Fecal occult blood testing (FOBT), flexible sigmoidoscopy

Transcription

1 GASTROENTEROLOGY 2007;132: Comparing Attendance and Detection Rate of Colonoscopy With Sigmoidoscopy and FIT for Colorectal Cancer Screening NEREO SEGNAN,* CARLO SENORE,* BRUNO ANDREONI, ALBERTO AZZONI, LUIGI BISANTI, ALESSANDRO CARDELLI, GUIDO CASTIGLIONE, # CRISTIANO CROSTA,** ANDREA EDERLE, ALBERTO FANTIN, ARNALDO FERRARI, MARIO FRACCHIA, FRANCO FERRERO, STEFANO GASPERONI, SERAFINO RECCHIA, MAURO RISIO, ## TIZIANA RUBECA,*** GIORGIO SARACCO, MARCO ZAPPA, and the SCORE3 Working Group Italy *CPO Piemonte, Torino; Surgery Unit II, Istituto Europeo di Oncologia, Milano; Gastroenterology Unit, Infermi Hospital, ASL 12, Biella; Epidemiology Unit, ASL Citta di Milano, Milano; Gastroenterology Unit, Infermi Hospital, AUSL Rimini, Rimini; # Imaging Unit, CSPO, Firenze; **Gastroenterology Unit, Istituto Europeo di Oncologia, Milano; Gastroenterology Unit, Polo Ospedaliero Est Veronese (VR), Gastroenterology Unit, Maria Vittoria Hospital, Torino; Gastroenterology Unit, Mauriziano Umberto I Hospital, Torino; Gastroenterology Unit, S Giovanni Bosco Hospital, Torino; ## Pathology Unit, Istituto per la Ricerca e la Cura del Cancro, Candiolo (Torino); ***Analytical and Biomolecular Citopathology Unit, CSPO, Firenze; Gastrohepatology Unit, S. Giovanni Battista-Molinette Hospital, Torino; and Clinical Epidemiology Unit, CSPO, Firenze, Italy See editorial on page Background & Aims: We conducted a study to estimate population coverage and detection rate (DR) achievable through different strategies of colorectal cancer (CRC) screening. Methods: A populationbased multicenter randomized trial comparing 3 strategies was used: (1) biennial immunologic fecal occult blood test (FIT), (2) once only sigmoidoscopy (FS), and (3) once only colonoscopy (TC). A random sample of men and women, aged 55 to 64 years, was drawn from general practitioners (GP) rosters. Eligible subjects, randomized within GP, were mailed a personal invitation. Nonresponders in groups 2 and 3 were invited again at 12 and 24 months. Screenees with highrisk distal polyps (villous component >20%, highgrade dysplasia, CRC, size >10 mm, >2 adenomas) at FS, or with positive FIT, were referred for TC. Results: The attendance rate was 32.3% (1965/6075) for FIT, 32.3% (1944/6018) for FS, 26.5% (1597/6021) for TC. FIT detected 2 patients with CRC (0.1%) and 21 with an advanced adenoma (1.1%). The corresponding figures were as follows: 12 (0.6%) and 86 (4.5%) patients, respectively, for FS; 13 (0.8%) and 100 (6.3%) patients, respectively, for TC. To detect 1 advanced neoplasm, it would be necessary to invite 264 people with FIT, 60 with FS, 53 with TC. FS would have detected 27.3% of the proximal advanced neoplasms detected at TC. Assuming the same participation rate at TC as at FS, 48 TCs would be necessary to detect 1 additional advanced neoplasm missed by FS. Conclusions: When participants are offered 1 screening test, participation is lower in a TC than in an FS program. However, DR of advanced neoplasia is higher with TC. Fecal occult blood testing (FOBT), flexible sigmoidoscopy (FS), and total colonoscopy (TC) are recommended as screening tests for colorectal cancer, 1 and they are currently used in Italy, alone or combined in different screening protocols. 2,3 FOBT sensitivity for colorectal cancer (CRC) ranges between 50% and 70% for a single screening round, and it is lower (20% 25%) for large adenomas. 4,5 A screening approach based on sigmoidoscopy (FS) would result in missing a large proportion of CRCs arising in the proximal colon 6 8 because a sizeable proportion of patients with proximal advanced lesions do not harbor polyps in the distal colon. TC has been proposed to overcome the limitations of these screening tests 9 because it allows a complete examination of the large bowel. However, the expected benefit of TC screening depends on technical factors that may prevent a thorough examination, on the rate of complications associated with the procedure, on the distribution of adenomas and CRC throughout the colon, and on the acceptability of the test. Acceptability represents a critical determinant of the impact of an organized program. More acceptable tests would pick up a higher proportion of prevalent lesions, even if their sensitivity were low because more people would attend screening. The natural history of the disease and of its precursor lesions represents another crucial issue influencing the impact of screening. A minority of the population develops CRC before 60 years of age, and there are more adenomas and CRCs in the left than in the right bowel before age 70 years Some studies suggested that the prevalence of distal adenomas shows a plateau at around age 60 years, 14 whereas the prevalence of proximal adenomas appears to plateau at older age. 15 The rationale for designing a multicenter comparative trial is therefore related to the need to obtain reliable Abbreviations used in this paper: CI, confidence interval; CRC, colorectal cancer; DCBE, double contrast barium enema; DR, detection rate; GP, general practitioner; FIT, immunochemical FOBT; FOBT, fecal occult blood testing; FS, flexible sigmoidoscopy; OR, odds ratio; TC, total colonoscopy by the AGA Institute /07/$32.00 doi: /j.gastro

2 June 2007 COLONOSCOPY SCREENING FOR COLORECTAL CANCER 2305 Figure 1. Recruitment, randomization, and attendance rate by center. *Previous diagnosis of CRC, polyps, or IDB accounted for 9.5% of the 1595 exclusions. The proportion of persons excluded because they had undergone a CRC screening test within the previous 2 years was similar for FOBT (19.3%) and for endoscopy (18.6%); patients suffering from severe disease or severe psychiatric symptoms represented 13.3% and 10.9% of the exclusions, respectively, whereas 28.4% of the patients had moved before randomization. & Undeliverable letters excluded. information concerning potential benefits and risks of the adoption of colonoscopy in the context of a mass screening intervention. We conducted a study among average risk population aged 55 to 64 years, comparing the relative performance of immunochemical FOBT (FIT), TC, and FS. The objectives of the study are to assess the attendance and to compare the detection rate (DR) and acceptability of each strategy. We report the results concerning attendance and DR at the initial screening round. Data about adverse effects and acceptability of these strategies will be reported in a separate paper. Materials and Methods Study Design The study was conducted between October 2002 and January 2004 in 6 centers in Italy (Biella, Firenze, Milano, Rimini, Torino, and Verona), following the same design adopted in a previous population-based, multicenter, randomized, controlled trial of different screening strategies. 16 Approval of the study was obtained by the local ethics review committees. Briefly, in each participating center, we draw a population sample from the general practitioners (GP) rosters or the population registers, taking into account the need to avoid including patients already enrolled in the previous studies of CRC screening. In Torino and Milano, all patients in the targeted age range, listed in the rosters of a sample of GPs, were included (GPs already included in the previous trials were excluded from the sampling list). In the other centers, all residents in specific districts or city neighborhoods, aged 55 to 64 years, who had not been recruited in the previous trials were targeted for recruitment (Figure 1). Because the proportion of people involved in the ongoing trials was higher in the older age group, subjects younger than 60 years of age are overrepresented in Biella, Firenze, and Rimini. In all locations, GPs were asked to exclude patients unable to give informed consent or with terminal illness, inflammatory bowel disease, or personal history of polyps or CRC; those with 2 first-degree relatives with CRC; and those who had a colorectal endoscopy or FOBT within the previous 2 years. GPs were also asked to sign the invitation letters and the mail reminders. Between October 2002 and January 2004, all eligible people (men and women, aged 55 to 64 years) were randomized within GP to the 3 screening protocols (ratio 1:1:1): (1) biennial FIT delivered by GP or screening facilities; (2) once-only FS; (3) once-only TC (Figure 1). The randomization was performed in each center by the local coordinating unit using a computer-generated allocation algorithm based on randomized blocks scheme. Patients were randomized on an individual basis, but the algorithm assigned spouses to the same arm. Invitation Procedure Eligible subjects were mailed a personal invitation letter, signed by their GP (or by the local study coordinator if the GP refused to participate). The letter indicated that the subject had been randomized in a study comparing different screening tests. A leaflet containing a brief description of the screening procedure and the operating characteristics of the test to which the subject

3 2306 SEGNAN ET AL GASTROENTEROLOGY Vol. 132, No. 7 had been randomized, and its possible side or adverse effects, was included. A prefixed appointment was offered to individuals allocated to FS or TC screening. Those who agreed to fix a test date were advised to visit their GP, or the screening center, to get the enema for bowel preparation. The invitation letter for FIT invited the patient to contact the GP or the screening center to obtain the kit and the necessary information for performing the test. A reminder letter was mailed to all nonresponders. Two additional invitations were mailed at 12 and 24 months after the first invitation to subjects allocated to the FS or TC arms who had not responded to the previous letters. Test Performance FIT. We adopted an immunochemical test (Immudia-HemSp; Fujirebio Inc., Tokyo) performed on a single sample without any dietary restriction. Although it was suggested that a 2-day fecal collection method may represent a cost-effective and accurate option for CRC screening, 17 there is no clear understanding yet on the best number of fecal specimens to be collected in the context of a mass screening program to balance accuracy and cost-effectiveness. Previous studies conducted in Florence indicated that, compared with a 3-sample FOBT, a single sample FIT was more accurate, 18 and it was associated with an increased participation rate. 19 In addition, the results of the pilot mass screening program in Florence 20 confirmed the extended protective effect of FIT suggested by previous reports, 21 also when using a single sample FIT. All cards collected in each study area were stored at 4 C and shipped weekly to one central laboratory (Laboratorio di Citopatologia, CSPO-Florence, Italy). Patients with positive test were called by the study staff and were offered an appointment date for a TC. A standard response letter was mailed to people with a negative result. All patients randomized in this arm will be invited every second year until age 69 years. Sigmoidoscopy. Bowel preparation was limited to a single enema (133 ml of 22% sodium phosphate) selfadministered at home 2 hours before the test. No dietary restriction was recommended. Screening was undertaken by gastroenterologists in hospital endoscopy units. The aim of the examination was to advance the endoscope beyond the sigmoid-descending colon junction under adequate bowel preparation. If the examination could not be performed because of inadequate preparation, the screenee was invited to repeat the test at a later date. Polyps 10 mm detected during the FS were removed immediately and sent for histologic assessment. Subjects with polyps 10 mm as well as those detected with advanced adenomas (see Polyp Classification) were referred for TC. Subjects with suspected CRC or with polyps too large to be removed endoscopically were referred for surgery. Colonoscopy. Oral bowel preparation with sodium phosphate solution (2 L) was adopted in all centers. Mild dietary restrictions (ie, to increase the uptake of water and to reduce consumption of foods rich in fiber the day before the test) were recommended. Colonoscopy was performed by gastroenterologists in hospital endoscopy units. The aim of the examination was to advance the scope as far as it could be achieved without producing pain or stress to the patient. No standard protocol for sedation was introduced. However, all centers were required to discuss with the patient the advantages and drawbacks of performing sedated colonoscopy. If the baseline colonoscopy could not be completed to the cecum, the patients were referred for a double contrast barium enema (DCBE) whenever advanced adenomas (see Polyp Classification) were detected in the segments examined. DCBE was not routinely indicated in the case of a negative incomplete colonoscopy because of patient intolerance. Polyp Classification Histologic classification of polyps and cancers was based on the World Health Organization criteria. 22 Advanced adenoma was defined as an adenoma with any of the following features: size 10 mm, high-grade dysplasia, villous component 20%. Cancer was defined as the invasion of malignant cells beyond the muscolaris mucosa. Patients with intramucosal carcinoma or carcinoma in situ were classified as having high-grade dysplasia. Polyps detected at FS were defined as distal, even if they were located beyond the sigmoid-descending colon junction, according to the endoscopist report. Polyps detected on TC were defined as distal if they were located in the rectum or sigmoid colon, according to the endoscopist report. A complete TC was reported if the cecum could be visualized or, in the case of failure, when a subsequent TC performed within 6 months after the index one was able to reach the cecum. The combined results of the 2 examinations were included in the analysis. Polyp size was classified according to the diameter of the largest polyp; for each polyp, we used the largest measurement indicated either by the endoscopist or by the pathologist. Patients were classified on the basis of their most advanced lesion to determine the DR of neoplasms. Study Size We planned to enroll 18,000 subjects (3000 in each center). Assuming a 30% attendance rate in the comparison group (FIT) and the conventional 5% (2- tailed) level of statistical significance, the planned groups size allows for an 80% power to detect an absolute 6% difference within groups in each center; a 3% absolute difference in attendance across groups (given the same assumptions) can be declared significant when combining the results from all centers.

4 June 2007 COLONOSCOPY SCREENING FOR COLORECTAL CANCER 2307 Based on the data from the Screening COlon REtto (SCORE) and SCORE2 trials, 16,23 the overall prevalence of advanced neoplasms (advanced adenomas and CRC) in the targeted age range is assumed to be 5% at FS screening, compared with an expected prevalence at TC ranging between 6.9% and 8.7%. 6,7,24 Assuming the conventional 5% (2-tailed) level of statistical significance, the expected attendees group size allows for an 80% power to declare as statistically significant a 2.3% absolute difference in the prevalence of advanced proximal lesions between FS and TC groups (ie, an increase in the overall prevalence from 5.0% to 7.3%). Statistical Methods Odds ratio (OR) and its 95% confidence interval (CI) was used as a measure of association between the outcome of interest and the variables under evaluation in both unadjusted and multivariable analyses. We presented also estimates of attendance standardized on the age distribution of the source population because younger age groups are overrepresented in the study population as a consequence of the sampling restrictions adopted in some centers to avoid contamination in other ongoing studies (see above). For multivariable analyses, a logistic regression model was fitted to the data using the SAS statistical package (SAS-Stat User s guide; version 6. 1st ed., 1989; SAS Institute Inc, Cary, NC). In this analysis, attendance to screening and the presence of advanced adenomas or CRC were modelled as a function of screening modality, gender, age at screening, and trial center to allow for variability in attendance and adenoma or CRC detection rates. Separate models were fitted for FIT-, FS-, and TC-based screening. All statistical tests were 2-sided and were considered statistically significant at P.05. Results Recruitment and Randomization Of 20,042 subjects aged 55 to 64 years listed in the rosters of the 172 GPs included in the study, 1595 (8.0%) were excluded and 18,447 were randomized (Figure 1). The proportion of randomized subjects by gender in each group corresponded to the demographic structure of the source population. 25 On the contrary, the age distribution is shifted toward the younger age group. Attendance The attendance rate was calculated over the 18,114 persons who received the invitation letter, after excluding 333 subjects who could not be traced (1.8%; range across study arms, 1.7% 1.9%). The attendance rate was 32.3% (1965/6075) for FIT, 32.3% (1944/6018) for FS, and 26.5% (1597/6021) for TC (Table 1). After adjusting for screening center, age, and gender (Table 2), the proportion of attendees in the TC arm was significantly lower compared with the FS arm (OR, 0.74; 95% CI: ), whereas the response rate was similar in those invited for FIT compared with FS screening (OR, 1.01; 95% CI: ). Compared with women, men showed a lower response rate (OR, 0.87; 95% CI: ) when invited for FIT screening, whereas their response was higher in the FS (OR, 1.23; 95% CI: ) and TC groups (OR, 1.14; 95% CI: ). Screening Results and Management of Trial Participants FIT. Of 1965 people who performed an FIT, 92 (4.7%) had a positive result. Of these 92 people, 81 (88.0%) underwent TC, and the examination could be completed to the cecum in 89% (72/81) of these cases. Table 1. Attendance Rates by Gender and Age y y Total Invited Attendees Invited Attendees Invited Attendees n n (%) n n (%) n n (%) TC once only Women (24.9) (25.0) (25.0) Men (28.2) (28.3) (28.2) Total a (26.5) (26.5) (26.5) FS once only Women (30.2) (29.6) (29.9) Men (34.7) (35.2) (34.9) Total a (32.4) (32.2) (32.3) FIT Women (32.0) (35.7) (33.5) Men (29.7) (33.4) (31.1) Total a (30.8) (34.6) (32.3) NOTE. Number of patients (n) and proportion of attendees (%). a Because the participation to FOBT was higher among older people, we can estimate that the participation rate to FOBT, standardized to the actual age distribution of the source population, would be 32.8%, whereas the participation to FS and TC would not be changed.

5 2308 SEGNAN ET AL GASTROENTEROLOGY Vol. 132, No. 7 Table 2. Attendance by Age, Gender, and Screening Arm in FIT, FS, and TC Arms and in all Study Population OR a 95% CI FIT Arm y, n y, n Women, n Men, n FS Arm y, n y, n Women, n Men, n TC Arm y, n y, n Women, n Men, n All study population FS once only, n FIT, n TC once only, n y, n y, n Women, n Men, n NOTE. Multivariable ORs, 95% CI. a ORs adjusted for screening center and for all the other variables in the table. Two CRCs (0.1%) and 21 (1.1%) advanced adenomas were detected (Table 3). The positive predictive value for advanced neoplasia (CRC or advanced adenoma) was 28.4% (23/81). FS. Among 1944 attendees, 22 (1.1%) patients, with inadequate bowel preparation, refused to fix another test date. Therefore, FS was performed in 1922 patients: 1730 had a complete (89.0%) and 192 a partial (9.9%) examination of the distal bowel. A total of 138 (7.2%) people were referred for TC, and 124 (89.9%) attended. Colonoscopy could not be completed to the cecum in 7 subjects (5.7%): 4 had a repeat TC within 12 months, 2 underwent a DCBE, and 1 refused further assessments. The DR of distal polyps was 18.9% (Table 3). Adenomas were detected in 214 (11.2%) subjects in the distal colon and distal CRC in 12 (0.6%) subjects. The prevalence of distal advanced neoplasia was 5.2% (100/1922). The prevalence of adenomas in the proximal colon among subjects undergoing TC was 19.5% (25/124); 8 (6.5%) were detected with advanced proximal adenomas. TC. Of 1597 attendees, 33 (2.1%) were offered a new appointment because the examination could not be performed because of inadequate preparation, and 32 reattended. Colonoscopy was reported to be complete to the cecum in 86.7% (1383/1595; range across centers, 85.2% to 88.8%) of the patients examined. Inadequate bowel preparation accounted for 9% of all incomplete examinations; pain and bowel adhesions were reported as the reason for stopping the examination in 44% and 38% of the cases, respectively. Of the 212 people with incomplete TC, 159 (75.0%) were discharged, 25 were referred for a DCBE, 23 were offered an endoscopic follow-up, and 5 were referred for surgery. Colonoscopy was performed with conscious sedation in 33.7% of the cases. The adoption of sedation showed a wide variability Table 3. Most Advanced Lesions, Distal or Proximal, by Gender and Age, by Screening Strategy Women Men y y Total n % n % n % n % n % FIT Patients examined FOBT positive CT performed Nonneoplastic polyps Low-risk adenomas Advanced adenomas CRC FS Patients examined No findings Nonneoplastic polyps Low-risk adenomas Advanced adenomas CRC TC Patients examined No findings Nonneoplastic polyps Low-risk adenomas Advanced adenomas CRC

6 June 2007 COLONOSCOPY SCREENING FOR COLORECTAL CANCER 2309 Table 4. Age-Specific Prevalence of Distal and Proximal Adenomas Among People Attending TC Screening Age, y All distal adenomas (%) a Advanced distal adenomas (%) All proximal adenomas (%) b across centers (range, 12.9% 82.5%), but it was not correlated to the observed completion rate (data not shown). The overall prevalence of colonic polyps was 31.1% (Table 3). The prevalence of adenomas was 11.7% (n 187) in the distal colon and 8.9% (n 144) in the proximal colon, respectively. The prevalence of advanced neoplasia was 5.1% (n 81), including 8 CRCs (0.5%), in the distal colon; and 2.8% (n 44), including 5 CRCs (0.3%), in the proximal colon. Advanced proximal neoplasms were detected in 4.0% of the cases among men and in 1.6% among women (OR, 2.64; 95% CI: ). In the TC arm, the proportion of advanced proximal neoplasia detected adopting the FS screening protocol would have been 27.3% (95% CI: ); this proportion would have been the same for men and women, with a nonsignificant trend toward an increase among people aged 60 years and over (32.0%) compared with those younger than 60 (21.1%) years. In the TC arm, the proportion of patients with proximal advanced neoplasia who did not harbor adenomas in the distal colon was 62.5% among men and 58.3% among women. The prevalence of adenomas in the rectum or sigmoid appeared to plateau after age 59 years, whereas a marked increase of the prevalence of adenomas with age was observed in the proximal colon (Table 4). Comparison of Neoplasia Yield Advanced proximal adenomas (%) NOTE. Distal refers to adenomas detected in the rectum or sigmoid colon. a 2 For linear trend, 1.73; P.18. b 2 For linear trend, 26.97; P.000. The proportion of people detected with advanced neoplasia showed an increase from 5.1% in the FS group to 7.1% among people examined with TC. After adjusting by age, gender, and screening center (Table 5), screening with TC resulted in a 42% increase (OR, 1.42; 95% CI: ) in the DR of advanced neoplasia compared with FS. This gain in neoplasia yield associated almost entirely with the adoption of TC was mainly explained by a marked increase in the DR of advanced neoplasia among people aged 60 years and over (OR, 2.00; 95% CI: ), given that the DR was similar for FS and TC in the younger age group (OR, 1.08; 95% CI: ). In addition, the DR for advanced neoplasia in the distal colon was the same for TC as for FS (OR, 1.02; 95% CI: ); it tended to be higher with TC compared with FS in the older age group, but this trend did not reach the level of statistical significance. Compared with FS, FIT screening resulted in a marked lower DR of advanced neoplasia (OR, 0.22; 95% CI: ). Discussion This is the first large trial comparing uptake and neoplasia yield of TC with other recommended screening methods FS and FIT in an average risk population. Overall, approximately 30% of the people invited were screened. To increase the participation in screening programs, time is needed to allow for the diffusion of the reputation of the program and of the awareness of the advantages and disadvantages of the proposed intervention at the community level. On the other hand, the degree of GPs involvement plays a role in favoring par- Table 5. Prevalence ORs of Advanced Adenomas and CRC by Screening Strategies OR a 95% CI Distal neoplasms y FS, n FIT, n TC, n Women, n Men, n All neoplasms y FS, n FIT, n TC, n Women, n Men, n Distal neoplasms y FS, n FIT, n TC, n Women, n Men, n All neoplasms y FS, n FIT, n TC, n Women, n Men, n All neoplasms All subjects FS, n FIT, n TC, n y, n y, n Women, n Men, n NOTE. Multivariable ORs, 95% CI. a ORs adjusted for screening center and for all other variables in the table.

7 2310 SEGNAN ET AL GASTROENTEROLOGY Vol. 132, No. 7 ticipation. In one center (Verona), all GPs agreed to sign the invitation letters (53% of GPs also available for distributing FIT kits and the bowel preparation for endoscopy). This high degree of GPs involvement was associated with a comprehensive information campaign promoting the project, involving local media (television, radio, newspapers), local administrations, nonprofit organizations, as well as volunteer groups at the community level. Such effort resulted in a 52% attendance rate, although the differences across the proposed strategies remained substantially unchanged (FIT, 56%; FS, 52%; TC, 47%). This result confirms that a higher response rate may be achieved when implementing well-designed information campaigns. The results of the first survey of Italian CRC screening programs, 3 showing a higher attendance rate both in programs adopting FIT (approximately 50%) and in those adopting FS (32% following a single invitation), would support this conclusion. However, because subjects were individually randomized within GP, the observed rates likely reflect the actual relative differences among the strategies tested. Consistent with the results of the previous pilot project conducted in 5 centers involved in the present study, 16 the cumulative attendance achieved with 3 consecutive invitations for FS screening over a 2-year period was about the same as the attendance to 1 FIT screening round. The adoption of TC screening was associated with a decrease in the participation rate in all centers. As in our previous study, the compliance with the invitation for endoscopic screening was higher among men, whereas women and people aged 60 years and over showed a higher response to the invitation for FIT screening. The DR of advanced neoplasia in the distal colon was the same for FS as for TC, suggesting that the potential loss of lesions occurring in those cases with incomplete FS is balanced by the ability of FS to detect lesions in the descending colon (approximately 7% of the advanced neoplasms detected at FS were located in the descending colon according to endoscopist report) when the examination can be adequately completed. The 2.8% prevalence of advanced neoplasia in the proximal colon among subjects examined with TC is consistent with the findings from other studies conducted among average risk people. 24 Adopting the same criteria for referring patients to TC as in the FS group (only patients with CRC or with high-risk distal polyps, ie, size 10 mm, or villous component 20% or high-grade dysplasia, or with 3 adenomas), approximately 27% of these proximal lesions would be detected (11/39 advanced adenomas and 1/5 CRCs). Therefore, based on the observed overall prevalence of advanced neoplasms in the TC group (7.1%) and in the FS group (5.2%), it can be estimated that our FS screening protocol would result in the identification of 72% of people with advanced neoplasia in our target population. This result is consistent with previous studies, which estimated that FS might detect 70% of the advanced neoplasms detected by TC screening 6 and approximately 25% of advanced proximal neoplasms. 15 Other criteria for referral to TC have been adopted in other studies, 6,8,26,27 allowing to achieve a slightly higher yield of advanced proximal neoplasia; however, this was associated with a nearly 3-fold increase in the proportion of people undergoing TC. The prevalence of distal adenomas shows a plateau after age 58 years, whereas the additional contribution of TC below age 60 years seems limited. In addition, contrary to what has been reported in a recent screening study among women, 28 the yield of screening FS would not have been different among women compared with men because the proportion of patients detected at TC screening with a proximal advanced neoplasia who did not harbor adenomas in the distal segments was similar for men and women. A possible explanation for such difference may be the stricter age range of our target population, which showed also a lower prevalence of advanced proximal neoplasia both among men and among women. Our findings support, therefore, the hypothesis that FS may represent an efficient screening strategy for average risk people before age 60 years. 15 Based on the observed participation rate and on the yield of each strategy, we can estimate that offering screening to 10,000 people would result in detecting 168 patients with advanced neoplasm using FS and 191 using TC. This difference should be balanced against the resources necessary for TC screening compared with FS screening. In our setting, to identify 23 additional patients with advanced neoplasms in a population of 10,000 persons would require performance of 2651 TC instead of 3193 FS and 206 TC. Assuming the same participation rate at TC as at FS screening, we can estimate that 48 additional TCs would be necessary to detect 1 additional advanced neoplasm missed by FS. The detection rate of advanced neoplasms has been proposed as a relevant intermediate outcome of screening studies. However, even if our study shows a marked increase in the yield of these lesions associated with the adoption of FS or TC, compared with FIT, the size and the duration of the protective effect of these strategies cannot be inferred from these data. In addition, as expected, a single screening round with FIT resulted in a low yield of advanced neoplasia compared with other tests. The yield of FIT cannot be estimated from a single screening round; to get more accurate estimates of the yield of advanced lesions achievable with FIT, we need to take into account the results of the subsequent tests offered every 2 years. We have also to take into account the cumulative participation rate over an interval equivalent to the interval between 2 FS or TC because previous studies have shown that the proportion of regular attendees to an FOBT tends to decrease over time. 29,30 This information will allow us to determine the number of

8 June 2007 COLONOSCOPY SCREENING FOR COLORECTAL CANCER 2311 FIT tests, as well as the cost and the endoscopic workload, needed to detect the same number of advanced lesions. In conclusion, although we need to follow over several screening rounds people allocated to the FIT arm to get an accurate estimate of the relative yield of advanced neoplasia achievable with this screening strategy, compared with FS or TC, our study suggests that either FIT or FS represent methods capable of being done and accepted by the population in screening programs for CRC. Compared with FS screening, TC would result in a 42% increase in the DR of advanced neoplasia and in a 26% reduction in the response rate. We found as well that the impact of these different strategies, both in terms of uptake and of neoplasia yield, may be influenced by patients preferences and by the natural history of the disease in different age groups. These findings can be useful when planning population-based programs or when building models to estimate the cost-effectiveness ratio of different strategies. Appendix 1 The following are the contributing members of the SCORE3 working group and coauthors of this article: Biella: Angelo Penna (Quality Assurance Unit, Azienda Sanitaria Locale 12, Fondo E Tempia Biella); Giuseppe Malfitana (Gastroenterology Unit, Ospedale degli Infermi, Azienda Sanitaria Locale 12 Biella); Mauro Giudici (Pathology Unit, Ospedale degli Infermi, Azienda Sanitaria Locale 12, Biella); Giovanna Genta and Anna Marutti (Fondo E Tempia, Azienda Sanitaria Locale 12, Biella). Florence: Grazia Grazzini (Imaging Unit, CSPO Florence); Serena Taddei (Imaging Unit, CSPO Florence); Massimo Confortini (Analytical and Biomolecular Citopathology Unit, CSPO Florence); Luca Messerini (Pathology Department, University of Florence). Milan: Giancarla Fiori (Unit of Gastroenterology, European Institute of Oncology); Luisa Marai (Azienda Sanitaria Locale Citta di Milano ); Stefano Bargiggia (Endoscopy Unit, Sacco Hospital, Milano); Cristina Gerosa (Unit of Gastroenterology, S. Carlo Hospital, Milano); Roberto Penagini (Endoscopy Unit, Policlinico Hospital, Milano); Zenia Pirone (Endoscopy Unit, Fatebenefratelli Hospital, Milano); Alfredo Rossi (Endoscopy Unit, Niguarda Hospital, Milano); Edi Viale (Endoscopy Unit, S. Raffaele Hospital, Milano). Rimini: Orietta Giuliani (Gastroenterology Unit, Ospedale degli Infermi, Azienda Unita Sanitaria Locale Rimini); Paolo Rinaldi (Pathology Unit, Ospedale degli Infermi, Azienda Unita Sanitaria Locale Rimini). Turin: Angelo Pera, Marco Tabone (Gastroenterology Unit, Ospedale Mauriziano Umberto I ); Delio Turco (Gastroenterology Unit, Ospedale M. Vittoria, Azienda Sanitaria Locale 3, Torino); Stefano Taraglio (Pathology Unit, Ospedale Ospedale M. Vittoria, Azienda Sanitaria Locale 3, Torino); Mario Rizzetto (Gastrohepatology Unit, Azienda Sanitaria Ospedaliera S. Giovanni Battista, Molinette, Torino); Paola Cassoni (Pathology Unit, University of Turin, and Azienda Sanitaria Ospedaliera S. Giovanni Battista, Molinette, Torino); Maurizio Cosimato, Roberta DiPlacido (Gastroenterology Unit, Ospedale G. Bosco, Azienda Sanitaria Locale 4, Torino); Alessandra Santarelli and Marco Silvani, (Epidemiology Unit, Centro per la Prevenzione Oncologica, Piemonte). Verona: Anna Tomezzoli (Pathology Unit, Azienda Ospedaliera Borgo Trento, Verona) Roberta Benedetti (Gastrohepatology Unit, Polo Ospedaliero Est Veronese). References 1. Winawer SJ, Zauber AG, Fletcher RH, Stillman JS, O Brien MJ, Levin B, Smith RA, Lieberman DA, Burt RW, Levin TR, Bond JH, Brooks D, Byers T, Hyman N, Kirk L, Thorson A, Simmang C, Johnson D, Rex DK; US Multi-Society Task Force on Colorectal Cancer; American Cancer Society. Guidelines for colonoscopy surveillance after polypectomy: a consensus update by the US Multi-Society Task Force on Colorectal Cancer and the American Cancer Society. Gastroenterology 2006;130: Ministero della sanità. Linee Guida elaborate dalla Commissione Oncologica Nazionale, in applicazione di quanto previsto dal Piano sanitario nazionale per il triennio , relativo all azione programmata Prevezione e cura delle malattie oncologiche, concernenti l organizzazione della prevenzione e dell assistenza in oncologia. Gazzetta Ufficiale Della Repubblica Italiana. June 1, Supplemento Ordinario Alla GU, N June 1, Zorzi M, Grazzini G, Senore C, Vettorazzi M. Screening for colorectal cancer in Italy: survey Epidemiol Prev 2006; 30(Suppl 3): Rozen P, Knaani J, Samuel Z. Comparative screening with a sensitive guaiac and specific immunochemical occult blood test in an endoscopy study. Cancer 2000;89: Aalquist DA, Shuber AP. Stool screening for colorectal cancer: evolution from occult blood to molecular markers. Clin Chim Acta 2002;315: Lieberman DA, Weiss DG. One-time screening for colorectal cancer with combined fecal occult-blood testing and examination of the distal colon. N Engl J Med 2001;345: Lieberman DA, Weiss DG, Bond JH, Ahnen DJ, Garewal H, Chejfec G. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. Veterans Affairs Cooperative Study Group 380. N Engl J Med 2000;343: Imperiale TF, Wagner DR, Lin CY, Larkin GN, Rogge JD, Ransohoff DF. Risk of advanced proximal neoplasms in asymptomatic adults according to the distal colorectal findings. N Engl J Med 2000;343: Podolsky DK. Going the distance the case for true colorectal cancer screening. N Engl J Med 2000;343: Rex DK, Chak A, Vasudeva R, Gross T, Lieberman D, Bhattacharya I, Sack E, Wiersema M, Farraye F, Wallace M, Barrido D, Cravens E, Zeabart L, Bjorkman D, Lemmel T, Buckley S. Prospective determination of distal colon findings in average risk patients with proximal colon cancer. Gastrointest Endosc 1999; 49: Clark JC, Collan Y, Eide TJ, Esteve J, Ewen S, Gibbs NM, Jensen OM, Koskela E, MacLennan R, Simpson JG, et al. Prevalence of polyps in an autopsy series from areas with varying incidence of large-bowel cancer. Int J Cancer 1985;36:

9 2312 SEGNAN ET AL GASTROENTEROLOGY Vol. 132, No Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J, eds. Cancer incidence in five continents. Volume VII. IARC Scientific Publications N.143 Lyon. IARC, Ferlay J, Bray F, Pisani P, Parkin DM. GLOBOCAN 2002: cancer incidence, mortality and prevalence worldwide ( IARC Cancer Base No. 5. Version 2.0. Lyon: IARC Press, Atkin WS, Cuzick J, Northover JM, Whynes DK. Prevention of colorectal cancer by once-only sigmoidoscopy. Lancet 1993;341: Geul KW, Bosman FT, Van Blankenstein M, Grobbee DE, Wilson JH. Prevention of colorectal cancer. Costs and effectiveness of sigmoidoscopy. Scand J Gastroenterol 1997;2(Suppl): Segnan N, Senore C, Andreoni B, Arrigoni A, Bisant L, Cardell A, Castiglione G, Crosta C, DiPlacido R, Ferrari A, Ferraris R, Ferrero F, Fracchia M, Gasperoni S, Malfitana G, Recchia S, Risio M, Pizzetto M, Saracco G, Spandre M, Turco D, Turco P, Zappa M, and SCORE2 Working Group Italy. Randomized trial of different screening strategies for colorectal cancer: patient response and detection rates. J Natl Cancer Inst 2005;97: Nakama H, Zhang B, Fattah AS. A cost-effective analysis of the optimum number of stool specimens collected for immunochemical occult blood screening for colorectal cancer. Eur J Cancer 2000;36: Castiglione G, Zappa M, Grazzini G, Mazzotta A, Biagini M, Salvadori P, Ciatto S. Immunochemical vs. guaiac faecal occult blood tests in a population-based screening programme for colorectal cancer. Br J Cancer 1996;74: Grazzini G, Castiglione G, Isu A, Mantellini P, Rubeca T, Sani C, Turco P, Zappa M. Colorectal cancer screening by fecal occult blood testing: results of population based experience. Tumori 2000;86: Zappa M, Castiglione G, Paci E, Grazzini G, Rubeca T, Turco P, Crocetti E, Ciatto S. Measuring interval cancers in populationbased screening using different assays of fecal occult blood testing: the district of Florence experience. Int J Cancer 2001; 92: Saito H, Soma Y, Koeda J, Wada T, Kawaguchi H, Sobue T, Aisawa T, Yoshida Y. Reduction in risk of mortality from colorectal cancer by fecal occult blood screening with immunochemical hemagglutination test. A case control study. Int J Cancer 1995; 61: Jass JR, Sobin LH. Histological typing of intestinal tumors. 2nd edition. New York, Berlin: Springer-Verlag, Segnan N, Senore C, Andreoni B, Aste H, Bonelli L, Crosta C, Ferraris R, Gasperoni S, Penna A, Risio M, Rossini FP, Sciallero S, Zappa M, Atkin WS; SCORE Working Group Italy. Baseline findings of the Italian Multicenter Randomized Controlled Trial of Once-Only Sigmoidoscopy CORE. J Natl Cancer Inst 2002;94: Lewis JD, Ng K, Hung KE, Bilker WB, Berlin JA, Brensinger C, Rustgi AK. Detection of proximal adenomatous polyps with screening sigmoidoscopy. A systematic review and meta-analysis of screening colonoscopy. Arch Intern Med 2003;163: Censimento Generale della Popolazione e delle Abitazioni: October 21, Sistema statistico Nazionale, Istituto Nazionale di Statistica. Roma: Istat, the National Institute of Statistics in Italy, Gondal G, Grotmol B, Hofstad M, Bretthauer M, Eide TJ, Hoff GS. The Norvegian Colorectal Cancer Prevention (NORCCAP) screening study. Baseline findings and implementation of clinical work-up in age groups years. Scand J Gastroenterol 2003; 38: Weissfeld JL, Schoen RE, Pinsky PF, Bresalier RS, Church T, Jurgalevitch S, Austin JH, Prorok PC, Gohagan JK, and PLCO Project Team. Flexible sigmoidoscopy in the PLCO cancer screening trial: results from the baseline screening examination of a randomized trial. J Natl Cancer Inst 2005;97: Schoenfeld P, Cash B, Flood A, Dobhan R, Eastone J, Coyle W, Kikendall JW, Kim HM, Weiss DG, Emory T, Schatzkin A, Lieberman D; CONCeRN Study Investigators. Colonoscopy screening of average risk women for colorectal neoplasia. N Engl J Med 2005; 352: Neilson AR, Whynes DK. Determinants of persistent compliance with screening for colorectal cancer. Soc Sci Med 1995;41: Tazi M, Faivre J, Dassonville F, Lamour J, Milan C, Durand G. Participation in faecal occult blood screening for colorectal cancer in a well-defined French population: results of five screening rounds from 1988 to J Med Screen 1997;4: Received September 11, Accepted February 22, Address requests for reprints to: Nereo Segnan, MD, MS Epi, Unita di Epidemiologia, CPO-Piemonte, ASO S. Giovanni Battista Via S. Francesco da Paola 31, Torino, Italy. nereo.segnan@ cpo.it; fax: (39) Supported by a grant from the Italian League against Cancer (LILT); and by the Istituto Oncologico Romagnolo (IOR), the Fondo E Tempia, the University of Milan, the ULSS 20, and the Piedmont Regional Health Authority for implementation of the study in Rimini, Biella, Milan, Verona, and Turin, respectively; and by SOFAR s.p.a. for providing the enemas for the bowel preparation. The authors thank Federica Gallo for editorial assistance and the members of the Steering Committee of the Italian trials of screening colonoscopy for their support and for their useful suggestions in the design phase of the study: W. S. Atkin, St. Mark s Hospital (London); P. Bruzzi, IST (Genova); F. Pacini, Azienda Ospedaliera Careggi (Firenze); S. J. Winawer, Memorial Sloan Kettering Cancer Center (New York). The following contributed to the implementation of the trial and are members of the SCORE3 working group: Cristina Sani, Analytical and Biomolecular Citopathology Unit, CSPO Florence; Franco Coppola, Roberto Ferraris, Giovanni Galatola, Gastroenterology Unit, Ospedale Mauriziano Umberto I ; Edoardo Formento, Alessandro Repici, Gastrohepatology Unit; Azienda Sanitaria, Ospedaliera S. Giovanni Battista, Molinette Torino; and Antonietta Garripoli, Alessandra Mondardini, Paola Secreto, Gastroenterology Unit, Ospedale M. Vittoria, Azienda Sanitaria, Locale 3 Torino; Rosa Amerio, Tiziano Bonatti, Daniela Brunetti, Fabrizio Cosso, Unit of Epidemiology, Centro per la Prevenzione Oncologica, Torino, Italy. All authors declare that they have no conflict of interest to disclose.