Aliment Pharmacol Ther 2004; 19: 1225–1233.
`
`doi: 10.1111/j.1365-2036.2004.02005.x
`
`Review article: the evolution to stool DNA testing for colorectal cancer
`
`K . S . T A G O R E * , T . R . L E V I N  & M . J . LA W SO Nà
`*University of California Davis Medical Center, Sacramento, CA, USA;  Kaiser Permanente, Walnut Creek, CA and Division
`of Research, Oakland, CA, USA; àKaiser Permanente, Sacramento, CA, USA
`
`Accepted for publication 13 April 2004
`
`S U MM AR Y
`
`Despite a variety of screening strategies and recent trends
`showing death rate stabilization, colorectal cancer still
`remains the second leading cause of overall cancer death.
`Current screening tools suffer from performance limita-
`tions, low patient acceptability, and marginal reliable
`access within the health care system. Noninvasive
`strategies present
`the lowest risk with the highest
`potential for patient satisfaction. However, serious imple-
`mentation barriers exist requiring consistent program-
`matic screening, strict patient adherence, and poor
`sensitivity for adenomas. Colonoscopy remains an inva-
`sive screening test with the best sensitivity and specificity,
`but faces large financial costs, manpower requirements,
`
`patient access and adherence. Development of advanced
`molecular techniques identifying altered DNA markers in
`exfoliated colonocytes signify early or precancerous
`growth. Stool-based DNA testing provides an entirely
`noninvasive population-based screening strategy which
`patients can perform easier than faecal occult blood
`testing (FOBT). Large-scale prospective randomized con-
`trol trials currently pending should help characterize
`accurate test performance, screening intervals, cost-
`effectiveness, direct comparison to FOBT and analysis of
`patient adherence. As tumour development pathways
`and potential target genes are further elucidated, refine-
`ments in multi-assay stool-based DNA testing portend
`enhanced test characteristics to detect and treat this
`genetically heterogeneous disease.
`
`I N T R O DU C T IO N
`
`Colorectal cancer (CRC) remains the second leading
`cause of overall cancer death in the United States
`approaching an estimated 150 000 new cases and
`60 000 deaths per year.1 An average lifetime risk in the
`population of 6%, a 90% cure rate in early disease
`(stage 1), and a highly accessible target organ comprise
`sensible reasons mass CRC screening reduces cancer
`mortality. Early detection of CRC reduces CRC mortal-
`ity.2 Finding and removing precursor adenomas reduces
`CRC incidence and, in high risk groups, CRC mortality.3
`However, only 37% of cases are diagnosed while still
`localized,4 suggesting that current screening options are
`
`Correspondence to: Dr M. J. Lawson, The Permanente Medical Group, Inc.,
`2025 Morse Avenue, Sacramento, CA 95825, USA.
`E-mail: michael.j.lawson@kp.org
`
`ineffective either because of inadequate sensitivity, low
`acceptability to patients, or high resource demands
`limiting widespread availability. The lack of consensus
`on a single best screening option reflects the limitations
`of all currently available strategies.
`As molecular understanding of disease has enhanced
`approaches
`to modern medicine, population-based
`screening for CRC will similarly evolve. In addition to
`point mutations in the oncogene k-ras, allele losses of
`tumour suppressor genes are involved in the pathogen-
`esis of CRC. These and other specific DNA alterations
`further identified in CRC progression became known as
`the adenoma-carcinoma sequence proposed in 1990.5
`Thereafter,
`several
`investigators recovered mutated
`DNA from the stool of patients with advanced colorectal
`neoplasia.6 The inability to identify one single mutation
`expressed uniformly across all colorectal neoplasia
`
`Ó 2004 Blackwell Publishing Ltd
`
`1225
`
`Geneoscopy Exhibit 1039, Page 1
`
`

`

`1226
`
`K . S. TA G O R E et al.
`
`reflects the genetic heterogeneity of the disease. Target-
`ing multiple DNA mutations has been essential
`in
`improving detection rates and false-positive rates have
`been lowered by using neoplasm-specific markers.
`Developments in understanding the molecular path-
`ways of colorectal tumorigenesis will lead to modifica-
`tions that can only further enhance the utility of such a
`screening modality.
`Proponents of stool-based DNA testing describe the
`sound biological rationale for targeting DNA markers in
`stool.7–12 This review will briefly examine limitations of
`current screening techniques, summarize the develop-
`ment of stool-based DNA testing and present the data
`from early clinical studies.
`
`C U R R E NT S C R E E N IN G S T R ATE G I E S
`
`Performance limitations of existing screening practices
`have resulted in a limited overall
`impact on CRC
`mortality. Reliance on symptoms suggestive of CRC
`can delay diagnosis and have a very low predictive
`value.13 Programmatic faecal occult blood testing
`(FOBT) every 1–2 years leads to a 15–33% reduction
`in cumulative CRC mortality.2, 14, 15 However, FOBT
`sensitivity for advanced adenomas and early-stage
`lesions limits its impact on mortality. Adenomas rarely
`bleed placing FOBT sensitivities in the low 20% range.16
`False-positive or indeterminate results lead to costly
`unnecessary and invasive tests. This may explain the
`failure of doctors to recommend colonoscopy in greater
`than half of the patients with positive FOBT, markedly
`reducing screening effectiveness.
`Compared with guaiac FOBTs, immunochemical FOBTs
`(iFOBTs) have higher sensitivity and equivalent specif-
`icity.17 Advantages of iFOBT include fewer specimens
`needed,
`less stool handling, and the test specifically
`detects human haemoglobin requiring no dietary
`restrictions.18 Blood released from a supra-colonic
`source will not cause a falsely positive test. An
`automated tracking system is used along with this test
`to ensure that participants are reminded to screen
`continuously on an annual basis.
`Flexible sigmoidoscopy (FS)
`identifies 50–70% of
`patients with advanced neoplasms. Four case–control
`studies found a reduction in distal CRC mortality with
`the protective effect persisting up to 10 years.19–22
`Current recommendations for screening CRC include an
`option for FS every 5 years; however, repeat perform-
`ance in this unsedated procedure is less preferred over
`
`other methods23 and refused in up to 10% of patients24
`who perceive discomfort.
`In most
`series, 50% of
`significant proximal neoplasms were not associated
`with a distal marker lesion detectable by FS.16, 25 The
`combination of a single application of FOBT and FS has
`sensitivity for advanced colonic neoplasia of 77%.26
`Repeated annual applications of FOBT would improve
`sensitivity, but depends on patient and doctor adher-
`ence.
`Colonoscopy is typically recommended to be performed
`at 10-year intervals for screening.27 The data support-
`ing this interval come indirectly from sigmoidoscopy
`studies.19 Undisputed as the diagnostic gold standard,
`the ability to additionally perform intervention on pre-
`malignant lesions seems to make colonoscopy an ideal
`screening modality. Unfortunately,
`the
`available
`resources, cost, risk, considerable expertise required
`and inconvenience to the patient limit colonoscopy’s
`appeal as a mass screening tool for CRC. Evidence from
`several polyp prevention trials demonstrates higher
`yields for subsequent cancer than would be expected
`despite a relatively high use of surveillance procedures
`in follow-up.28–30 It is unclear whether these data
`represent a higher miss rate in the community than that
`seen in expert series or variable tumour biology with
`aggressive interval tumours.
`Radiographic modalities for screening CRC include
`double-contrast barium enema (DCBE) and virtual
`colonoscopy (VC). The DCBE has been recommended
`as a screening option, but is becoming less commonly
`utilized by practitioners. DCBE is a cost-effective option,
`but the low sensitivity for polyps and need for colon-
`oscopic confirmation limits its acceptability for screen-
`ing.31 The main role of DCBE is to visualize the right
`colon in cases where endoscopic visualization is not
`attained. VC has attracted some interest with its non-
`invasive nature and safety. The potential of VC to detect
`extracolonic disease is a double-edged sword as it adds
`cost and risk to subjects without any clear-cut evidence
`of benefit. Initial studies with VC showed mixed results
`for polyp detection and did not evaluate an average-risk
`asymptomatic population.32–35 Methods using prep-less
`cleansing techniques may improve acceptability, but
`sensitivity may be compromised.36 Recent results using
`a primary three-dimensional approach have shown
`improved polyp detection 6–10 mm in size prompting
`serious consideration for mass population screening.37
`The appropriate patient population and optimal means
`of training examiners requires further investigation.
`
`Ó 2004 Blackwell Publishing Ltd, Aliment Pharmacol Ther 19, 1225–1233
`
` 13652036, 2004, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2036.2004.02005.x, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
`
`Geneoscopy Exhibit 1039, Page 2
`
`

`

`R EV I E W A R TI CL E: ST OO L D N A TE ST IN G FO R COL O RE CTA L CA N CE R
`
`1227
`
`Multi-slice CT scanners and special software present
`highly resource-intensive implementation obstacles not
`available in excess supply at this time.
`
`S T OO L - B A SE D D N A T ES T I NG
`
`Rationale
`
`Clearly, there is a need for a more cancer-specific, non-
`invasive and targeted screening test to directly assess
`colonic epithelium changes. A successful
`test must
`reflect the entire length of
`the colorectum and be
`exempt from limiting factors such as patient-perceived
`discomfort, dietary or medicinal restrictions, and inter-
`mittent leaking of markers. If the test was acceptable
`enough to patients, and cost low, it could be adminis-
`tered more frequently, possibly allowing detection of
`interval aggressive tumours.
`Renewal of colonocytes results in exfoliation of at
`least 1010 cells every 24 h from the colonic surface.
`These cells resist degradation and can be isolated from
`dispersed human faeces.38 Stool
`samples obtained
`from CRC patients may contain more exfoliated cells
`or DNA than material from healthy people.39 Known
`specific DNA mutations previously characterized5, 40
`in exfoliated cells have been targeted to screen for
`precursor lesions and malignant disease. Advances in
`powerful molecular biology techniques,
`such as
`polymerase chain reaction (PCR), allow DNA to be
`amplified a billion-fold. Selection of characteristic
`markers will continue to evolve as the carcinogenetic
`pathways are better understood.
`Interestingly,
`the
`amount of faecal human DNA in cancer patients does
`not correlate with tumour size suggesting either the
`presence of a field effect or degradation-resistant
`DNA.41
`
`Current understandings of carcinogenesis
`
`Genomic instability is seen in virtually all CRCs. This
`disruption of genomic replication creates the necessary
`alterations for carcinogenesis. While the responsible
`mechanisms remain unknown, three different patho-
`genetic pathways are described. Although potentially
`overlapping, each pathway gives rise to mutations that
`can be strategically targeted for screening purposes.
`These processes are likely to be initiated after the loss of a
`‘gatekeeper’ gene, most likely the tumour suppressor APC
`gene. The APC gene inhibits b-catenin protein, an
`
`important regulator of cellular proliferation and adhe-
`sion.
`In 80–85% of colorectal neoplasms, point mutations
`at some loci produce a large number of
`loss of
`heterozygosity events. The chromosomal
`instability
`pathway was first conceptualized in relation to loss of
`both alleles in tumour suppressor genes, such as the
`APC gene. Sporadic mutation in one allele of the APC
`gene with inactivation of the second allele through
`genomic instability provides a growth advantage to a
`cell
`line allowing a polyp to undergo clonal expan-
`sion. Ninety per cent of the mutations occur in the
`the APC-coding region,42 providing
`first half of
`predictable specific targets for stool-based DNA testing.
`However,
`loss of APC function is not
`the sole
`determining factor
`in chromosomal
`instability in
`colorectal adenomas and the existence of multiple
`independent chromosomal
`instability pathways has
`been demonstrated.43
`The second pathway known as microsatellite instabil-
`ity (MSI) occurs in 12–15% of CRC patients. MSI was
`characterized in 1993 by loss of the DNA mismatch
`repair (MMR) system causing many point mutations
`and a large number of mutations at simple repeat
`sequences known as microsatellites. Several key growth
`regulators are altered in this process. MSI is known to
`cause mutations in large poly-A regions known as big
`adenine tracts (BAT) and dinucleotide repeat sequences
`known as CA-repeats. Targeting these sequences in
`DNA shed into stool has made it possible to detect
`MSI-containing cancers. Both of the above pathways
`lead to CRC and are illustrated in Figure 1.
`A small minority of CRC, 2–3%, does not develop
`through the above two pathways (it may be higher – as
`the multi-target assay panel
`(MTAP)
`is only 80%
`sensitive on tumour tissue).
`In 1999, one group
`demonstrated hypermethylation in promoter regions
`composed of clusters of cytosine-guanosine residues
`(CpG islands) could inactivate the genes.44 Methylation
`of tumour suppressor genes, especially APC, has been
`described in addition to a number of other genes
`including TIMP-3, HIC-1, PTEN, p16, p14, RARB and
`MGMT. Even as a distinct pathway, overlap exists when
`hypermethylation of the hMLH1 gene leads to MSI.
`Essentially, multiple pathways to CRC exist with each
`pathway containing its own mechanism of genomic
`instability. For example,
`the chronic inflammatory
`condition of ulcerative colitis may induce MSI through
`adaptive increases in the base excision repair enzymes.45
`
`Ó 2004 Blackwell Publishing Ltd, Aliment Pharmacol Ther 19, 1225–1233
`
` 13652036, 2004, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2036.2004.02005.x, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
`
`Geneoscopy Exhibit 1039, Page 3
`
`

`

`1228
`
`K . S. TA G O R E et al.
`
`Chromosomal
`instability
`(ras, p53
`mutations)
`
`LOH
`aneuploidy
`
`APC mutation or
`pathway inactivation
`
`p53 LOH
`
`Normal
`
`Adenoma
`
`Cancer
`
`Somatic or
`germline
`mutation (FAP)
`
`*Microsatellite instability
`(hMLH1 hypermethylation
`or
`MMR gene mutation)
`
`Hypermutable
`phenotype
`
`R-II, BAX, MMR genes
`MBD4, TCF-4, IGF2R, E2F4
`frameshifts
`
`*NOTE: Whether the two distinct pathways of Chromosomal Instability and Microsatellite Instability
`
`diverge before or after an initiating “gatekeeper” event is unclear. The diagram shows that this most
`
`commonly occurs with the mutational inactivation of both APC alleles, however APC is usually normal in
`
`the Microsatellite Instability pathway.
`
`Figure 1. The chromosomal instability and
`microsatellite instability pathways for colo-
`rectal cancer.
`
`Development
`
`Tumour-derived mutations in oncogenes and suppres-
`sor genes provide specific tumour markers. The discov-
`ery of these genetic alterations has raised the possibility
`of detecting colorectal neoplasms through examination
`of stool DNA from shed colonocytes. The challenge has
`been to identify these genetic mutational targets by
`isolating faecal colonocytes and purifying sufficient
`quantities of human DNA diluted amongst micro-
`organisms,
`food and mucus.
`Initial attempts were
`cumbersome and included cloning of PCR-amplified
`DNA products in a bacteriophage vector followed by
`hybridization to mutant-specific oligonucleotides.
`Early studies focused on k-ras, a mutated intracellular
`signalling protein found in up to 50% of malignant and
`benign tumours.5–12 Mutations in k-ras stabilize the
`protein in the guanosine triphosphate bound form,
`creating an oncogene that sends an unremitting stimulus
`for cellular proliferation. More than 80% of k-ras muta-
`tions are confined to codons 12 or 13 of k-ras46 allowing
`for technically easier gene PCR amplification producing
`large numbers of copies for examination. However, k-ras
`mutations are identified in controls,47 in addition to
`nondysplastic aberrant crypt foci and small hyperplastic
`polyps which often contain k-ras mutations.48–50
`Attention turned to other mutational markers inclu-
`ding p53 mutations which occur in approximately 50%
`
`of all human cancers.51 Activated with DNA damage,
`the tumour suppressor gene makes a p53 protein to
`control the cell cycle, repair DNA, halt synthesis of
`mutant DNA and cause apoptosis. Known as the
`‘guardian of the genome,’ the ultimate effect of p53 to
`prevent clonal expansion of mutant cells. Missense
`mutations of this tumour suppressor gene occur in
`predictable portions in exons 5–9, leading to its use as a
`stool DNA marker. While mutations in p53 are thought
`to be present only in the later stages of colorectal
`neoplasia, we have identified such changes in up to 64%
`of severely dysplastic polyps.52
`The ‘gatekeeper’ APC gene’s role in early cell trans-
`formation makes it most favourable for detecting early-
`stage colorectal neoplasia. Although large, this gene
`does have a good potential for isolating mutations as
`83% occur in the first part of the coding sequence.
`Missense mutations producing stop codons lead to a
`truncated APC protein. However, the mutations are not
`as easily detected in faecal DNA due to a low presence of
`the APC gene. Traverso et al. amplified the main region
`where the majority of mutations take place within a
`single PCR product.53 Magnetic capture beads were
`coated with oligonucleotides corresponding to the
`region between codons 1210 and 1581. This allowed
`for increased detection of
`the APC gene through
`identification of abnormal proteins produced from
`mutant genes, known as digital protein truncation.53
`
`Ó 2004 Blackwell Publishing Ltd, Aliment Pharmacol Ther 19, 1225–1233
`
` 13652036, 2004, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2036.2004.02005.x, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
`
`Geneoscopy Exhibit 1039, Page 4
`
`

`

`R EV I E W A R TI CL E: ST OO L D N A TE ST IN G FO R COL O RE CTA L CA N CE R
`
`1229
`
`the
`The results showed that stools from over half
`patients with CRC or polyps contained such genes, but
`normal subjects had no detectable mutant APC genes.
`Proximal and distal tumours may have inherent and
`acquired differences that provide the right colon with a
`predilection to undergo DNA mismatch repair.54 The
`inclusion of MSI in a panel of markers screening for CRC
`reflects the entire colorectum. Mutations in DNA-MMR
`produce changes in the coding of genes involved in the
`regulation of cell growth, including cell programmed
`death or apoptosis. Inactivation of DNA-MMR genes
`results in MSI in 15% of all CRCs. The phenotype can be
`detected as frequent alterations in certain microsatellite
`sequences such as BAT-26.55
`Colonocytes are sloughed in the intestinal lumen and
`undergo a process of physiological cell death termed
`apoptosis. Apoptotic cells shed from normal mucosa
`contain DNA cleaved by endonucleases
`in short
`fragment lengths of 180–200 bp. Observation of long
`PCR products, ‘long’ DNA (L-DNA) or high-molecular
`weight DNA, identifies the presence of non-apoptotic
`colonocytes which are characteristically exfoliated
`from neoplasms. Stools
`from patients with CRC
`contain subsets of both non-apoptotic L-DNA from
`dysplastic cells and ‘short’ DNA from normal mucosa.
`This longer template DNA is an epigenetic phenom-
`enon consistent with the known loss of apoptosis that
`occurs with CRCs.47 Differential and quantitative
`analysis of DNA fragments isolated from stools of
`patients with CRC have higher ‘integrity’ than DNA
`isolated from stools of patients with healthy colonic
`mucosa. Hence, the presence of L-DNA, as demon-
`strated by an assay of faecal DNA integrity, may be a
`sensitive and specific marker for detecting the presence
`of CRC.56
`
`Clinical studies
`
`Combinations of the above markers as a panel have
`been applied in clinical situations to assess sensitivity
`and specificity. Estimates of usefulness of these markers
`for CRC screening can be obtained by collectively
`examining the results. The initial blinded pilot study of
`33 tumours by Ahlquist et al. analysed point mutations
`at any of 15 sites on k-ras, p53, APC genes, BAT-26 and
`highly amplifiable DNA.47 The study demonstrated an
`impressively high sensitivity of 91% for CRC and 82%
`for
`large adenomas
`from archived stool
`samples.
`Adenomas ‡ 1 cm initially had an equally impressive
`specificity of 93%. By excluding k-ras from the panel,
`cancer sensitivity remained unaffected while adenoma
`sensitivity decreased to 73% and adenoma specificity
`increased to 100%. BAT-26 was positive in over half the
`malignant neoplasms above the splenic flexure and
`L-DNA was positive in 83% of cancers distal to the
`splenic flexure.
`Subsequent prospective studies have not been able to
`replicate such impressive results. Our study expanded
`the number of point mutations to a total of 21 sites and
`examined 80 in vivo neoplasms diagnosed at sigmoid-
`oscopy with an overall cancer sensitivity of 63.5%.57
`Advanced adenomas ‡ 1 cm were detected in 57.1% of
`cases and 85.7% of polyps with high-grade dysplasia.
`The study was limited to mostly symptomatic patients,
`the investigators were not blinded, and no direct
`comparisons were made with the other main modality
`of stool screening (i.e. FOBT). Comparisons with other
`prospective studies are strikingly similar as shown in
`Table 1.
`found similar results using the same
`Calistri et al.
`markers k-ras, APC, p53, BAT-26 and L-DNA.58 Despite
`
`Table 1. Performance of stool DNA mark-
`ers in the detection of cancer and advanced
`adenomas
`
`Authors
`
`Ahlquist et al.47
`Brand et al.67
`Syngal et al.66
`Tagore et al.57
`Calistri et al.58
`Imperiale et al.
`(pers. comm.)
`Overall reported
`experience
`
`Sensitivity-invasive
`cancers
`
`Specificity
`
`Sensitivity for advanced
`adenomas
`
`20/22 (91)
`11/17 (65)
`35/56 (62)
`33/52 (63)
`33/53 (62)
`16/31 (52)
`
`26/28 (93)
`N/A
`N/A
`204/212 (96)
`37/38 (97)
`1343/1423 (94)
`
`9/11 (82)
`N/A
`5/16 (31)
`16/28 (57)
`N/A
`61/403 (15)
`
`148/231 (64)
`
`1610/1701 (95)
`
`91/458 (20)
`
`Values in parentheses are expressed as percentage.
`
`Ó 2004 Blackwell Publishing Ltd, Aliment Pharmacol Ther 19, 1225–1233
`
` 13652036, 2004, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2036.2004.02005.x, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
`
`Geneoscopy Exhibit 1039, Page 5
`
`

`

`1230
`
`K . S. TA G O R E et al.
`
`labelling and detection methods,
`slightly different
`molecular alterations were found in 62% of the samples.
`This study attempted a panel that would be less time-
`consuming and labour-intensive by only analysing four
`fragments on the APC gene. As suggested by Ahlquist,
`L-DNA amplification detected more than 50% of
`tumours alone. K-ras was the next most
`frequent
`alteration detected in stool leading the researchers to
`suggest that these two markers are potentially capable
`of detecting two-thirds of cancers. More sensitive
`techniques would need to be developed and employed,
`but such an approach could make the screening test
`much less labour-intensive and more cost-effective.
`Preliminary data from a recent prospective trial
`presented at the 2003 American Society of Gastroen-
`terology (ASGE) took 4022 average risk patients to stool
`DNA testing and FOBT prior to screening colonoscopy.
`Thirty-one advanced neoplasms were identified. DNA
`mutations were identified in 52% of tumour-bearing
`patients but only 13% were FOBT-positive (personal
`communication). Results from a similar-sized NIH-
`funded trial are pending.
`Among the specific markers, L-DNA amplification has
`emerged as the most
`frequent event. Longer DNA
`represents exfoliated non-apoptotic neoplastic cells. This
`DNA has escaped the cleaving effects of endonucleases
`which normally create short DNA. Stool BAT-26 in
`most studies appears to be a marker of MSI which
`identifies the presence of half the advanced right-sided
`neoplasms. Performance characteristics of
`individual
`stool DNA markers are listed in Table 2.
`
`Future of DNA testing
`
`Population-based screening of average risk individuals
`will require a highly sensitive test capable of detecting
`early disease. Stool DNA mutations are inherently
`reliable markers of molecular carcinogenesis, which
`may provide to both patients and primary care doctors a
`
`clear reason why screening and follow-up for CRC is
`necessary. While preliminary studies are promising, the
`problems of sensitivity and cost need to be addressed. As
`more markers are added to improve sensitivity, the cost
`of the assay increases. The current iteration of the
`EXACT assay costs between $500 and 800. Compared
`with immunochemical or guaiac-based FOBTs which
`cost between $5 and 30, the sensitivity benefit has yet
`to demonstrate cost-effectiveness.
`Refinements in detection techniques, targeting mul-
`tiple markers, and developing novel
`strategies of
`targeting new mutations may improve sensitivity and
`specificity. EffipureTM is a novel technique that allows
`five-fold greater amplification of stool human DNA
`using a highly efficient capture technique.59 The new
`sample prep method gives an average 5.4-fold increase
`in the quantity of human DNA that can routinely be
`retrieved from faecal samples compared with the bead
`capture method. This resulted in a 16% increase (14 of
`89) in detection of confirmed cancers, with no loss of
`specificity (personal communication). Preliminary stud-
`ies show far greater tumour sensitivities on previously
`negative tumour samples.
`It
`is
`likely that
`future
`technological improvements will further enhance the
`sensitivity of stool DNA testing.
`A plasma or serum test to detect sporadic colorectal
`neoplasia would involve DNA markers
`including
`k-ras,60–62 APC63 and hypermethylated p16,64 leaked
`into the bloodstream during tissue invasion. The overall
`detection rates in these studies have ranged from 15 to
`70%, but show higher detection rates with more
`advanced stage neoplasia,
`limiting its value as a
`screening test. One recent report describes gynaecologi-
`cal and breast cancer detection rates of 62% overall and
`50% for stage I cancers using an assay of L-DNA in
`serum.65 If tumour marker exfoliation into the colon
`likely precedes tissue invasion, stool assay of DNA
`markers for adenomas and early-stage cancers will prove
`more sensitive than the plasma assay. No comparison
`
`Table 2. Sensitivity performance of individual markers in the stool DNA assay panel for invasive cancers only
`
`Authors
`
`Ahlquist et al.47
`Dong et al.68
`Tagore et al.57
`Calistri et al.58
`Range (%)
`
`APC
`
`5/22 (23)
`N/A
`7/52 (14)
`2/53 (4)
`4–23
`
`K-ras
`
`4/22 (18)
`8/51 (16)
`9/52 (17)
`6/53 (11)
`11–18
`
`P53
`
`3/22 (14)
`30/51 (59)
`17/52 (33)
`3/53 (6)
`6–59
`
`BAT-26
`
`5/22 (23)
`3/51 (6)
`2/52 (4)
`3/53 (6)
`4–23
`
`L-DNA
`
`14/22 (61)
`N/A
`19/52 (37)
`27/53 (51)
`37–61
`
`Values in parentheses are expressed as percentage.
`
`Ó 2004 Blackwell Publishing Ltd, Aliment Pharmacol Ther 19, 1225–1233
`
` 13652036, 2004, 12, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/j.1365-2036.2004.02005.x, Wiley Online Library on [25/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
`
`Geneoscopy Exhibit 1039, Page 6
`
`

`

`R EV I E W A R TI CL E: ST OO L D N A TE ST IN G FO R COL O RE CTA L CA N CE R
`
`1231
`
`studies have been performed of stool-based DNA tests
`with plasma-based DNA tests for detection of cancer.
`The cost of the procedure is high and far exceeds that of
`FOBT. Rationalization of the test to one or two markers
`could significantly lower the cost. It appears likely that
`markers germane to all tumours will be more suitable
`than a large panel. Candidates for this streamlined
`approach include L-DNA and hypermethylation markers
`because of their sensitivity. Application as an interval
`test between endoscopies may be an attractive option for
`patients who prefer to avoid or minimize repeat colon-
`oscopy. Up to 93% of
`follow-up post-polypectomy
`colonoscopies fail to find advanced adenomas and in
`the elderly the risks of the procedure outweigh the low
`benefit. To rationalize resources and minimize risk, this
`group could be better served by DNA screening.
`Another potential area of utilization for stool DNA
`testing is for tumour follow-up and risk assessment.
`Syngal et al. did post-resection stool DNA testing for
`patients with colonic neoplasms.66 Most mutational
`markers resolved post-resection with the exception of
`L-DNA suggesting the presence of a pervasive field
`effect. These findings suggest that stool DNA mutational
`analysis could have potential use in monitoring CRC
`patients after treatment.
`Stool DNA testing offers a sensitive and non-invasive
`alternative for colon cancer screening likely to appeal to
`patients who have avoided screening because of fear or
`inconvenience. If the sensitivity and cost limitations can
`be overcome, DNA markers offer the potential for a less
`risky screening procedure which could target popula-
`tions who are currently underserved by screening.
`
`A C K N OW L E DG E M E N T S
`
`The authors with to thank Mark Henderson MD, Mike
`Ross MD, Barry Berger MD, Kathy Morel RN for their
`critical review of this manuscript. The authors are not
`employed by Exact Laboratories and do not have any
`personal financial investment in the company.
`
`R E F E R E N C E S
`
`1 American Cancer Society. Cancer Facts and Figures 2004.
`Atlanta: American Cancer Society, 2004.
`2 Mandel JS, Bond JH, Church TR, et al. Reducing mortality
`from colorectal cancer by screening fecal occult blood. Min-
`nesota Colon Cancer Control Study. N Engl J Med 1993; 328:
`1365–71.
`
`for colorectal
`
`3 Mandel JS, Church TR, Bond JH, et al. The effect of fecal
`occult-blood screening on the incidence of colorectal cancer.
`N Engl J Med 2000; 343: 1603–7.
`4 Ries L, Eisner M, Kosary C, et al. SEER Cancer Statistics
`Review, 1973–1997. Bethesda, MD: National Cancer Insti-
`tute, 2000.
`5 Fearon ER, Vogelstein B. A genetic model
`tumorigenesis. Cell 1990; 61: 759–67.
`6 Sidransky D, Tokino T, Hamilton SR, et al. Identification of ras
`oncogene mutations in the stool of patients with curable
`colorectal tumors. Science 1992; 256: 102–5.
`7 Hasegawa Y, Takeda S, Ichii S, et al. Detection of k-ras
`mutations in DNAs isolated from feces of patients with colo-
`rectal tumors by mutant-allele-specific amplification (MASA).
`Oncogene 1995; 10: 1441–5.
`8 Smith-Raven J, England J, Talbot IC, et al. Detection of c-Ki-ras
`mutations in faecal samples from sporadic colorectal cancer
`patients. Gut 1995; 36: 81–6.
`9 Villa E, Dugani A, Rebecchi AM, et al. Identification of subjects
`at risk for colorectal carcinoma through a test based on K-ras
`determination in the stool. Gastroenterology 1996; 110:
`1346–53.
`10 Nollau P, Moser C, Weinland G, et al. Detection of K-ras
`mutations in stools of patients with colorectal cancers by
`mutant-enriched PCR. Int J Cancer 1996; 66: 332–6.
`11 Eguchi S, Kohara N, Komuta K, et al. Mutations of the p53
`gene in stool of patients with respectable colorectal cancer.
`Cancer 1996; 77: 1707–10.
`12 Puig P, Urgell E, Capella G, et al. Detection of k-ras mutations
`in the stool of patients with pancreatic adenocarcinoma and
`pancreatic ductal hyperplasia. Cancer Res 1994; 54:
`3568–73.
`13 Thompson MR, Heath I, Ellis BG, et al. Identifying and man-
`aging patients at low risk of bowel cancer in general practice.
`BMJ 2003; 327: 263–5.
`14 Hardcastle JD, Chamberlain JO, Robinson MH, et al. Rand-
`omised controlled trial of
`faecal-occult-blood screening for
`colorectal cancer. Lancet 1996; 348: 1472–7.
`15 Kronborg O, Fenger C, Olsen J, et al. Randomised study of
`screening for colorectal cancer with faecal-occult-blood test.
`Lancet 1996; 348: 1467–71.
`16 Lieberman DA, Weiss DG, Bong JH, et al. Use of colonoscopy to
`screen asymptomatic adults
`for colorectal cancer. VA
`Cooperative Study Group 380. N Engl J Med 2000; 343:
`162–8.
`17 Allison JE, Tekawa IS, Ransom LJ, Adrain AL. A comparison
`of
`fecal occult-blood tests for colorectal-cancer screening.
`N Engl J Med 1996; 334: 155–9.
`18 Cole SR, Y