geographic distribution was skewed towards North
America with endoscopists from the United States, Europe and Canada providing the
majority of responses.
Our survey found that only 15.8 % of endoscopists use the resect-and-discard strategy
in their current practice and 59.9 % thought that implementation was not feasible
of the resect-and-discard strategy in its current form. The most important reasons
why the resect-and-discard strategy was not feasible included fear of making an incorrect
diagnosis leading to incorrect surveillance interval assignment and medicolegal issues.
Our results were similar to those found by Soudagar et al. 2016, where medicolegal
concerns were the main barrier for implementation of the resect-and-discard strategy
for the 105 Gastroenterologists surveyed during a national conference in the United
States [14 ]. These reasons seem to point towards a concern of potential interval CRCs when using
a resect-and-discard strategy, however, 80.3 % of endoscopists voiced the opinion
that a resect-and-discard strategy would not increase CRC risk. This could be caused
by endoscopists feeling that interval cancer, while not more frequent with the resect-and-discard
strategy, would be more difficult to explain in a possible future medicolegal pursuit.
Very few endoscopists cited complexity of resect-and-discard and training requirements
as barriers for implementation. While the consensus for most regions is that resect-and-discard
was not feasible, European endoscopists showed an increased adoption of the strategy
(54.3 %). Endoscopist practice can be dependent upon current healthcare culture, such
as fear of medico-legal issues, acquiring new technology, and early adoption of new
trends, which potentially explains the varying dispositions observed between geographic
regions.
While the survey found that endoscopists did not completely trust their capacity to
make accurate diagnoses, recent meta-analyses have shown that adequately trained endoscopists
can achieve > 90 % concordance with histology-based diagnosis and > 90 % negative
predictive value during optical diagnosis [8 ]
[9 ]. The American Society for Gastrointestinal Endoscopy (ASGE) and the European Society
of Gastrointestinal Endoscopy (ESGE) currently recommend the use of resect-and-discard
if these thresholds can be met. However other studies have shown that endoscopists
can be unable to reach these benchmarks even after optical diagnosis training [10 ]
[15 ]
[16 ]. Most gastroenterology and endoscopy societies across the world have not yet officially
made statements on resect-and-discard in their guidelines because of this contradictory
data, contributing to the low implementation rate we found in this study. Further,
recent research in imaging technologies and optical diagnosis have led to the emergence
of many optical diagnosis-based classification systems, each with their own criteria,
which can be confusing and overwhelming for endoscopists [17 ]. Our survey shows a definite concern from endoscopists about the feasibility of
making optical diagnoses and highlights the need for simplifying decision-making or
removing the need for optical diagnosis altogether [18 ]. A recent study has proposed a simplified and/or location-based strategy to reduce
the need for optical diagnosis [19 ]. These simplified strategies achieved > 90 % surveillance interval agreement compared
to pathology and allowed for providing more patients with surveillance interval planning
on the same day as the colonoscopy [19 ]. Other recent studies have proposed using only number and size of polyps as criteria
for surveillance interval assignment [20 ]
[21 ]. This approach was able to achieve a 90 % surveillance interval agreement with pathology
in one instance and an 89.3 % agreement in another [20 ]
[21 ]. More than 80 % of patients could be provided with a surveillance interval on the
same day using such an approach [20 ]
[21 ]. These alternative approaches thus appear promising even though endoscopists tended
to prefer shorter intervals [20 ]. Limitations of simplified polyp (number and size only) or location-based strategies
include the problem of not being able to distinguish between early signs of cancer
(i. e. NICE-3 morphology) in optical diagnosis.
There was no consensus between endoscopists on cancer risk of diminutive polyps, with
about 50 % believing that leaving them unresected increased cancer risk. This result
differs from a survey by Gellad et al. 2013, which reported that the majority of respondents
would be somewhat agreeable to leave diminutive polyps in place if guidelines would
support this practice [22 ]. Half of endoscopists in our study believed that CT-colonography increased patient
risk. Studies on CT-colonography performance show a sensitivity of > 90 % for the
detection of polyps > 10 mm and > 95 % for the detection of colorectal cancer [23 ]
[24 ]
[25 ]. CT-colonography was, however, shown to have lower detection of high-risk SSAs when
compared to colonoscopy [26 ] and exhibits poor sensitivity in diminutive polyp detection [27 ]. However, risk of malignancy of diminutive colon polyps was shown to be extremely
low: 2 % having advanced histology, and approximately 0.05 % containing high grade
dysplasia or neoplasia [1 ]
[2 ]. Studies with large numbers of small and diminutive polyps found no CRC present
in any of these polyps [2 ]
[28 ]. Recent studies on the natural history of diminutive polyps showed a very indolent
course for these polyps on follow-ups [29 ]
[30 ]. Current literature therefore suggests that risk of CRC arising from diminutive
polyps is extremely low.
It is interesting to note that most endoscopists in our survey reported leaving polyps
unresected in their patients and assigning 3- to 5-year maximum surveillance intervals
when doing so. A recent meta-analysis has shown poor worldwide adherence for post-colonoscopy
surveillance intervals [31 ]. Because resect-and-discard allows for providing more patients on the same day as
the colonoscopy with a decision on when the next surveillance interval can be scheduled,
clinical introduction of resect-and-discard might not only reduce immediate pathology
costs but also upstream costs through avoiding unnecessary short follow-up surveillance
intervals.
One important study limitation is the low response rate causing a potential for selection
bias. However, since endoscopists can be members of multiple societies our response
rate might be underestimated and represented a worst-case scenario. The advertising
of our survey through Facebook/Twitter could also have skewed our sample towards younger
endoscopists with more of an online presence or encourage double clicking from some
of the participants. Furthermore, survey study always raise the issue of possible
participation bias where survey participants are more incline to be up to date with
recent guidelines and recommendations. Even in this selected population, resect-and-discard
uptake was very low; suggesting the true uptake in the general population is probably
even lower. The majority of responses were from the United States, Canada and Europe,
which potentially limits our interpretation and generalizability to other regions.
The survey was only available in English, which could have led to selection bias for
certain regions of the world and for anglophone participants. However, we present
the largest survey available to date on the topic.
Conclusion
In conclusion, current uptake of resect-and-discard is very low (15.8 %) with most
endoscopists agreeing that such strategies are not feasible. Fear of making the wrong
diagnosis and potential medicolegal repercussions are cited amongst the main reasons
for difficulty of implementation. The development of simplified resect-and-discard
models will likely provide a solution to these barriers for implementing resect-and-discard
in clinical practise.
