GASTROENTEROLOGY 2008;134:1570 –1595
`
`AGA INSTITUTE
`
`The following article contains new recommendations for colorectal cancer screening, the first set we have published since 2003 (Winawer S,
`Fletcher R, Rex D, et al. Colorectal cancer screening and surveillance: clinical guidelines and rationale— update based on new evidence.
`Gastroenterology 2003;124:544 –560.) The current recommendations have emerged through the participation of multiple national societies,
`taking into consideration newly emerging technologies. Please note the US Multi-Society Task Force (USMTF) represents the American
`Gastroenterological Association (AGA) Institute, the American Society for Gastrointestinal Endoscopy, and the American College of Gastro-
`enterology. Commissioned originally by the American Cancer Society, this compendium will be published concurrently in CA: A Cancer Journal
`for Clinicians and reprinted in the June issue of Radiology.
`
`Screening and Surveillance for the Early Detection of Colorectal Cancer
`and Adenomatous Polyps, 2008: A Joint Guideline From the American
`Cancer Society, the US Multi-Society Task Force on Colorectal Cancer,
`and the American College of Radiology
`
`BERNARD LEVIN,* DAVID A. LIEBERMAN,‡ BETH MCFARLAND,§ KIMBERLY S. ANDREWS,储 DURADO BROOKS,储
`JOHN BOND,## CHIRANJEEV DASH,¶ FRANCIS M. GIARDIELLO,# SETH GLICK,** DAVID JOHNSON,***
`C. DANIEL JOHNSON,‡‡‡ THEODORE R. LEVIN,‡‡ PERRY J. PICKHARDT,§§ DOUGLAS K. REX,储 储 ROBERT A. SMITH,储
`ALAN THORSON,¶¶ and SIDNEY J. WINAWER§§§ for the American Cancer Society Colorectal Cancer Advisory Group, the
`US Multi-Society Task Force, and the American College of Radiology Colon Cancer Committee
`
`*The University of Texas MD Anderson Cancer Center, Houston, Texas; ‡Division of Gastroenterology, Oregon Health and Science University, Portland Veterans Medical
`Center, Portland, Oregon; §Mallinckrodt Institute of Radiology, St Luke’s Hospital, Diagnostic Imaging Associates, Chesterfield, Missouri; 储Cancer Control Science
`Department, American Cancer Society, Atlanta, Georgia; ¶Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, Georgia; #Johns
`Hopkins University School of Medicine, Baltimore, Maryland; **University of Pennsylvania Health System, Philadelphia, Pennsylvania; ‡‡Gastroenterology Department, Kaiser
`Permanente Walnut Creek Medical Center, Walnut Creek, California; §§Radiology Department, University of Wisconsin Hospital and Clinics, Madison, Wisconsin;
`储 储Indiana University, Indianapolis, Indiana; ##Gastroenterology Section, Minneapolis Veterans Affairs Medical Center, Minneapolis, Minnesota; ***Eastern Virginia
`Medical School, Norfolk, Virginia; ‡‡‡Radiology Department, Mayo Clinic, Scottsdale, Arizona; ¶¶Section of Colon and Rectal Surgery, Creighton University School of
`Medicine, and University of Nebraska College of Medicine, Omaha, Nebraska; and §§§Memorial Sloan-Kettering Cancer Center, New York, New York
`
`In the United States, colorectal cancer (CRC) is the
`third most common cancer diagnosed among men
`and women and the second leading cause of death
`from cancer. CRC largely can be prevented by the
`detection and removal of adenomatous polyps, and
`survival is significantly better when CRC is diagnosed
`while still localized. In 2006 to 2007, the American
`Cancer Society, the US Multi-Society Task Force on
`Colorectal Cancer, and the American College of Ra-
`diology came together to develop consensus guide-
`lines for the detection of adenomatous polyps and
`CRC in asymptomatic average-risk adults. In this up-
`date of each organization’s guidelines, screening tests
`are grouped into those that primarily detect cancer
`early and those that can detect cancer early and also
`can detect adenomatous polyps, thus providing a
`greater potential for prevention through polypec-
`tomy. When possible, clinicians should make patients
`aware of the full range of screening options, but at a
`minimum they should be prepared to offer patients a
`choice between a screening test that primarily is ef-
`fective at early cancer detection and a screening test
`that is effective at both early cancer detection and
`cancer prevention through the detection and removal
`of polyps. It is the strong opinion of these 3 organi-
`
`zations that colon cancer prevention should be the
`primary goal of screening.
`
`I n the United States, colorectal cancer (CRC) is the
`
`third most common cancer diagnosed in men and
`women and the second leading cause of death from
`cancer.1 In 2008, it is estimated that 148,810 men and
`women will be diagnosed with CRC and 49,960 will die
`from this disease.1 Five-year survival is 90% if the disease
`is diagnosed while still localized (ie, confined to the wall
`of the bowel) but only 68% for regional disease (ie, disease
`
`Abbreviations used in this paper: ACR, American College of Radiol-
`ogy; ACRIN, American College of Radiology Imaging Network; ACS,
`American Cancer Society; CRC, colorectal cancer; CSPY, colonoscopy;
`CT, computed tomography; CTC, computed tomographic colonogra-
`phy; DCBE, double-contrast barium enema; DIA, DNA integrity analysis;
`FIT, fecal immunochemical test; FOBT, fecal occult blood test; FSIG,
`flexible sigmoidoscopy; gFOBT, guaiac-based fecal occult blood test;
`HPNCC, hereditary nonpolyposis colon cancer; MRI, magnetic reso-
`nance imaging; NRDR, National Radiology Data Register; OC, optical
`colonoscopy; sDNA, stool DNA test; 2D, 2-dimensional; 3D, 3-dimen-
`sional; USMSTF, US Multi-Society Task Force on Colorectal Cancer.
`© 2008 by the AGA Institute and American Cancer Society, Inc
`0016-5085/08/$34.00
`doi:10.1053/j.gastro.2008.02.002
`
`AGA
`
`INSTITUTE
`
`Geneoscopy Exhibit 1013, Page 1
`
`

`

`May 2008
`
`AGA INSTITUTE 1571
`
`INSTITUTE
`
`AGA
`
`with lymph node involvement) and only 10% if distant
`metastases are present.2 Recent trends in CRC incidence
`and mortality reveal declining rates, which have been
`attributed to reduced exposure to risk factors, the effect
`of screening on early detection and prevention through
`polypectomy, and improved treatment.3 However, in the
`near term, even greater incidence and mortality reduc-
`tions could be achieved if a greater proportion of adults
`received regular screening. Although prospective ran-
`domized trials and observational studies have demon-
`strated mortality reductions associated with early detec-
`tion of
`invasive disease, as well as
`removal of
`adenomatous polyps,4 –7 a majority of US adults are not
`receiving regular age- and risk-appropriate screening or
`have never been screened at all.8,9
`The goal of cancer screening is to reduce mortality
`through a reduction in incidence of advanced disease. To
`this end, modern CRC screening can achieve this goal
`through the detection of early-stage adenocarcinomas
`and the detection and removal of adenomatous polyps,
`the latter generally accepted as a nonobligate precursor
`lesion. Adenomatous polyps are common in adults over
`age 50 years, but the majority of polyps will not develop
`into adenocarcinoma; histology and size determine their
`clinical importance.10,11 The most common and clinically
`important polyps are adenomatous polyps, which repre-
`sent approximately one half to two thirds of all colorectal
`polyps and are associated with a higher risk of CRC.
`Thus, most CRC screening studies evaluate the detection
`rate of invasive CRCs as well as advanced adenomas,
`which conventionally are defined as polyps ⱖ10 mm or
`histologically having high-grade dysplasia or significant
`villous components. The evidence for the importance of
`colorectal polyps in the development of CRC is largely
`indirect, but nonetheless extensive and convincing, and
`has been described in detail.11–13
`Today there is a range of options for CRC screening in
`the average-risk population, with current technology fall-
`ing into 2 general categories: stool tests, which include
`tests for occult blood or exfoliated DNA, and structural
`exams, which include flexible sigmoidoscopy (FSIG),
`colonoscopy (CSPY), double-contrast barium enema
`(DCBE), and computed tomographic colonography
`(CTC). Stool tests are best suited for the detection of
`cancer, although they also will deliver positive findings
`for some advanced adenomas, while the structural exams
`can achieve the dual goals of detecting adenocarcinoma
`as well as identifying adenomatous polyps.14 These tests
`may be used alone or in combination to improve sensi-
`tivity or, in some instances, to ensure a complete exam-
`ination of the colon if the initial test cannot be com-
`pleted. Although screening tests for CRC vary in terms of
`the degree of supporting evidence, potential efficacy for
`incidence and mortality reduction, cost-effectiveness, and
`acceptability, any one of these options applied in a sys-
`
`tematic program of regular screening has the potential to
`significantly reduce deaths from CRC.
`Beginning in 1980, the American Cancer Society (ACS)
`first issued formal guidelines for CRC screening in aver-
`age-risk adults.15 Since then, the ACS has periodically
`updated its CRC guidelines,16 –19 including adding rec-
`ommendations for high-risk individuals in 1997.17 Other
`organizations also have issued recommendations for
`CRC screening, most notably the US Preventive Services
`Task Force,20,21 the American College of Radiology
`(ACR),22,23 and the US Multi-Society Task Force on Colo-
`rectal Cancer (USMSTF).12,24 Recently, the ACS and the
`USMSTF collaborated on an update of earlier recommen-
`dations for postpolypectomy and post-CRC resection
`surveillance in response to reports suggesting significant
`deviation from existing recommendations.25,26 Since
`1997,
`the organizational guidelines for average-risk
`adults have grown increasingly similar and represent a
`broad organizational consensus on the value, options,
`and methods for periodic screening for CRC.
`In the last decade, there has been an increase in the
`number of technologies available for CRC screening, and
`in the case of stool tests, there has been growth in the
`number of commercial versions of guaiac-based and im-
`munochemical-based stool tests (gFOBT and FIT). This
`growth in options also has been accompanied by chang-
`ing patterns in the proportion of adults using different
`tests, with FSIG rates declining, CSPY rates increasing,
`use of stool blood tests remaining somewhat constant,
`and use of the DCBE for screening now becoming very
`uncommon.8
`There are pros and cons to having a range of options
`for CRC screening. Despite the fact that the primary
`barriers to screening are lack of health insurance, lack of
`physician recommendation, and lack of awareness of the
`importance of CRC screening,27 the historical evidence
`shows that adults have different preferences and patterns
`of use among the available CRC screening tests.28 –31
`Although population preferences or resistance to a par-
`ticular technology may change over time or may be in-
`fluenced by referring physicians, it also may be true that
`over time some adults may persist in choosing one tech-
`nology and rejecting another. Furthermore, at this time
`not all options are available to the entire population, and
`transportation, distance, and financial barriers to some
`screening technologies may endure for some time. Al-
`though in principle all adults should have access to the
`full range of options for CRC screening, the fact that
`simpler, lower-cost options are available in most settings,
`whereas other more costly options are not universally
`available, is a public health advantage. However, for av-
`erage-risk adults, multiple testing options challenge the
`referring physician to support an office policy that can
`manage a broad range of testing choices, their follow-up
`requirements, and shared decision making related to the
`options. Shared decision making for multiple screening
`
`Geneoscopy Exhibit 1013, Page 2
`
`

`

`1572 LEVIN ET AL
`
`GASTROENTEROLOGY Vol. 134, No. 5
`
`choices is both demanding and time consuming and is
`complicated by the different characteristics of the tests
`and the test-specific requirements for individuals under-
`going screening.31 In addition, the description of benefits
`is complicated by different performance characteristics of
`the variants of the occult blood tests and uncertain
`differences between test performance in research settings
`and test performance in clinical practice. These chal-
`lenges have been discussed in the past,19,32 and they still
`are with us today.
`In this guideline review, we have reassessed the indi-
`vidual test evidence and comparative evidence for stool
`tests, including gFOBT, FIT, and stool DNA test (sDNA),
`and the structural exams, including FSIG, CSPY, DCBE,
`and CTC, the latter also known as virtual colonoscopy.
`We have sought to address a number of concerns about
`the complexity of offering multiple screening options
`and the degree to which the range of screening options
`and their performance, costs, and demands on individu-
`als poses a significant challenge for shared decisions. An
`overriding goal of this update is to provide a practical
`guideline for physicians to assist with informed decision
`making related to CRC screening. These guidelines are
`for individuals at average risk. Individuals with a per-
`sonal or family history of CRC or adenomas, inflamma-
`tory bowel disease, or high-risk genetic syndromes should
`continue to follow the most recent recommendations for
`individuals at increased or high risk.24 –26
`
`Guidelines Development, Methods, and
`Framework
`
`The guidelines update process was divided into 2
`phases. The first phase focused on the stool tests, includ-
`ing gFOBT, FIT, and sDNA. The second phase of the
`guidelines update process focused on the structural ex-
`ams, including FSIG, colonoscopy, DCBE,and CTC. De-
`liberations about evidence and presentations from ex-
`perts took place during 2 face-to-face meetings of the the
`collaborating organizations and invited outside experts
`and through periodic conference calls. The process relied
`on earlier evidence-based reviews.12,16 –21,24 Literature re-
`lated to CRC screening and specific to individual tests
`published between January 2002 and March 2007 was
`identified using MEDLINE (National Library of Medi-
`cine) and bibliographies of identified articles. Expert
`panel members also provided several unpublished ab-
`stracts and manuscripts. Where evidence was insufficient
`or lacking to provide a clear, evidence-based conclusion,
`final recommendations were based on expert opinion and
`are so indicated.
`While there is clear experimental evidence that screen-
`ing for CRC with gFOBT is associ-ated with reduced
`incidence and mortality from CRC screening,5,6,33 most
`of the information supporting the use of the other colo-
`rectal screening tests is based on observational and infer-
`ential evidence. In this review, priority was placed on
`
`studies of asymptomatic average-risk or higher-risk pop-
`ulations that were followed by testing with colonoscopy
`in all or nearly all study participants as a validation
`measure.
`
`Summary of the Recommendations
`
`In this update of guidelines for CRC screening in
`average-risk adults, the expert panel concluded that a
`screening test must be able to detect the majority of
`prevalent or incident cancers at the time of testing. Here
`we are drawing a new, important distinction between test
`sensitivity and program sensitivity, the former being the
`sensitivity achieved in a single test and the latter being
`the sensitivity achieved over time through serial testing in
`a program. While cancer screening tests are expected to
`achieve acceptable levels of sensitivity and specificity,34
`no specific acceptance threshold for either measure, alone
`or in combination, has been established for any screening
`test.35,36 Thus, this criterion is based on expert opinion
`and the following considerations. First, in the judgment
`of the panel, recent evidence has revealed an unacceptably
`wide range of sensitivity among some gFOBT strategies,
`with some practices and tests performing so poorly that
`the large majority of prevalent cancers are missed at the
`time of screening.37–39 The observation of very low sen-
`sitivity for cancer and advanced neoplasia associated with
`in-office gFOBT led Sox to speculate that CRC mortality
`rates might be considerably lower today if the quality of
`gFOBT testing during the previous decade had been
`higher.40 While the literature on other CRC screening
`tests also reveals a range of sensitivities, even in the
`presence of significant, correctable, quality-related short-
`comings, the majority of invasive cancers still will be
`detected. Second, a test like gFOBT that demonstrates
`poor test sensitivity but good program sensitivity de-
`pends on high rates of adherence with regular screening.
`However, many patients have only one test and do not
`return the following year for programmatic testing.41,42
`Given the lack of systems to ensure or at least facilitate
`adherence with recommended regular screening intervals,
`as well as evidence of suboptimal awareness and engage-
`ment of primary care in supporting adherence with
`screening recommendations,43 the panel concluded that
`it was not realistic at this time to rely on program
`sensitivity to overcome limitations in test sensitivity. Phy-
`sicians and institutions should select stool blood tests
`that have been shown in the scientific literature to detect
`the majority of prevalent CRCs in an asymptomatic pop-
`ulation. If there is not evidence that an available test has
`met that benchmark, it should not be offered to patients
`for CRC screening.
`Individuals and health care professionals should also
`understand that screening tests for CRC broadly fall into
`2 categories. In one category are the fecal tests (ie,
`gFOBT, FIT, and sDNA), which are tests that primarily
`are effective at identifying CRC. Some premalignant ad-
`
`AGA
`
`INSTITUTE
`
`Geneoscopy Exhibit 1013, Page 3
`
`

`

`May 2008
`
`AGA INSTITUTE 1573
`
`INSTITUTE
`
`AGA
`
`Table 1. Testing Options for the Early Detection of
`Colorectal Cancer and Adenomatous Polyps for
`Asymptomatic Adults Aged 50 Years and Older
`
`Tests that detect adenomatous polyps and cancer
`FSIG every 5 years, or
`CSPY every 10 years, or
`DCBE every 5 years, or
`CTC every 5 years
`Tests that primarily detect cancer
`Annual gFOBT with high test sensitivity for cancer, or
`Annual FIT with high test sensitivity for cancer, or
`sDNA, with high sensitivity for cancer, interval uncertain
`
`enomatous polyps may be detected, providing an oppor-
`tunity for polypectomy and the prevention of CRC, but
`the opportunity for prevention is both limited and inci-
`dental and is not the primary goal of CRC screening with
`these tests. In the second category are the partial or full
`structural exams (ie, FSIG, CSPY, DCBE, and CTC)44
`which are tests that are effective at detecting cancer and
`premalignant adenomatous polyps. These tests differ in
`complexity and accuracy for the detection of CRC and
`advanced neoplasia. When performed properly, each of
`these structural exams has met the standard of detecting
`at least half of prevalent or incident cancers at the time of
`testing.
`It is the strong opinion of this expert panel that colon
`cancer prevention should be the primary goal of CRC
`screening. Tests that are designed to detect both early
`cancer and adenomatous polyps should be encouraged if
`resources are available and patients are willing to un-
`dergo an invasive test. These tests include the partial or
`full structural exams mentioned above. These tests re-
`quire bowel preparation and an office or hospital visit
`and have various levels of risk to patients. These tests also
`have limitations, greater patient requirements for suc-
`cessful completion, and potential harms. Significant pos-
`itive findings on FSIG, DCBE, and CTC require follow-up
`CSPY.
`The panel recognized that some patients will not want
`to undergo an invasive test that requires bowel prepara-
`tion, may prefer to have screening in the privacy of their
`home, or may not have access to the invasive tests due to
`lack of coverage or local resources. Collection of fecal
`samples for blood or DNA testing can be performed at
`home without bowel preparation. However, providers
`and patients should understand the following limita-
`tions and requirements of noninvasive tests:
`
`&# These tests are less likely to prevent cancer com-
`pared with the invasive tests;
`
`&# These tests must be repeated at regular intervals to
`be effective;
`
`&# If the test is abnormal, an invasive test (CSPY) will
`be needed.
`
`If patients are not willing to have repeated testing or
`have CSPY if the test is abnormal, these programs will
`not be effective and should not be recommended.
`Based on our review of the historic and recent evi-
`dence, the tests in Table 1 are acceptable options for the
`early detection of CRC and adenomatous polyps for
`asymptomatic adults aged 50 years and older (also see
`Table 2).
`
`Screening Tests for the Detection of
`CRC
`Stool Blood Tests— gFOBT and FIT
`
`Stool blood tests are conventionally known as
`fecal occult blood tests (FOBT) because they are designed
`to detect the presence of occult blood in stool. FOBT fall
`into 2 primary categories based on the detected analyte:
`gFOBT and FIT. Blood in the stool is a nonspecific
`finding but may originate from CRC or larger (⬎1 to 2
`cm) polyps. Because small adenomatous polyps do not
`tend to bleed and bleeding from cancers or large polyps
`may be intermittent or simply not always detectable in a
`single sample of stool, the proper use of stool blood tests
`requires annual testing that consists of collecting speci-
`mens (2 or 3, depending on the product) from consecu-
`tive bowel movements.18,24,45 FIT generally are processed
`only in a clinical laboratory, whereas gFOBT are pro-
`cessed either in the physician’s office or in a clinical
`laboratory. When performed for CRC screening, a posi-
`tive gFOBT or FIT requires a diagnostic workup with
`CSPY to examine the entire colon in order to rule out the
`presence of cancer or advanced neoplasia.
`
`gFOBT
`
`gFOBT are the most common stool blood tests in
`use for CRC screening and the only CRC screening tests
`for which there is evidence of efficacy from prospective,
`randomized controlled trials. Guaiac-based tests detect
`blood in the stool through the pseudoperoxidase activity
`of heme or hemoglobin, while immunochemical-based
`tests react to human globin. The usual gFOBT protocol
`consists of collecting 2 samples from each of 3 consecu-
`tive bowel movements at home. Prior to testing with a
`sensitive guaiac-based test, individuals usually will be
`instructed to avoid aspirin and other nonsteroidal anti-
`inflammatory drugs, vitamin C, red meat, poultry, fish,
`and some raw vegetables because of diet-test interactions
`that can increase the risk of both false-positive and false-
`negative (specifically, vitamin C) results.46 Collection of
`all 3 samples is important because test sensitivity im-
`proves with each additional stool sample.14
`gFOBT—Efficacy and Test Performance. Three
`large, prospective, randomized controlled trials with
`gFOBT have demonstrated that screened patients have
`cancers detected at an early and more curable stage than
`unscreened patients. Over time (8 –13 years), each of the
`
`Geneoscopy Exhibit 1013, Page 4
`
`

`

`1574 LEVIN ET AL
`
`GASTROENTEROLOGY Vol. 134, No. 5
`
`Table 2. Guidelines for Screening for the Early Detection of Colorectal Cancer and Adenomas for Average-risk Women and
`Men Aged 50 Years and Older
`
`The following options are acceptable choices for colorectal cancer screening in average-risk adults beginning at age 50 years. Since each of
`the following tests has inherent characteristics related to prevention potential, accuracy, costs, and potential harms, individuals should
`have an opportunity to make an informed decision when choosing one of the following options.
`In the opinion of the guidelines development committee, colon cancer prevention should be the primary goal of colorectal cancer screening.
`Tests that are designed to detect both early cancer and adenomatous polyps should be encouraged if resources are available and
`patients are willing to undergo an invasive test.
`
`Tests that Detect Adenomatous Polyps and Cancer
`
`Test
`
`FSIG with insertion to 40 cm or to
`splenic flexure
`
`Interval
`
`Every 5 years
`
`CSPY
`
`DCBE
`
`CTC
`
`Every 10 years
`
`Every 5 years
`
`Every 5 years
`
`Tests that Primarily Detect Cancer
`
`Test
`
`gFOBT with high sensitivity for cancer
`
`Interval
`
`Annual
`
`FIT with high sensitivity for cancer
`
`Annual
`
`sDNA with high sensitivity for cancer
`
`Interval uncertain
`
`Key Issues for Informed Decisions
`
`● Complete or partial bowel prep is required
`● Sedation usually is not used, so there may be some discomfort during the
`procedure
`● The protective effect of sigmoidoscopy is primarily limited to the portion of
`the colon examined
`● Patients should understand that positive findings on sigmoidoscopy
`usually result in a referral for CSPY
`● Complete bowel prep is required
`● Conscious sedation is used in most centers; patients will miss a day of
`work and will need a chaperone for transportation from the facility
`● Risks include perforation and bleeding, which are rare but potentially
`serious; most of the risk is associated with polypectomy
`● Complete bowel prep is required
`● If patients have one or more polyps ⬎6 mm, CSPY will be recommended;
`follow-up CSPY will require complete bowel prep
`● Risks of DCBE are very low; rare cases of perforation have been reported
`● Complete bowel prep is required
`● If patients have one or more polyps ⬎6 mm, CSPY will be recommended;
`if same day CSPY is not available, a second complete bowel prep will be
`required before CSPY
`● Risks of CTC are very low; rare cases of perforation have been reported
`
`Key Issues for Informed Decisions
`
`● Depending on manufacturer’s recommendations, 2 to 3 stool samples
`collected at home are needed to complete testing; a single sample of
`stool gathered during a digital exam in the clinical setting is not an
`acceptable stool test and should not be done
`● Positive tests are associated with an increased risk of colon cancer and
`advanced neoplasia; CSPY should be recommended if the test results are
`positive
`● If the test is negative, it should be repeated annually
`● Patients should understand that one-time testing is likely to be ineffective
`● An adequate stool sample must be obtained and packaged with
`appropriate preservative agents for shipping to the laboratory
`● The unit cost of the currently available test is significantly higher than
`other forms of stool testing
`● If the test is positive, CSPY will be recommended
`● If the test is negative, the appropriate interval for a repeat test is
`uncertain
`
`FSIG, flexible sigmoidoscopy; CSPY, colonoscopy; DCBE, double-contrast barium enema; CTC, computed tomography colonography; gFOBT,
`guaiac-based fecal occult blood test; FIT, fecal immunochemical test; sDNA, stool DNA test.
`
`trials demonstrated significant reductions in CRC mor-
`tality of 15% to 33%.5,6,34 Moreover, incidence reduction
`of 20% was demonstrated in one trial (Minnesota) after
`18 years of follow-up, which has been attributed to rela-
`tively higher rates of CSPY in the study (38% of subjects
`in the screened group).7
`The sensitivity and specificity of a gFOBT has been
`shown to be highly variable and varies based on the brand
`
`or variant of the test;47 specimen collection technique;38
`number of samples collected per test;14 whether or not the
`stool specimen is rehydrated (ie, adding a drop of water to
`the slide window before processing);48; and variations in
`interpretation, screening interval, and other factors.46
`The reported sensitivity of a single gFOBT varies con-
`siderably. In a review by Allison and colleagues, sensitivity
`for cancer ranged from 37.1% for unrehydrated Hemoc-
`
`AGA
`
`INSTITUTE
`
`Geneoscopy Exhibit 1013, Page 5
`
`

`

`May 2008
`
`AGA INSTITUTE 1575
`
`INSTITUTE
`
`AGA
`
`cult II to 79.4% for Hemoccult SENSA.47 Lieberman and
`Weiss
`compared one-time
`testing with rehydrated
`Hemoccult II and observed 35.6% sensitivity for cancer.14
`In a study comparing gFOBT (unrehydrated Hemoccult
`II) with sDNA, sensitivity for cancer was only 12.9%.37
`More recently, Allison and colleagues compared a high-
`sensitivity gFOBT (Hemoccult SENSA) with an FIT and
`observed 64.3% sensitivity for cancer and 41.3% for ad-
`vanced adenomas.49 Thus, the data reveal a range of
`performance among gFOBT variants that allows them to
`be grouped into low and high test sensitivity groups. The
`specificity of gFOBT also is variable, with low test sensi-
`tivity gFOBT (such as Hemoccult II) tending to have very
`high specificity and high test sensitivity gFOBT (such as
`Hemoccult SENSA) having lower specificity. In a com-
`parison of various stool blood tests, Allison and col-
`leagues observed specificity for cancer and advanced ad-
`enomas of 97.7% and 98.1%, respectively, for Hemoccult
`II, with a combined specificity for cancer and advanced
`adenomas of 98.1%. For Hemoccult SENSA, which had
`greater sensitivity for cancer and advanced adenomas
`compared with Hemoccult II, specificity for cancer and
`advanced adenomas was 86.7% and 87.5%, respectively,
`with a combined specificity for cancer and advanced
`adenomas of 87.5%.47
`A significant limitation of the potential of testing with
`gFOBT is that it is commonly performed in the physi-
`cian’s office as a single-panel test following a digital rectal
`exam.39 In a recent national survey of primary care phy-
`sicians, 31.2% reported using only the in-office method of
`gFOBT, and an additional 41.2% of physicians reporting
`using both the in-office method or the take-home
`method. While this approach may seem pragmatic, Col-
`lins et al demonstrated that sensitivity is only 4.9% for
`advanced neoplasia and only 9% for cancer.38 The accu-
`racy of this method is so low that it cannot, under any
`circumstances or rationale of convenience, be endorsed as
`a method of CRC screening.
`An additional limitation observed in the current use of
`gFOBT is inadequate follow-up of a positive test. Despite
`the fact that all existing CRC screening guidelines rec-
`ommend CSPY follow-up of a positive gFOBT, in the
`same survey that revealed high rates of in-office gFOBT,
`nearly one third of physicians reported that they followed
`up a positive gFOBT with a repeat gFOBT, and a sub-
`stantial percentage reported that they referred patients to
`sigmoidoscopy rather than CSPY after a positive gFOBT.
`Similar patterns of testing and response to positive test
`results have been reported by patients undergoing at-
`home screening.39
`gFOBT—Benefits, Limitations, and Harms. An-
`nual testing with gFOBT has been shown to reduce both
`CRC mortality and incidence. Testing for occult blood is
`simple and is associated with minimal harm, although
`any testing with gFOBT is associated with a possibility of
`a positive test result that will require follow-up with
`
`CSPY, which is associated with a greater risk of harm.
`The limitation of gFOBT is that many of the individual
`tests have limited test sensitivity under the best of cir-
`cumstances, and this sensitivity may be further compro-
`mised by poor and incomplete specimen collection and
`inadequate or improper processing and interpretation.
`Program sensitivity (ie, the outcome of repeat annual
`testing) is considerably higher, but the systems to ensure
`regular annual testing often are not in place to support
`either the patient or his or her physician to be adherent.
`Further, testing in the office following a digital rectal
`exam, which is highly inaccurate, has been common and
`still may persist at significant levels today. When the test,
`the testing procedure, or both have very low test sensi-
`tivity, and when positive tests are not followed up with
`CSPY, the potential is high for patients to have a false
`sense of reassurance after testing. Finally, patients who
`choose gFOBT for CRC screening must understand that
`annual testing is required.
`Quality Assurance. If patients and their provid-
`ers select gFOBT for CRC screening, they should be
`aware of several quality issues based on programmatic
`performance in clinical trials. First, the test must be
`performed properly with 3 stool samples obtained at
`home. A single stool sample FOBT collected after digital
`rectal exam in the office is not an acceptable screening
`test, and it is not recommended. Prior to testing with a
`sensitive guaiac-based test,
`individuals should be in-
`structed to avoid nonsteroidal anti-inflammatory drugs
`such as ibuprofen, naproxen, or aspirin (more than one
`adult aspirin per day) for 7 days prior to testing unless
`they are on a cardioprotective regimen. There has been
`debate as to whether additional dietary restrictions re-
`duce compliance with testing and are necessary to reduce
`the risk of both false-negative and false-positive results.
`Results of a meta-analysis that examined completion and
`positivity results found little support for the influence of
`dietary restrictions on