GASTROENTEROLOGY 1996;110:1346–1353
`
`Identification of Subjects at Risk for Colorectal Carcinoma
`Through a Test Based on K-ras Determination in the Stool
`
`ERICA VILLA,* AISHA DUGANI,* ANNA MARIA REBECCHI,* ANNALISA VIGNOLI,*
`ANTONELLA GROTTOLA,* PAOLA BUTTAFOCO,* LORENA LOSI,‡ MARIO PERINI,* PAOLO TRANDE,*
`ANNALISA MERIGHI,* RITA LEROSE,* and FEDERICO MANENTI*
`*Division of Gastroenterology, Department of Internal Medicine, and ‡Department of Pathology, University of Modena, Modena, Italy
`
`Background & Aims: The gold standard for screening
`for colorectal carcinoma is colonoscopy. The aim of this
`study was to compare endoscopic results with those
`obtained using the noninvasive screening test of K-
`ras determination in the stool in a large population of
`patients undergoing colonoscopy. Methods: Two hun-
`dred thirty consecutive patients were studied by K-ras
`amplification on stool-derived DNA using polymerase
`chain reaction and oligomer-specific hybridization. Re-
`sults: Wild-type K-ras was amplified in 103 of 230 pa-
`tients (44.8%), the rate of amplification being directly
`proportional to the presence of an organic disease of
`the intestine characterized by hyperproliferating mu-
`cosa. In 30 of these 103 patients (29.1%), a K-ras
`mutation was found. Four of 5 with early colorectal
`carcinoma, all who had K-ras mutations in the tumor,
`were identified. In first-degree relatives of patients with
`colorectal carcinoma, all subjects either carrying ade-
`nomas ú1 cm in diameter or multiple smaller adeno-
`mas were identified.
`In patients with inflammatory
`bowel disease, the test identified the only patient with
`neoplastic transformation. Conclusions: The sensitivity
`and specificity of K-ras determination on stool-derived
`DNA in patients with colorectal carcinoma, in first-de-
`gree relatives of patients with colorectal carcinoma,
`and in patients with inflammatory bowel disease sup-
`port the opportunity of a large-scale trial to validate its
`use as a screening test.
`
`Sporadic colorectal cancer (CRC) occurs in Western
`
`countries at a frequency second only to lung cancer.1
`If diagnosed at an early stage, CRC can be cured either
`by endoscopic excision or by surgery. The use of occult
`blood tests as a screening test for early identification
`of neoplastic lesions has shown controversial results2 – 8;
`although all trials showed an increased discovery of early
`colorectal cancers, only one showed a reduction in mortal-
`ity.3 Therefore, there is still uncertainty about the sensi-
`tivity and specificity of this test, with some authors sug-
`gesting that a significant proportion of cancers detected
`may be chance discoveries.7,9 Colonoscopy remains the
`gold standard for identification of both preneoplastic and
`
`neoplastic lesions; however, it cannot be used for mass
`screening because of its cost and invasiveness. Studies
`using colonoscopy for screening are in progress in selected
`populations such as patients older than 60 years of age.10
`Apart from genetically determined colonic tumors,
`several changes in proto-oncogenes and tumor-suppressor
`genes have been shown in sporadic colorectal carcinogen-
`esis11 and, recently, also in colitis-associated tumors.12
`Although the model proposed for tumorigenesis suggests
`that it is the accumulation of multiple alterations rather
`than the specific order of them that is important for the
`development of tumor, K-ras mutations occur earlier
`than p53 mutations and are observed with a similar prev-
`alence in carcinomas and in intermediate-stage adeno-
`mas.11 Furthermore, most of the reported mutations clus-
`ter in a rather short fragment of DNA (i.e., the first
`exon of K-ras gene), making the exploration of possible
`mutations in genomic material derived from intestinal
`mucosa easier, whereas mutations of APC or p53 gene
`are dispersed in a rather longer fragment of DNA.11
`Sidransky et al.13 were first to report the possibility
`of using DNA extracted from the stool for identifying
`mutations of DNA derived from intestinal mucosa; they
`found mutations in 8 patients with colorectal tumor. We
`tested this method as a noninvasive screening test for
`colorectal tumors or colorectal adenomas in a large popu-
`lation of nonselected patients undergoing diagnostic co-
`lonoscopy.
`Materials and Methods
`Patients
`
`Stool specimens were collected from 230 nonselected
`patients undergoing diagnostic colonoscopy from September
`1993 to June 1994. The indications for colonoscopy were rectal
`bleeding in 6 patients (1 first-degree relative of a patient with
`
`Abbreviations used in this paper: CRC, colorectal cancer; PCR,
`polymerase chain reaction; SDS, sodium dodecyl sulfate; SSC, stan-
`dard saline citrate.
`䉷 1996 by the American Gastroenterological Association
`0016-5085/96/$3.00
`
`/ 5e0d$$0040
`
`04-11-96 15:07:21
`
`gasa
`
`WBS-Gastro
`
`Geneoscopy Exhibit 1054, Page 1
`
`

`

`May 1996
`
`K-ras DETERMINATION IN THE STOOL 1347
`
`Table 1. Demographic Characteristics of Patients in the Different Categories
`
`CRC
`
`5
`62 { 16
`3/2
`
`0/5
`1/5
`
`0/5
`
`History of
`CRC
`
`31
`67 { 8
`21/10
`
`0/31
`4/31
`
`0/31
`
`Adenomas
`(including hyperplastic
`polyps)
`
`55
`59 { 11
`37/18
`
`1/42a
`20/42
`
`History of
`adenomas
`
`67
`62 { 10
`50/17
`
`1/67b
`10/67
`
`2/42a,d
`
`4/67a,b,d
`
`IBD
`
`20
`41 { 13
`11/9
`
`0/20
`0/20
`
`0/20
`
`Lymphocytic
`colitis
`
`No organic
`disease
`
`6
`67 { 6
`2/4
`
`0/6
`0/6
`
`0/6
`
`46
`50 { 14
`14/32
`
`1/46c
`19/46
`
`2/46b,e
`
`No. of patients
`Age (yr, mean { SD)
`M/F
`Personal history of tumors
`other than CRC
`Family history of CRC
`Family history of tumors
`other than CRC
`
`aProstate.
`bLung.
`cUterus.
`dStomach
`ePancreas.
`
`CRC); surveillance endoscopy after surgery for CRC in 31
`patients (4 first-degree relatives of a patient with CRC); sur-
`veillance endoscopy for previously excised adenomas in 67
`patients (10 first-degree relatives of patients with CRC); sur-
`veillance endoscopy for inflammatory bowel disease (IBD) fol-
`low-up in 19 patients; symptoms of irritable bowel syndrome,
`diarrhea, or constipation in 68 patients; and screening for
`family history of CRC in 39 patients.
`The goal of the study was explained to the patients, and
`they were asked to bring a sample of stool 1 week after colonos-
`copy before excision of the tumor or polyp. Samples were
`collected and coded by personnel unaware of the results of
`endoscopy. Results of the molecular analysis (performed under
`code) and endoscopy were matched only at the end of the
`investigation. Stool specimens were frozen at 080⬚C until use.
`All patients agreed to participate in the study; in a subgroup
`of patients with CRC, bioptic specimens were obtained for
`comparison with the results of the stool test. All of these
`patients gave written informed consent for use of the bioptic
`specimens not for diagnostic purposes but for use in the present
`study. The study was approved by the Institutional Human
`Research Committee of the University of Modena.
`Each patient completed a questionnaire with the following
`information: age, sex, results of previous colonoscopies (when
`present), histology of adenomas (when present), previous sur-
`gery in the gastrointestinal tract for tumors, family history of
`intestinal tumors, and family history of tumors elsewhere than
`the gastrointestinal tract. These data are summarized in Table
`1.
`
`Detection of K-ras Mutations
`
`DNA was extracted from stool specimens according
`to the method described by Sidransky et al.13 with minor
`modifications. Different amounts of stool specimens (100 mg,
`500 mg, 1 g, and 5 g) were extracted in a pilot experiment
`(data not shown) to evaluate the optimal quantity of stool to
`use. The best results, in terms of number of positive amplifica-
`tions of both K-ras and b-globin, were obtained with 100 mg.
`
`Therefore, this amount was routinely used. Samples of stool
`were diluted in 300 mL lysis buffer (500 mmol/L Tris-HCl,
`16 mmol/L ethylenediaminetetraacetic acid, and 10 mmol/
`L NaCl (pH 9.0), and particulate material was removed by
`centrifugation at 12,000g for 2 minutes. Proteins were digested
`with sodium dodecyl sulfate (SDS)–proteinase K (20 mg/mL)
`overnight at 37⬚C or for 1 hour at 65⬚C, extracted twice with
`phenol-chloroform, and ethanol precipitated. The extracted
`DNA (0.5–9 mg) was incubated in a total volume of 100 mL
`of a solution containing 10 mmol/L Tris-HCl (pH 7.2); 50
`mmol/L KCl; 2 mmol/L MgCl2; 500 mmol/L each of deoxy-
`adenosine triphosphate, deoxyguanine triphosphate, deoxycyti-
`dine triphosphate, and deoxyribosylthymine triphosphate; 100
`ng of primers located in the first ras exon (sense primer: 5ⴕ
`AGG AAT TCA TGA CTG AAT ATA AAC TTG 3ⴕ; anti-
`sense primer: 5ⴕ ATC GAA TTC CTC TAT TGT TGG ATC
`ATA TTC 3ⴕ, giving rise to an amplified fragment of 202
`base pairs), and 1 U of Taq DNA polymerase (Boehringer
`Mannheim, Milan, Italy). Forty-five cycles (30 minutes at
`95⬚C, 3 minutes at 58⬚C, and 2 minutes at 70⬚C) were per-
`formed. Amplification products were electrophoresed through
`1.8% agarose gel and ethidium bromide stained. Electropho-
`resed polymerase chain reaction (PCR) products were trans-
`ferred to nylon membranes (Hybond N; Amersham, Milan,
`Italy) and prehybridized in a mixture containing 50% for-
`mamide, 51 standard saline citrate (SSC) (0.75 mol/L NaCl
`and 0.075 mol/L sodium citrate), 51 Denhardt’s solution,
`0.1% Ficoll (type 400; Pharmacia, Milan, Italy), 0.1% polyvi-
`nylpyrrolidone, 0.1% bovine serum albumin (fraction V;
`Sigma, Milan, Italy), and 50 mmol/L NaHPO4 (pH 7.0) at
`42⬚C for at least 2 hours. Hybridization was performed under
`stringent conditions in a mixture containing 50% formamide,
`10% dextran sulfate, 51 SSC, 11 Denhardt’s, and 20 mmol/
`L NaPO4 (pH 7.0) in the presence of 100 mg/mL of denatured
`salmon sperm DNA at 42⬚C overnight. As probes, specific
`oligonucleotides located between the amplification primers and
`specific for each mutation, terminally labeled with [g-32P]
`deoxyadenosine triphosphate using T4 polynucleotide kinase,
`
`/ 5e0d$$0040
`
`04-11-96 15:07:21
`
`gasa
`
`WBS-Gastro
`
`Geneoscopy Exhibit 1054, Page 2
`
`

`

`1348 VILLA ET AL.
`
`GASTROENTEROLOGY Vol. 110, No. 5
`
`were used. Stringent washing conditions were used; after rins-
`ing the nylon membranes in 31 SSC plus 0.1% SDS at room
`temperature for 5 minutes, they were washed in 31 SSC plus
`0.1% SDS for 30 minutes at approximately 10⬚C below the
`estimated melting temperature for each probe. Autoradiogra-
`phy was performed for 6, 12, and 36 hours at 080⬚C with an
`intensifying screen; samples were considered positive when a
`clear band was present after 36 hours of exposure. No cross-
`hybridization was observed between the mutated probes.
`As internal control for DNA origin and quality, in an initial
`group of 30 patients, b-globin was amplified using the follow-
`ing primers: sense, 5ⴕ GGT TGG CCA ATC TAC TCC AGG
`3ⴕ; antisense, 5ⴕ TGG TCT CCT TAA ACC TGT CTT G 3ⴕ,
`which give rise to an amplified fragment of 270 base pairs.
`To reduce the risk of contamination and false-positive re-
`sults, a number of precautions were taken as already de-
`scribed.14 As positive control, DNA extracted from surgical
`specimens of normal and neoplastic colon mucosa, known to
`be positive for K-ras mutation, were used in each run of experi-
`ment.
`In 5 patients with CRC at colonoscopy, bioptic specimens
`of the tumor were obtained and frozen at 080⬚C until pro-
`cessed for DNA extraction. In 10 patients who had undergone
`surgery for CRC, tumor tissue embedded in paraffin was avail-
`able for amplification; DNA was extracted as described pre-
`viously14 and amplified as described in Materials and Methods.
`
`Statistical Analysis
`Data were analyzed using the x2 test with Yates’ correc-
`tion when appropriate or using two-tailed Fisher’s Exact Test.
`A P value of õ0.05 was considered significant.
`Results
`DNA obtained from 100 mg of stool ranged from
`0.5 to 9 mg. Differences were present among the various
`groups of patients studied concerning both the purity
`and the total quantity recovered (Table 2).
`K-ras fragments were amplified from DNA extracted
`from stool in 103 of 230 patients (44.7%). A K-ras
`mutation was identified in 30 of these 103 patients
`(29.1%); the mutation identified was Asp12 in 21 of 30
`
`Table 2. Characteristics of DNA Extracted From Stool
`
`A260/A280
`
`CRC
`Patients undergoing surgery for CRC
`Without recurrence of polyps
`With recurrence of polyps
`Adenomas
`Active IBD
`Quiescent IBD
`No organic disease
`
`2.0
`
`1.9
`1.8
`1.9
`2.0
`1.9
`1.8
`
`Total amount of
`DNA obtained
`from 100 mg
`stool (mg)
`7.3 { 1.1
`3.4 { 2.2
`5.9 { 2.1
`6.4 { 2.8
`7.5 { 1.6
`2.9 { 0.8
`1.0 { 0.5
`
`Figure 1. (A) Ethidium bromide staining and (B) Southern blot analy-
`sis of K-ras fragments on DNA from the stool. DNA was amplified by
`PCR as described in Materials and Methods and hybridized with Asp12-
`specific probe. Lanes 1 and 2, patients with colorectal carcinoma;
`lane 3, patient with ulcerative colitis; lanes 4 and 5, patients with
`adenomas, belonging to the subgroup of CRC relatives. M, molecular
`weight marker; pBR 322 DNA, Hae III digested.
`
`patients (70.0%; Figure 1), Val12 in 5 (16.6%), and Asp13
`in 4 (13.3%). Table 3 shows the percentages of amplified
`K-ras fragments and identified mutations in the patients
`divided according to endoscopic diagnosis.
`In a subgroup of 30 patients (4 with CRC, 8 with
`adenomas [5 with a diameter of ú1 cm, 3 with õ1 cm],
`4 with IBD, and 14 without endoscopic abnormalities),
`DNA was also amplified with b-globin primers. b-Glo-
`bin was successfully amplified in all patients with CRC,
`in the 5 patients with adenomas ú1 cm in diameter,
`and in the 4 patients with IBD. No amplification was
`obtained in patients with adenomas õ1 cm in diameter,
`whereas a positive amplification was found in 4 of 14
`patients (28.5%) without endoscopic abnormalities.
`In 4 of 5 patients in whom a carcinoma (4 in T1 stage
`and 1 in T2 stage according to Hutter et al.15) was found,
`K-ras fragments were amplifiable; Asp12 mutation was
`found in 3 of these patients, and Asp13 mutation was
`found in 1 patient.
`In patients who had undergone surgery for CRC, the
`rate of amplification of K-ras fragments was significantly
`higher in those who had a recurrence of adenomas during
`follow-up than in patients without adenomas found dur-
`ing follow-up (P õ 0.01) (Table 3). A higher, although
`not significantly different, number of K-ras mutations
`was found in patients with recurrent adenomas compared
`with patients without recurrences (3 of 7 vs. 3 of 24).
`Among 42 subjects with adenomas at colonoscopy, K-
`ras was amplifiable in 30 (71.4%). In 19 of these patients,
`the polyp was single and the diameter was ¢1; in the
`other 11 patients, several small adenomas (4–7 mm in
`diameter) were present. A mutation was present in 12
`patients (28.5%): 11 with adenomas ¢1 cm and 1 with
`multiple small adenomas (4–7 mm). In the 12 subjects
`with adenomas in whom K-ras fragments were not am-
`plified, polyps were single and õ8 mm in diameter.
`In 67 subjects with an adenoma excised in the past,
`
`/ 5e0d$$0040
`
`04-11-96 15:07:21
`
`gasa
`
`WBS-Gastro
`
`Geneoscopy Exhibit 1054, Page 3
`
`

`

`May 1996
`
`K-ras DETERMINATION IN THE STOOL 1349
`
`Table 3. Results of K-ras Amplification in All Patients Studied
`
`Patients with colon cancer
`Patients undergoing surgery for colon cancer
`With adenomas during follow-up
`Without adenomas during follow-up
`Patients with adenomas
`Patients with history of adenomas
`Patients with hyperplastic polyps
`IBD
`Lymphocytic colitis
`No organic diseases
`Total
`aPõ 0.01.
`
`n
`
`5
`31
`7
`24
`42
`67
`13
`20
`6
`46
`230
`
`K-ras fragments (%)
`
`Mutated K-ras (%)
`
`5 (100)
`19 (61.2)
`7 (100)a
`12 (50.0)a
`30 (71.4)
`11 (16.4)
`5 (38.4)
`13 (65.0)
`6 (100)
`15 (32.6)
`104 (45.2)
`
`4 (80.0)
`6 (19.3)
`3 (42.8)
`3 (12.5)
`12 (28.5)
`2 (2.9)
`1 (7.7)
`3 (15.0)
`0
`2 (4.3)
`30 (13.0)
`
`no adenomas were found; K-ras fragments were ampli-
`fiable in only 11 and a mutation was present in only 2
`(2.9%). In 13 other patients with an adenoma excised in
`the past, only hyperplastic polyps were found; K-ras was
`amplified in 5 (38.4%), but a mutation (Asp13) was de-
`tected in only 1.
`
`First-Degree Relatives of Patients With
`Colon Cancer
`
`Fifty-four subjects were first-degree relatives of
`patients with colon cancer. The distribution of results in
`this subgroup of patients is shown in Table 4. The differ-
`ence in the rate of amplification of K-ras fragments is
`particularly evident, although not significant, in patients
`with previous surgery for colon cancer and in patients
`without organic disease at colonoscopy. The difference
`becomes significant both in the rate of amplification of
`K-ras (P õ 0.05) and in the frequency of K-ras mutations
`(P õ 0.01) when the whole group of relatives of patients
`with colon cancer is compared with that of nonrelatives
`(Table 4).
`In patients with adenomas, the rate of amplification
`
`of K-ras fragments was not significantly different be-
`tween the two subgroups (relatives and nonrelatives),
`whereas the difference in the rate of demonstration of K-
`ras mutations was statistically significant (P õ 0.01)
`(Table 4). The diameter of the adenomas found at colon-
`oscopy in CRC kindred was ¢1 cm in 12 of 20 patients;
`10 of these 12 patients had K-ras mutations (83.3%). In
`nonrelatives, only 7 of 22 subjects had polyps ú1 cm
`and only 1 subject had mutations (14.0%) (P õ 0.05,
`relatives vs. nonrelatives).
`
`IBD
`
`In patients with IBD, independently from having
`ulcerative colitis or Crohn’s disease, there was a marked
`difference between patients with active and widespread
`disease and patients with inactive disease; K-ras frag-
`ments were amplified in a very high percentage of pa-
`tients with active disease and in none with inactive dis-
`ease (Table 5). Mutations were found in only 3 patients,
`all with pancolitis. One patient had Crohn’s disease diag-
`nosed 14 years earlier (Asp13). The other 2 patients (both
`with evidence of Asp12 mutation) had ulcerative colitis;
`
`Table 4. Results of K-ras Amplification in DNA From Stool From Different Categories of Patients Divided According to
`Presence of History of Colon Cancer in a First-Degree Relative
`
`CRC kindred
`
`Nonrelatives
`
`n
`
`1
`
`4
`20
`10
`19
`54
`
`K-ras fragments
`(%)
`
`Mutated K-ras
`(%)
`
`1 (100)
`
`1 (100)
`
`4 (100)
`13 (65.0)
`2 (20.0)
`9 (47.3)
`29 (53.7)b
`
`2 (50.0)
`10 (50.0)a
`1 (10.0)
`1 (5.2)
`15 (27.7)c
`
`n
`
`4
`
`27
`22
`57
`27
`137
`
`K-ras fragments
`(%)
`
`Mutated K-ras
`(%)
`
`3 (75.0)
`
`15 (55.5)
`17 (77.2)
`9 (15.8)
`6 (22.2)
`50 (36.9)b
`
`3 (75.0)
`
`4 (14.8)
`2 (9.0)a
`1 (1.7)
`1 (3.7)
`11 (8.0)c
`
`Patients with colon cancer
`Patients undergoing surgery for colon
`cancer
`Patients with adenomas
`Patients with history of adenomas
`No organic diseases
`Total
`a,cPõ 0.01.
`bPõ 0.05.
`
`/ 5e0d$$0040
`
`04-11-96 15:07:21
`
`gasa
`
`WBS-Gastro
`
`Geneoscopy Exhibit 1054, Page 4
`
`

`

`1350 VILLA ET AL.
`
`GASTROENTEROLOGY Vol. 110, No. 5
`
`Table 5. Percentage of K-ras Fragments and Mutated K-ras
`in Patients With IBD and Lymphocytic Colitis
`
`Ulcerative colitis
`Active disease
`Inactive disease
`Crohn’s disease
`Lymphocytic colitis
`
`n
`
`15
`10
`5
`5
`6
`
`K-ras fragments
`(%)
`
`Mutated K-ras
`(%)
`
`9 (90.0)
`0
`4 (80.0)
`6 (100)
`
`2 (20.0)
`0
`1 (20.0)
`0
`
`NOTE. Among the patients with ulcerative colitis, all but 1 with active
`disease had pancolitis.
`
`1 patient, a 46-year-old man with ulcerative colitis diag-
`nosed 13 years earlier, underwent colectomy for deep
`impairment of clinical conditions. He did not have endo-
`scopic evidence of neoplasm, but microscopic examina-
`tion of the excised colon showed three small, plaque-like
`carcinomas about 2 cm in diameter in the right colon.
`The other patient, a 55-year-old man, was diagnosed at
`first colonoscopy performed for rectal bleeding. Histolog-
`ical analysis showed polymorphs infiltrating the epithe-
`lium, mucin depletion, and a slight increase in the num-
`ber of nuclei, which were also hyperchromatic. These
`changes were considered indefinite for dysplasia, and co-
`lectomy was delayed.
`Six patients (2 men and 4 women; mean age, 67 { 6
`years) had a diagnosis of lymphocytic colitis16,17; these
`patients had a normal endoscopic appearance apart from
`slight hyperemic changes. Histologically, mild inflam-
`matory infiltrates of the surface epithelium and of the
`lamina propria were present. K-ras fragments were am-
`plifiable in all of these patients, but none had evidence
`of K-ras mutation (Table 5).
`On the whole, K-ras fragments were amplified in 54
`of 56 patients (96.4%) with conditions associated with
`hyperproliferating mucosa and increased cellular shed-
`ding into the intestinal lumen (5 CRCs, 19 adeno-
`mas ú1 cm, 11 multiple adenomas, 15 active IBD, and
`6 lymphocytic colitis). Instead, this gene has been ampli-
`fied in only 15 of 63 patients (23.8%) either with no
`organic disease (46 subjects) or with substantially benign
`conditions (12 small adenomas õ8 mm in diameter or
`5 inactive IBD). The difference is highly significant
`(P õ 0.0001). K-ras mutations were also differently dis-
`tributed in the two subgroups (19 of 56 vs. 2 of 62; P õ
`0.001).
`
`Presence of K-ras Mutation in DNA From
`Stools and in Cancer Tissue
`
`Colonic tissue derived from the tumoral lesion
`was available for investigation of K-ras mutation in all
`5 patients with CRC at colonoscopy and in 10 of those
`
`who had undergone surgery for CRC, 6 of whom had a
`K-ras mutation. In patients with actual CRC at colonos-
`copy, results of amplification on DNA from stool and
`from colonic tissue were the same, whereas in patients
`with previous surgery for CRC, PCR on DNA from stool
`failed to identify a K-ras mutation in 1 of 6 (16.6%)
`(Table 6). The only patient characterized by Asp13 muta-
`tion had a recurrence during the period of the study,
`whereas all others remained tumor-free.
`Discussion
`K-ras mutations have been found to occur during
`colorectal carcinogenesis in a stage that roughly corre-
`sponds to transition from early to intermediate adenoma;
`they could be involved in the conversion of small adeno-
`mas to more dysplastic ones with progression to carci-
`noma.11,18 They are mostly localized on codon 12 and,
`at a lesser degree, on codon 13. This clustering in a short
`fragment of DNA, about 150 base pairs in size, is very
`convenient for reliable and easy PCR amplification, even
`on DNA derived from stool (which could be more de-
`graded than that derived directly from the intestinal
`mucosa).13
`
`Table 6. Comparison Between K-ras Findings in the DNA
`Extracted From Stool and From Tumor Tissue of
`Patients With CRC and Patients With Previous
`Surgery for CRC
`
`Patients with CRC
`1
`2
`3
`4
`5
`Patients with previous surgery
`for CRC
`1
`2
`3
`4
`5
`6
`7a
`8a
`9a
`10
`
`K-ras
`
`Tumor tissue
`
`Stool
`
`Wild type
`Asp12
`Asp12
`Asp12
`Asp13
`
`Wild type
`Wild type
`Wild type
`Wild type
`Asp12
`Asp12
`Asp12
`Asp12
`Asp13
`Val12
`
`Wild type
`Asp12
`Asp12
`Asp12
`Asp13
`
`Wild type
`Wild type
`Wild type
`Wild type
`Asp12
`Asp12
`Asp12
`Asp12
`Asp13
`Negative
`
`NOTE. Patients with CRC were studied by PCR on DNA derived from
`fresh-frozen tumor tissue and patients with previous surgery for CRC
`by DNA extracted from paraffin-embedded tumor tissue.
`aPatients 7–9 had a recurrence of adenomas during follow-up. Patient
`9 had also local tumor recurrence diagnosed 1 1/2 years after surgery
`and 1 year after collection of stool for K-ras testing. No tumor tissue
`was available for analysis in an additional patient who had rasmuta-
`tion on stool-derived DNA.
`
`/ 5e0d$$0040
`
`04-11-96 15:07:21
`
`gasa
`
`WBS-Gastro
`
`Geneoscopy Exhibit 1054, Page 5
`
`

`

`May 1996
`
`K-ras DETERMINATION IN THE STOOL 1351
`
`The K-ras amplification test on the stool proved to be
`very effective in the identification of subjects with early
`CRC carrying K-ras mutations. In this category of pa-
`tients, the main limit of the test on stool-derived DNA
`is represented by the fact that only about 50% of colon
`cancers have K-ras mutations.19,20 However, our results
`show that the test can be most useful in selected popula-
`tions such as in first-degree relatives of patients with
`colon cancer or patients with a personal history of colon
`cancer and recurrence of adenomas. The latter patients
`have an increased prevalence of ras mutations as shown
`by McLellan et al.21
`In the group of first-degree relatives of patients with
`CRC, all patients but 2 with adenomas ú1 cm in diame-
`ter were identified by the presence of K-ras mutation, a
`much higher percentage than that evidenced by the fecal
`occult blood test in a similar series of patients.3,7 It should
`be noted that the high percentage of K-ras mutation
`found in relatives is consistent with the recognized higher
`risk of these individuals for developing CRC.22 – 24 The
`high percentage of K-ras mutations found agrees with
`the figures reported in class II adenomas that have exten-
`sive high-grade dysplasia and large diameters but that
`have not progressed to carcinoma; these figures may range
`between 67% and 100%.18
`The finding in patients with IBD can be related with
`the recent finding of an accumulation of genetic alter-
`ations similar to that occurring in sporadic colorectal
`carcinogenesis also in colitis-associated neoplasms.12 Re-
`garding K-ras, we have shown mutations in 50% of the
`cases studied; comparable percentages have been reported
`in other series.25,26 The ease of K-ras fragment amplifica-
`tion in IBD, which in turn facilitates the identification
`of possible mutations, strongly supports the possibility
`of using K-ras for surveillance of CRC in this subgroup
`of patients, considering that the fecal occult blood test
`in these patients is useless and that the areas that have
`undergone neoplastic transformation may be macroscopi-
`cally unrecognizable.
`The rate of amplification of K-ras fragments was very
`low in subjects with single adenomatous lesions õ1 cm
`in diameter and in general in subjects without abnormal
`colonoscopic findings (such as subjects who had a small
`adenoma excised in the past or subjects without any
`organic disease). This is not surprising because only 10%
`of adenomas õ1 cm carry K-ras mutations.22,27 Similarly,
`very low percentages of K-ras mutations have been found
`in patients who have undergone surgery for colon cancer
`without recurrent adenomas on follow-up (P õ 0.05 vs.
`patients with recurrent adenomas). This finding is not a
`false-negative result of the test but depends on the fact
`that the primary tumors of these patients are character-
`
`ized by a very low percentage of K-ras mutation, as
`already reported by Losi et al.28 and Benhattar et al.29
`On the other hand, it is not surprising that it was possible
`to evidence K-ras mutations in the stool even after resec-
`tion of the tumor; Minamoto et al.30 detected ras muta-
`tions in the normal mucosa of about 20% of patients
`undergoing surgery for CRC, and Tobi et al.31 identified
`ras mutations in the effluent fluids of 2 of 5 patients
`with previous surgery for CRC. The abnormal prolifera-
`tive pattern that has been found in the unaffected mucosa
`of these subjects, determining a higher cellular shedding
`into the intestinal lumen, probably favors the amplifica-
`tion reaction.32,33
`The rate of amplification of wild-type K-ras was very
`high and approached 100% in patients with IBD (both
`ulcerative colitis and Crohn’s disease) and lymphocytic
`colitis. It has been shown that patients with IBD have
`a shortening of the turnover time and an enlargement of
`the proliferative compartment.34,35 This can result in a
`higher shedding of mucosal cells into the intestinal lu-
`men with facilitated amplification of K-ras. The low per-
`centage of amplification obtained in conditions usually
`associated with normally or nearly normal proliferating
`mucosa, i.e., such as in subjects without endoscopic ab-
`normalities or subjects with history of polyps without
`recurrences,36 may therefore be attributable to an insuffi-
`cient content of intestinal cells in stool. Alternatively, it
`can be hypothesized that in normal subjects, DNA can
`undergo higher degradation by apoptosis, whereas in neo-
`plastic cells, this mechanism is not fully operative and
`more DNA remains available for amplification.13 This
`second hypothesis is substantiated also by the results of
`b-globin amplification, which was amplified much more
`easily in patients with either CRC or adenomas than in
`normal subjects. In patients with benign or malignant
`adenomas, it has been repeatedly shown that epithelial
`cells, during progression of carcinogenic process, lose
`normal growth regulatory constraints and acquire an ab-
`normal proliferative behavior with greatly increased
`shedding into the intestinal lumen.28,29 In this context,
`even the significantly higher percentage of amplification
`of K-ras fragments in subjects with large adenomas with
`high-grade dysplasia, in relatives of patients with CRC,
`or in patients with IBD than in subjects without any
`organic abnormality at the time of collection of stool
`may be considered important, because it is the result of
`an abnormally proliferating mucosa. On the other hand,
`an abnormal pattern of cell proliferation has been repeat-
`edly found in each of the different categories of patients
`(first-degree relatives of patients with CRC and patients
`with IBD) in whom K-ras fragments have been easily
`amplified.32 – 38
`
`/ 5e0d$$0040
`
`04-11-96 15:07:21
`
`gasa
`
`WBS-Gastro
`
`Geneoscopy Exhibit 1054, Page 6
`
`

`

`1352 VILLA ET AL.
`
`GASTROENTEROLOGY Vol. 110, No. 5
`
`In conclusion, the amplification of K-ras fragments,
`independently from having an associated mutation,
`seems to provide useful information about the status of
`the colon. Indeed, in the large majority of cancers, in
`adenomas (ú1 cm), in active IBD, and in lymphocytic
`colitis, wild-type K-ras gene was amplified in almost all
`patients (54 of 56; 96.4%), whereas in patients with no
`organic disease and with small adenomas and in patients
`with inactive IBD, this gene was amplified in 15 of 63
`subjects (23.8%). Although this gives a specificity of
`only 76%, the difference between the two groups is still
`highly significant (P õ 0.0001). A careful follow-up of
`these patients is needed to try to explain the exact mean-
`ing of K-ras amplification in absence of demonstrable
`intestinal disease.
`The finding of K-ras mutation would further reinforce
`the significance of amplification of K-ras fragments, espe-
`cially in subgroups with known higher risk of colorectal
`carcinoma (such as first-degree relatives of patients with
`CRC or patients with long-standing ulcerative colitis).
`The possible interference of K-ras mutations derived
`from other sources (e.g., from pancreas)39 should certainly
`be considered; in these subjects, in the presence of nega-
`tive colonoscopy findings, abdominal computerized to-
`mography would be advisable. However, the detection
`of K-ras in stool-derived DNA, as reported also by other
`investigators,31,40,41 seems sufficiently specific and sensi-
`tive to justify a large-scale trial to validate its use as a
`screening test.42
`
`References
`
`1. Wingo PA, Tong T, Bolden S. Cancer statistics. CA Cancer J Clin
`1995;45:8–30.
`2. Winawer SJ, Andrews MNA, Flehinger B, Sherlock P, Schottenfeld
`D, Miller DG. Progress report on controlled trial of fecal occult
`blood testing for the detection of colorectal neoplasia. Cancer
`1980;45:2959–2964.
`3. Allison JE, Feldman R, Tekawa IS. Hemoccult screening in de-
`tecting colorectal neoplasms: sensitivity, specificity and pre-
`dictive value. Long-term follow-up in a large group practice set-
`ting. Ann Intern Med 1990;112:328–333.
`4. Thomas WM, Hardcastle JD. An update on the Nottingham trial
`of faecal occult blood screening for colorectal carcinoma. In:
`Miller AB, Chamberlain J, Day NE, Hakama M, Prorok PC, eds.
`Cancer screening. Cambridge, England: Cambridge University,
`1991:106–115.
`5. Kewenter J, Asztely M, Engaras B, Haglind E, Svanvik J, Ahren
`C. A randomised trial of fecal occult blood testing for early detec-
`tion of colorectal cancer: result of screening and rescreening of
`51.325 subjects. In: Miller AB, Chamberlain J, Day NE, Hakama
`M, Prorok PC, eds. Cancer screening. Cambridge, England: Cam-
`bridge University, 1991:116–125.
`6. Kronborg O, Fenger C, Olsen J. Interim report on a randomised
`trial of screening for colorectal cancer with hemoccult-II. In: Miller
`AB, Chamberlain J, Day NE, Hakama M, Prorok PC, eds. Cancer
`screening. Cambridge, England: Cambridge University, 1991:
`126–130.
`7. Ahlquist DA, Wieand HS, Moertel CG, McGill DB, Loprinzi CL,
`
`O’Connell MJ, Maillard JA, Gerstner JB, Pandya K, Ellefson RD.
`Accuracy of fecal occult blood screening for colorectal neoplasia.
`A prospective study using Hemoccult and Hemoquant tests.
`JAMA 1993;269:1262–1267.
`8. Mandel JS, Bond JH, Church TR, Snover DC, Bradley GM, Schu-
`man LM, Ederer F. Reducing mortality from colorectal cancer by
`screening for fecal occult blood. N Engl J Med 1993;328:1365–
`1371.
`9. Lang CA, Ransohoff DF. Fecal occult blood screening for colo-
`rectal cancer: is mortality reduced by chance selection for screen-
`ing colonoscopy? JAMA 1994;271:1011–1013.
`10. Ransohoff DF, Lang CA. Cost-effectiveness of one-time colonos-
`copy screening to reduce colorectal cancer mortality (abstr). Gas-
`troenterology 1994;106:A24.
`11. Fearon ER, Vogel