| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, University of Wisconsin School of Medicine and Public
Health, E3/311 Clinical Science Center, 600 Highland Ave., Madison, WI
53792-3252.
2 Department of Radiology, Uniformed Services University of the Health Sciences,
Bethesda, MD.
3 Gastroenterology and Digestive Endoscopy Unit, "Nuovo Regina
Margherita" Hospital, Rome, Italy.
4 Department of Radiological Sciences, University of Rome La Sapienza
"Policlinico Umberto I," Rome, Italy.
Received April 3, 2008;
accepted after revision June 12, 2008.
The opinions and assertions contained herein are the private views of the
authors and are not to be construed as official or as reflecting the views of
the Department of the Navy or Department of Defense.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. A decision analysis model was constructed
incorporating the expected advanced neoplasia prevalence, frequency of
measurable growth, colorectal cancer (CRC) prevalence and risk, CTC
performance, and costs related to CRC screening and treatment. CRC risk was
assumed to be independent of advanced adenoma size, which intentionally
overestimates the risk related to small polyps. Clinical effectiveness and
costs for 3-year CTC surveillance versus immediate colonoscopic polypectomy
were compared for a concentrated cohort of patients with 6- to 9-mm polyps.
For the CTC surveillance strategy, only cases with measurable growth (
1
mm) at follow-up CTC were referred for polypectomy.
RESULTS. Without any intervention, the estimated 5-year CRC death rate from 6- to 9-mm polyps in this concentrated cohort was 0.08%, which is a sevenfold decrease over the 0.56% CRC risk for the general unselected screening population. The death rate was further reduced to 0.03% with the CTC surveillance strategy and to 0.02% with immediate colonoscopy referral. However, for each additional cancer-related death prevented with immediate polypectomy versus CTC follow-up, 9,977 colonoscopy referrals would be needed, resulting in 10 additional perforations and an incremental cost-effectiveness ratio of $372,853.
CONCLUSION. For patients with small (6- to 9-mm) polyps detected at
CTC screening, the exclusion of large polyps ( 10 mm) already confers a
very low risk of CRC. The high costs, additional complications, and relatively
low incremental yield associated with immediate polypectomy of 6- to 9-mm
polyps support the practice of 3-year CTC surveillance, which allows for
selective noninvasive identification of small polyps at risk.
Keywords: colorectal cancer • cost-effectiveness • CT colonography • polyps • screening • virtual colonoscopy
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Noninvasive strategies exploiting recent technologic advances may substantially improve the compliance rate without af fecting the sensitivity for detecting clinically relevant lesions. Among these emerging screening options, CT colonography (CTC) has shown excellent performance characteristics for the detection of advanced neoplasia when current techniques are applied [6, 7]. In fact, the recently revised colorectal screening guidelines of the American Cancer Society now include CTC as a recommended screening test, emphasizing its ability for prevention [3]. CTC offers the potential for selectively and noninvasively filtering out those patients who would benefit most from therapeutic colonoscopy [7–10].
Colonoscopic referral for diminutive polyps (
5 mm) detected at CTC has
been shown to be a highly ineffective approach because of the extremely low
likelihood of advanced neoplasia and the high costs associated with
polypectomy
[9–11].
In comparison, the clinical management of small (6- to 9-mm) polyps detected
at CTC remains controversial, in part because of the perceived lack of data
regarding the natural history of these lesions. Although the current clinical
practice is generally to offer polypectomy for small (6- to 9-mm) polyps
detected at CTC [3], many
patients will choose short-term CTC surveillance if presented with the option
[7,
12]. To date, no detrimental
effect has been shown from longitudinal follow-up of small colorectal polyps
in previous endoscopic and barium enema studies
[13–15],
suggesting that this may be a reasonable clinical approach. In fact, most
small polyps will either remain stable or decrease in size over time, which
agrees with our preliminary experience with CTC surveillance of 6- to 9-mm
polyps.
A previous decision analysis predicted an unexpectedly high rate of cancers resulting from short-term CTC follow-up of 6- to 9-mm polyps [16]. However, that analysis was largely based on data from high-risk or symptomatic cohorts and lacked appropriate natural history assumptions based on actual CTC observational surveillance studies. In addition, no cost-effectiveness analysis was performed, precluding evaluation of the cost efficiency for referring patients from CTC to colonoscopy.
To address the previous shortcomings in the knowledge of the natural history of small polyps, we recently developed a model in which the distribution of advanced neoplasia according to polyp size was projected on the CRC risk for the general population [10]. In addition, we now have initial input data from longitudinal CTC surveillance of 6- to 9-mm polyps that show that only a small percentage of small polyps, including the expected number of histologically advanced lesions, will show measurable interval growth (Pickhardt PJ et al., presented at the 2008 meeting of the Society of Gastrointestinal Radiologists). This observation should allow noninvasive distinction from stable lesions without concerning histologic features. It is highly unlikely that a nonadvanced subcentimeter tubular adenoma will progress to cancer over 5–10 years [17]; most cancers will likely arise from progression of adenomas that are already histologically advanced. To date, no assessment has been made of the cost-effectiveness of short-term CTC surveillance for small polyps, as currently recommended by radiology consensus guidelines [11]. The aim of this decision analysis was to compare the relative benefits, harms, costs, and resource utilization of short-term CTC surveillance for 6- to 9-mm lesions detected at CTC screening, compared with a policy of immediate colonoscopic referral for polypectomy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
To define the risk associated with 6- to 9-mm polyps, we related polyp
prevalence by age to the risk of developing cancer within a 5-year time
interval. We assumed that all potentially preventable CRCs arose from advanced
adenomas, which are defined by histologic features of a pro minent villous
component or high-grade dysplasia, or by size 10 mm
[19]. Advanced neoplasia in
small 6- to 9-mm polyps includes both histologically advanced adenomas, as
defined, and malignant polyps, in which cancer cells have penetrated beyond
the muscularis mucosae into the submucosa. CRC includes both malignant polyps
and frankly invasive masses. With regard to clinical staging of CRC, localized
disease refers to stage I or II; regional disease, to stage III; and distant
disease, to stage IV.
The CRC risk for advanced adenomas was assumed to be independent of lesion
size. Our baseline assumption of a 1.0% annual transition rate from a small
advanced adenoma to CRC (Table
1) was based on observational data with large polyps by Stryker et
al. [20], which showed a 2.5%
CRC progression rate over 5 years and an 8% rate over 10 years for large (
10 mm) unresected polyps. This conservative assumption of a 1% annual
progression rate was intentionally used to overestimate the risk related to
small polyps in order to ensure that the safety of the surveillance option is
not overstated. For example, a 6-mm tubulovillous adenoma in this model would
carry the same 5-year CRC risk as a 3-cm villous adenoma with high-grade
dysplasia. The prevalence data for polyps according to both size and histology
were based on a pooled data analysis of three large published screening
cohorts of asymptomatic adults who underwent colonoscopy, with or without
preceding CTC [6,
7,
21]
(Table 3). Because these series
represent the screening setting, they more closely reflect the true screening
prevalence of disease than do symptomatic or higher-risk cohorts.
|
Patients with a polyp were categorized and modeled according to their lesion of greatest clinical significance, which incorporates both lesion size and histology. A baseline 5-year CRC risk of 0.56% for the general 60-year-old population was derived from the Surveillance Epidemiology and End Results (SEER) data [22]. We assumed that 85% of CRCs would be preventable by detection and removal of advanced neoplasms, with a 15% residual risk due to either de novo, fast-growing, or missed lesions; or to an alternative unpreventable pathway distinct from the adenoma–carcinoma sequence [9]. As noted previously, the annual transition rate of 1% for small advanced adenomas developing into cancer was based on longitudinal natural history data from large polyps, thus somewhat overestimating the risk of small polyps [20].
In order not to underestimate the potential benefit of immediate polypectomy, the initial stage distribution for small (6- to 9-mm) malignant polyps was assumed to be localized in 95% of cases and regional in 5%, roughly corresponding to pooled data analyses [23, 24]. For prevalent small malignant polyps not immediately removed after CTC, we assumed a progressive up-staging related to the delay in cancer diagnosis (see Table 1). Such up-staging has been used in model simulations for other malignancies [25–28]. In order not to favor the CTC surveillance strategy, we did not simulate a slower rate of progression in the very early stages of a malignant polyp, as would be expected due to the exponential growth of cancers. The transition rates among the three CRC stages (localized, regional, distant) were extrapolated from previous cost-effectiveness models [29, 30] (Table 1). The same transition rates were also applied to cancers arising from advanced adenomas in the simulation period. Additional baseline assumptions are listed in Table 1.
Accuracy and Costs of the Diagnostic Procedures
The performance characteristics for CTC were predicated on the use of
current techniques, including 3D polyp detection and oral contrast tagging
[6,
31](Table
1). The baseline costs for colonoscopy and CTC were derived from
actual charges at our medical center and assume that costs are 40% of the
charge. Costs associated with endoscopic or CTC complications were also
accounted for in the model.
Clinical Efficacy and Cost-Effectiveness Analyses
The incremental cost-effectiveness ratio of 3-year CTC follow-up versus
immediate colonoscopy or no intervention for CTC-detected 6- to 9-mm polyps
was derived from a decision tree according to chance events that are
determined by the assigned probabilities
(Fig. 1). The product of the
probabilities along an outcome pathway determines the likelihood of each
outcome (i.e., life expectancy). The overall cost of an arm in the decision
tree can be calculated by weighting the cost of each outcome by its
probability and summing the results. In order not to underestimate the impact
of a newly diagnosed CRC on life expectancy, which would favor the CTC
surveillance strategy, we assumed all 5-year CRC mortality to occur within the
study interval, regardless of the timing of CRC detection.
A key cost-effectiveness measure was the difference in cost between the CTC follow-up and colonoscopy referral arms, divided by the difference in life expectancy, which gives the cost per life-year gained. We also considered the possibility of no intervention after initial polyp detection (i.e., neither follow-up CTC nor colonoscopy referral). An incremental cost-effectiveness ratio of $100,000 per life-year gained was considered to represent the threshold for determining a reasonably efficient versus costly option. No discounting was used because all the costs and benefits were assumed to occur within a few years from the beginning of the simulation.
Sensitivity Analysis
Sensitivity analysis was performed using two methods. First, a number of
key input variables were varied simultaneously and randomly for 10,000
interactions in a Monte Carlo simulation (Lumenaut software, version 3.4.9,
Lumenaut Ltd). This provides estimates of the variability in
cost-effectiveness, expressed as 10th–90th percentiles, that arise when
variables in the model are allowed to take on distributions. Second, a
systematic sensitivity analysis was performed based on the prevalence of
advanced adenomas and cancers among 6- to 9-mm polyps, cancer up-staging
rates, CTC performance characteristics, population age, and costs.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
When extrapolating these results to a general screening population of 100,000 subjects in whom the prevalence of 6- to 9-mm polyps is 8.5%, the numbers of CRC cases and CRC deaths related to small polyps are estimated at 10 (0.01%) and seven (0.007%) cases, respectively. This compares with 560 total CRC cases and 188 deaths from all cases of CRC in the general screening cohort, with nearly all of the remaining cancers arising from large polyps and masses.
Three-year CTC surveillance strategy— When simulating a policy of 3-year CTC follow up, 10,000 patients (10%) would be referred to colonoscopy for a growing 6- to 9-mm polyp. Considering the entire 5-year study interval, 87 (0.09%) patients would be diagnosed with CRC, corresponding to a CRC prevention rate of 23% (36% when excluding the theoretically unpreventable cancers). The CRCs in this arm were due to the 5-year progression of 162 CTC-undetected advanced adenomas, resulting in eight CRCs (9% of all CRCs); the 3-year progression of detected small advanced adenomas, leading to 40 CRCs (46%); and 39 small malignant polyps (45%).
The potential for cancer prevention through the removal of small advanced adenomas and for earlier detection of small malignant polyps with 3-year CTC surveillance positively affected the CRC-related mortality rate, which was reduced by 64% compared with no intervention, from 0.08% to 0.028%. The 28 deaths (representing a loss of 685 life-years) were due to CRC arising from advanced adenomas in 39% of the cases and to malignant polyps in the remaining cases. The corresponding numbers of 3-year CTC surveillance studies needed to be performed to prevent one case of cancer or one cancer-related death were 3,571 and 2,000, respectively.
The overall cost of the CTC follow-up program was $68,290,291 and was mainly due to procedural costs (CTC cost, 57%; colonoscopy cost, 35%) rather than to CRC treatment (8%) (Table 3). Performing a 3-year CTC follow-up for a 6- to 9-mm polyp compared with a policy of no intervention was a cost-effective option, with an incremental cost-effectiveness ratio of $50,418 per life-year gained. At Monte Carlo analysis, the 10th and 90th percentiles were $41,203 and $67,059, respectively, showing that even simultaneous large variations of all the included variables were unable to meaningfully affect the cost-effectiveness of this strategy.
Immediate colonoscopy referral—When simulating a policy of immediate colonos copy, the model predicts that only 50 (0.05%) patients would be diagnosed with CRC from small polyps in the 5-year study interval, corresponding to a reduction in CRC of 56% compared with no intervention and 42% compared with CTC follow-up. The higher prevention rate compared with CTC was related to the removal of detectable 6- to 9-mm advanced adenomas at the onset of the simulation, before the possibility of pro gression to CRC. Most of the CRC cases in this arm were due to malignant polyps (39 cases, 78%); the remaining were from undetected advanced adenomas that de veloped into CRC.
The 12 (0.012%) deaths specifically related to CRC in the colonoscopy referral arm resulted in the loss of 299 years of life. However, the increased utilization of colonoscopy resulted in six additional colonoscopy-related deaths. Therefore, the overall death rate for this approach was 0.018%, corresponding to overall reductions in the death rate of 76% and 35% compared with no intervention (0.08% death rate) and 3-year CTC surveillance (0.028% death rate), respectively. Compared with no intervention, the numbers of colonoscopy studies needed to be performed to prevent one case of cancer or one cancer-related death were 1,538 and 1,691, respectively. However, compared with the CTC surveillance approach, the numbers of colonoscopy studies needed to be performed to prevent one case of cancer or one cancer-related death were 2,684 and 9,977, respectively. The absolute benefit of the colonoscopy strategy in terms of life-years gained was fairly modest, especially when the results are projected to a general screening population.
The overall cost of the colonoscopy program was $157,984,989, and was mainly due to procedural costs (98.6%) and not to CRC treatment. Because of the high costs for the degree of absolute mortality reduction, performing an immediate colonoscopy for a 6- to 9-mm polyp when compared with a policy of no intervention was a relatively expensive option, with an incremental cost-effectiveness ratio of $104,456 (10–90% Monte Carlo analysis, $84,674–132,904). Immediate colonoscopy was not a cost-effective option compared with the 3-year CTC strategy, resulting in an incremental cost-effectiveness ratio of $372,853 per life-year gained (10–90% Monte Carlo analysis, $172,207–857,955) (Table 3).
Systematic Sensitivity Analysis
The incremental cost-effectiveness ratio between colonoscopy and 3-year CTC
follow-up depends on the simulated prevalence of advanced adenomas and
malignant polyps. A relatively large increase in the absolute prevalence of 6-
to 9-mm advanced neoplasms from the baseline assumption of 0.4% to 2.4% and
5.1% would be required to reduce the incremental cost-effectiveness ratio of
colonoscopy to less than $100,000 and $50,000, respectively. These increases
would correspond to a relative prevalence of advanced neoplasms among
individuals with 6- to 9-mm polyps of 28% and 60%, respectively, which is well
above actual clinical observations
[6,
7,
21]. Similarly, an increase in
the absolute prevalence of 6- to 9-mm malignant polyps from the baseline
assumption of 0.01% to 0.05% and 0.09%, corresponding to a relative prevalence
of 0.6% and 1.1%, was required to reduce the incremental cost-effectiveness
ratio of colonoscopy to less than $100,000 and $50,000, respectively.
Assuming a decrease in the percentage of advanced adenomas that show measurable growth from the baseline of 100% to 70% of lesions, the incremental cost-effectiveness ratio of colonoscopy relative to CTC surveillance would be lowered to $235,746. Under this condition, an increase in the absolute prevalence of either the 6- to 9-mm advanced adenomas to 1.3% or the 6- to 9-mm malignant polyps to 0.04% for the general screening population would be enough to reduce the incremental cost-effectiveness ratio to less than $100,000.
A 50% reduction of the annual 30% and 60% transition rates to regional and distant stages, respectively, would further reduce the benefit derived from colonoscopy referral, with an incremental cost-effectiveness ratio rising to $1,095,844. In this condition, only an absolute prevalence of either 3.7% (44% relative prevalence) for 6- to 9-mm advanced adenomas or an absolute prevalence of 0.1% (1.2% relative prevalence) for malignant polyps was sufficient to lower the incremental cost-effectiveness ratio to less than $100,000. Assuming a 75% reduction of the transition rates underlying cancer progression, an absolute prevalence of either 4.7% (55% relative) for 6- to 9-mm advanced polyps or 0.21% absolute (2.5% relative) prevalence for 6- to 9-mm malignant polyps was necessary to reduce the incremental cost-effectiveness ratio to less than $100,000. On the other hand, the output of the simulation was not meaningfully affected by the initial distribution of cancer stages in the malignant polyps.
Regarding procedural costs, a 50% reduct ion in endoscopic costs or, alternatively, a 270% increase in CTC costs, decreased the incremental cost-effectiveness ratio of colonoscopy compared with CTC to less than $100,000. With regard to the surgery rate of malignant polyps, even broad variations did not meaningfully affect the relative cost-effectiveness. With regard to patient age, variation of the age-specific 5-year CRC risk appeared to be more important than that of life expectancy. For instance, increasing the overall 5-year CRC risk to 1.09% and reducing life expectancy from 24 to 14 years, as is the case for a 70-year-old cohort, the incremental cost-effectiveness ratio of colonoscopy to CTC was reduced to $184,048. For a 50-year-old cohort, the incremental cost-effectiveness ratio was greater than $400,000.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
10 mm. We
thought that we must first show that 3-year CTC follow-up of small polyps is
safe because even this is a major deviation from a "leave no polyp
behind" mentality that stems primarily from optical colonoscopy
screening [32]. The high costs and additional complications associated with the immediate colonoscopy referral of all 6- to 9-mm polyps further support the practice of CTC surveillance. Our findings show that short-term CTC sur veillance of small polyps is a much more cost-effective strategy, with a staggering incre mental cost-effectiveness ratio for im mediate polypectomy of $372,853 per life-year-gained, corresponding to about 10,000 addi tional colonoscopy examin ations. In ad dition, any gain in clinical efficacy related to CRC mortality reduction would be largely negated by the endoscopic complications resulting from such a large number of invasive examinations. In fact, when the estimated 5-year CRC death rate from unresected 6- to 9-mm polyps in our study is applied to the general screening population as a whole, the absolute death rate (0.007%) is similar to the expected death rate related to comp lications at optical colonoscopy (0.006%). Therefore, immediate colonoscopy referral for all CTC-detected 6- to 9-mm polyps is not only an expensive strategy, but the low clinical yield provides an additional reason for seeking out a less aggressive strategy. In addition, optical colonoscopy is a limited resource that is already significantly tied up performing postpolypectomy surveillance. In comparison, colonoscopy referral for large polyps detected at CTC is a truly therapeutic indication that is both highly clinically efficacious and cost-effective [10].
The basic rationale behind CTC surveillance of 6- to 9-mm polyps is that this approach allows noninvasive identification of the small subset of these polyps that will show interval growth. The remainder of these polyps that are either stable in size or actually decrease in size probably carry a negligible CRC risk. Our preliminary findings from CTC surveillance of small polyps indicate that only about 10% show measurable interval growth and that all advanced adenomas (based on expected prevalence) show measurable growth at follow-up. These results, which provided key input assumptions for our decision analysis, share many general parallels with previous longitudinal endoscopic and barium enema natural history studies that evaluated subcentimeter polyps over time [13–15].
Our results differ significantly from the recent decision analysis by Hur
et al. [16]. One striking
difference was that their model estimated that 773 cancers would arise over a
3-year period from CTC surveillance of 100,000 subjects with 6- to 9-mm
polyps, whereas we found that only 115 cancers would arise over a longer
5-year period with no intervention. Key differences in a number of input
assumptions likely account for this sizable discrepancy. For one, Hur et al.
did not have the benefit of our actual CTC surveillance data, which were
critical to our model. For the relative prevalence of malignant polyps among
6- to 9-mm polyps, our assumption of 0.1% is well supported by a number of
studies evaluating screening cohorts
[6,
7,
21,
33–35]
that include more than 100,000 6- to 9-mm polyps. The 0.9% cancer rate assumed
by Hur et al. is an outdated figure that was largely based on high-risk and
symptomatic cohorts. Furthermore, we assumed a 1% annual progression of
advanced adenomas to localized cancer, compared with 5% assumed by Hur et al.
We based our 1% assumption on the barium enema surveillance work by Stryker et
al. [20], which found the 5-
and 10-year cancer risks for large ( 10 mm) unresected polyps to be 2.5%
and 8%, respectively, which implies that a 1% annual progression rate for 6-
to 9-mm polyps is considerably overestimated. However, the 5% annual cancer
risk for 6- to 9-mm polyps used by Hur et al. is double the cumulative 5-year
cancer rate for large polyps found by Stryker et al. At sensitivity analysis,
we confirmed that the relative prevalence of advanced neoplasia among 6- to
9-mm polyps and CRC progression transition rates were indeed key variables.
Finally, Hur et al. did not perform a cost-effectiveness analysis, which was a
critical component of our analysis. Including the cost considerations places
the extraordinarily high use of endoscopy into its proper context and shows
the downside of this approach, rather than simply reporting a net reduction in
cancer.
Our study has limitations. As with any model simulation, the results are highly dependent on the underlying input assumptions. By applying a number of conservative assumptions that tended to overestimate the importance of 6- to 9-mm polyps, we think that our results skewed toward a more aggressive nature for these lesions than that seen in actual practice. The number of cancers from 6- to 9-mm polyps may be considerably less in reality, but we sought to present a worst-case scenario. Our preliminary natural history data derived from CTC surveillance, which provided some key assumptions, represents ongoing research that must still be verified on a larger scale. Nonetheless, these are unique data that were previously unavailable.
In conclusion, our analysis shows that the very low CRC risk associated with 6- to 9-mm polyps detected at CTC screening argues against immediate colonoscopy referral. In addition, the high costs and additional complications associated with immediate colonoscopy referral for all 6- to 9-mm polyps also support the practice of CTC surveillance. On the basis of the findings from this hypothetic population of asymptomatic 60-year-old adults with 6- to 9-mm polyps, the most important factor for CRC risk reduction at CTC screening appeared to be exclusion of large polyps or masses at the baseline evaluation.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  P. J. Pickhardt and D. H. Kim Colorectal Cancer Screening With CT Colonography: Key Concepts Regarding Polyp Prevalence, Size, Histology, Morphology, and Natural History Am. J. Roentgenol., July 1, 2009; 193(1): 40 - 46. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. J. Pickhardt, C. Hassan, A. Laghi, and D. H. Kim CT Colonography to Screen for Colorectal Cancer and Aortic Aneurysm in the Medicare Population: Cost-Effectiveness Analysis Am. J. Roentgenol., May 1, 2009; 192(5): 1332 - 1340. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 How Should We Manage Patients with 6- to 9-mm Polyps Detected by CT Colonography? Journal Watch (General), November 18, 2008; 2008(1118): 3 - 3. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
