Abstract

Objective

The National Polyp Study is used as the basis of recommendations for colonoscopic surveillance after polypectomy, establishing an interval of 3 years after removal of newly diagnosed adenomas. The aim of this retrospective cohort study was to estimate the incidence of advanced neoplasia after initial colonoscopy and compare the differences among risk groups.

Methods

Patients over 40 years who were referred for initial colonoscopy at six institutes were selected. They were classified into four groups based on the initial colonoscopy: A, patients without any adenoma; B, with adenomas of <6 mm only; C, with adenomas of ≥6 mm; D, with any intramucosal cancer. The index lesion (IL) at follow-up colonoscopy was defined as large adenoma ≥10 mm, intramucosal/invasive cancer.

Results

A total of 5309 patients were enrolled in this study. Overall, median follow-up period was 5.1 years. The numbers of eligible patients in the various subgroups were A, 2006; B, 1655; C, 1123; D, 525. A total of 379 ILs were newly diagnosed during follow-up colonoscopy. The cumulative incidence of ILs in each group was A, 2.6%; B, 6.7%; C, 13.4%; and D, 12.6%.

Conclusions

Patients with any adenomas >6 mm or intramucosal cancer at the initial colonoscopy have a higher risk of advanced neoplasia during follow-up colonoscopy.

INTRODUCTION

Colorectal cancer (CRC) is the third most common cause of cancer mortality in Japan (1). The identification and removal of adenomatous polyps and post-polypectomy surveillance are considered to be crucial for the control of CRC (2,3). However, recommendations for post-polypectomy surveillance in Japan have not been established. In current practice, the intervals between colonoscopies after polypectomy are variable, often annual, and not based on data from randomized clinical trials.

The evolution of CRC from a precursor lesion, the adenoma, was first reported in studies by Morson (4) as the adenoma–carcinoma sequence. The introduction of colonoscopy provided an opportunity for clarifying this sequence because of its ability to examine the entire colon and remove polyps for pathological examination. The epidemiology and natural history of adenomas are not only important for choosing the optimal follow-up policy after polypectomy, but also for evaluating endoscopic screening for colorectal adenomas and cancer. The existence of flat and depressed lesions, including some with advanced histology, has been demonstrated in multiple recent series from several countries in the West and Japan (5–8). However, the clinical significance of flat and depressed (non-polypoid) lesions and whether they actually constitute alternative pathways to CRC is still controversial (9).

In the USA, the National Polyp Study (NPS) carried out since 1980 recommended an interval of at least 3 years between the colonoscopic removal of newly diagnosed adenomatous polyps and follow-up examination (2,3,10). However, the NPS was conducted prior to recent epidemiologic studies documenting the prevalence of non-polypoid lesions in the colorectum as well as other recent studies suggesting improvements in yield at colonoscopy with slower withdrawal times (11). Thus, the Japanese style colonoscopy, which consists of a bowel preparation using polyethylene glycol (PEG) solution given in the morning on the day of colonoscopy, and techniques such as chromoendoscopy required for the diagnosis of non-polypoid neoplasia (6,12,13) were not used and may at least in part explain the discrepancy between the results of NPS and those of the recent epidemiologic studies (14,15). The aim of this multicenter retrospective cohort study was to estimate the incidence of advanced neoplasia including the prevalence of non-polypoid lesions after initial colonoscopy using the Japanese style colonoscopy and to compare the differences among risk groups of such incidences.

PATIENTS AND METHODS

Subjects and Study Design

This multicenter retrospective cohort study was coordinated by the Japan Polyp Study Workgroup (JPSWG), which was set up in 2000 in Japan. Cases of screening patients over 40 years who were referred for initial total colonoscopy at the six institutes (National Cancer Center Hospital, National Cancer Center Hospital East, Akita Red Cross Hospital, Kitasato University East Hospital, Osaka Medical Center for Cancer and Cardiovascular Diseases, Hattori GI Endoscopy and Oncology Clinic) in Japan were followed up for >3 years from 1990 to 1995. Patients who did not have a familial or personal history of familial adenomatous polyposis, hereditary non-polyposis CRC, inflammatory bowel disease, a personal history of polypectomy or invasive CRC or a sessile adenoma with a base >30 mm where a piecemeal resection or closer follow-up would have been needed were selected for this retrospective cohort study. Written informed consent for examination and treatment were obtained from all of the studied patients prior to the procedures. We retrospectively reviewed colonoscopy reports and medical records for all patients.

They were classified into four groups according to the most advanced lesion found at initial colonoscopy: Group A, patients without any adenomatous polyp; Group B, patients with adenomas of <6 mm only; Group C, patients with adenomas of ≥6 mm; Group D, patients with any intramucosal (M) cancer. All adenomatous polyps of >6 mm and M cancers were removed at the initial colonoscopy. The index lesion (IL) diagnosed during follow-up colonoscopy was defined as follows: large adenomatous polyp ≥10 mm, M cancer and invasive cancer. In this study, we analyzed the cumulative incidence of ILs at follow-up colonoscopy for each patient based on the four groups.

Endoscopic Procedures

All patients were prepared for colonoscopy by administering 2–3 l of PEG on the examination day morning. Scopolamine butylbromide (10 mg) or glucagon (0.5 mg) was administered intravenously to patients with no contraindication prior to examination to avoid bowel movements. Medium-length colonoscopes were used, and one man method colonoscopy was performed. During colonoscopy, the location and the size of all detected lesions were documented and evaluated in real time and categorized as non-neoplastic or neoplastic using chromoendoscopy or magnifying chromoendoscopy. The size of the lesions was estimated using open biopsy forceps. Those diagnosed as non-neoplastic lesions were left untreated. If lesions were identified as neoplastic, hot biopsy, snare polypectomy or EMR was performed. Basically, polyps <6 mm were removed by coagulation biopsy (hot biopsy), and flat lesions or those ≥6 mm were treated with loop snare polypectomy or EMR. However, diminutive adenomatous polyps <6 mm were occasionally permitted to be left untreated. Finally, all neoplastic lesions with >6 mm and M cancers were completely removed at the initial colonoscopy. If lesions were diagnosed as invasive cancer, biopsy specimen was taken and patients were referred for surgery.

Histopathological Evaluation

Resected specimens were immediately fixed in 10% buffered formalin solution and subsequently stained with hematoxylin–eosin. Experienced gastrointestinal pathologists evaluated all pathological specimens. Histopathological diagnoses were determined according to the Japanese Research Society for Cancer of the Colon and Rectum (JRSCCR) and the World Health Organization (WHO) criteria (16,17).

Statistical Analysis

The cumulative incidence of ILs during the follow-up period was described by the Kaplan–Meier method. The Kaplan–Meier curves were compared in the four groups, and the cumulative incidence at 1-year, 3-year and the maximum follow-up period was estimated, respectively. For comparison, we re-categorized the above-mentioned four groups (A, B, C, D) into two (A + B, C + D), and the cumulative incidences for the maximum follow-up period between the two groups were compared by a log-rank test. A two-sided P value of <0.05 was considered statistically significant. When the differences of the baseline characteristics between ILs were examined, the chi-squared test was used for the proportion and t-test for continuous variables. All statistical analyses were performed with SPSS statistical software (SPSS, version 16.0J, for Windows, Tokyo, Japan).

RESULTS

Subjects and Outlines of Follow-up Colonoscopy

A total of 5309 patients, including 3328 (63%) male patients, were enrolled in this study as shown in Table 1. Eligible patients were classified into four groups as follows: Group A, 2006 (38%); Group B, 1655 (31%); Group C, 1123 (21%); and Group D, 525 (10%). The mean age was 60.2, 63.2, 63.7 and 65.1 in Groups A, B, C and D, respectively. Overall, the median follow-up period and the frequency of colonoscopy were 5.1 years and 4.1 times, respectively. There were no significant differences in the follow-up period and the number of times in each group. Moreover, the average interval of colonoscopy was 21.3, 17.2, 16.8 and 13.9 months in Groups A, B, C and D, respectively.

Table 1.

Patient characteristics and outlines of follow-up colonoscopy

 Group A Group B Group C Group D Total 
Patients [no. (%)] 2006 (38) 1655 (31) 1123 (21) 525 (10) 5309 
Male sex [no. (%)] 934 (47) 1145 (69) 849 (76) 400 (76) 3328 (63) 
Agea (years) 60.2 ± 9.8 63.2 ± 9.8 63.7 ± 9.1 65.1 ± 9.2 62.4 ± 9.8 
Follow-up periodb (years) 5.2 (3.0–12.3) 5.3 (3.0–10.7) 5.0 (3.0–11.0) 4.8 (3.0–10.2) 5.1 (3.0–12.3) 
Number of exam times of TCSa 3.8 ± 1.7 4.3 ± 1.9 4.1 ± 1.8 4.5 ± 1.7 4.1 ± 1.8 
Interval of TCSa (months) 21.3 ± 11.5 17.2 ± 8.4 16.8 ± 9.2 13.9 ± 6.7 18.3 ± 10.0 
 Group A Group B Group C Group D Total 
Patients [no. (%)] 2006 (38) 1655 (31) 1123 (21) 525 (10) 5309 
Male sex [no. (%)] 934 (47) 1145 (69) 849 (76) 400 (76) 3328 (63) 
Agea (years) 60.2 ± 9.8 63.2 ± 9.8 63.7 ± 9.1 65.1 ± 9.2 62.4 ± 9.8 
Follow-up periodb (years) 5.2 (3.0–12.3) 5.3 (3.0–10.7) 5.0 (3.0–11.0) 4.8 (3.0–10.2) 5.1 (3.0–12.3) 
Number of exam times of TCSa 3.8 ± 1.7 4.3 ± 1.9 4.1 ± 1.8 4.5 ± 1.7 4.1 ± 1.8 
Interval of TCSa (months) 21.3 ± 11.5 17.2 ± 8.4 16.8 ± 9.2 13.9 ± 6.7 18.3 ± 10.0 

aPlus-minus values are mean ± SD.

bMedian (range).

Incidence of IL According to Initial Colonoscopy

A total of 379 ILs were newly diagnosed during follow-up colonoscopy. In Table 2, the incidence of ILs (%) and total cases (in parenthesis) in each group were as follows: Group A, 2.6% (52); Group B, 6.7% (111); Group C, 13.4% (150); and Group D, 12.6% (66). In Groups A, B, C and D, the cumulative incidence of ILs at 1 and 3 years was 0.1/0.8%, 1.0/2.9%, 2.5/5.4% and 2.9/5.7%, respectively. When we re-categorized four groups into two, the cumulative incidence of ILs at 1 and 3 years was 0.5/1.9% and 2.7/5.6% in Group A + B (low-risk group) and Group C + D (high-risk group), respectively. A significant difference was found between the low- and high-risk groups (P < 0.0001) (Fig. 1).

Figure 1.

Comparison of cumulative incidence of index lesion and invasive colorectal cancer between risk groups.

Figure 1.

Comparison of cumulative incidence of index lesion and invasive colorectal cancer between risk groups.

Table 2.

Cumulative incidence of index lesions after initial colonoscopy

 Cumulative incidence (%)
 
n Total number of incidence cases 
 1-year 3-year Maximum follow-up period 
Group A 0.1 0.8 2.6 2006 52 
Group B 1.0 2.9 6.7 1655 111 
Group C 2.5 5.4 13.4 1123 150 
Group D 2.9 5.7 12.6 525 66 

 
Group A + B (low risk) 0.5 1.9 4.5 3661 163 
Group C + D (high risk) 2.7 5.6 13.1 1648 216 
 Cumulative incidence (%)
 
n Total number of incidence cases 
 1-year 3-year Maximum follow-up period 
Group A 0.1 0.8 2.6 2006 52 
Group B 1.0 2.9 6.7 1655 111 
Group C 2.5 5.4 13.4 1123 150 
Group D 2.9 5.7 12.6 525 66 

 
Group A + B (low risk) 0.5 1.9 4.5 3661 163 
Group C + D (high risk) 2.7 5.6 13.1 1648 216 

Clinicopathological Characteristics of ILs

There were 189 (50%), 125 (33%) and 65 (17%) right-sided, left-sided and rectal ILs, respectively, as shown in Table 3. Group A revealed right-sided ILs in 24 (46%), left-sided in 15 (29%) and rectal in 13 (25%). Similarly, Groups B, C and D exhibited right-sided ILs in 59 (53%), 74 (49%) and 32 (48%), left-sided in 32 (29%), 55 (37%) and 23 (35%) and rectal in 20 (18%), 21 (14%) and 11 (17%), respectively.

Table 3.

Clinicopathological characteristics of index lesions in each group

 Group A (n = 52) Group B (n = 111) Group C (n = 150) Group D (n = 66) Total (n = 379) 
Location [no. (%)] 
 Right colona 24 (46) 59 (53) 74 (49) 32 (48) 189 (50) 
 Left colonb 15 (29) 32 (29) 55 (37) 23 (35) 125 (33) 
 Rectum 13 (25) 20 (18) 21 (14) 11 (17) 65 (17) 
Histopathology [no. (%)] 
 Adenoma (≥10 mm) 28 (54) 56 (50) 80 (54) 33 (50) 197 (52) 
 Intramucosal cancer 13 (25) 46 (41) 59 (39) 25 (38) 143 (38) 
 Submucosal cancer 4 (8) 3 (3) 6 (4) 7 (11) 20 (5) 
 Advanced cancer 7 (13) 6 (6) 5 (3) 1 (1) 19 (5) 
 Group A (n = 52) Group B (n = 111) Group C (n = 150) Group D (n = 66) Total (n = 379) 
Location [no. (%)] 
 Right colona 24 (46) 59 (53) 74 (49) 32 (48) 189 (50) 
 Left colonb 15 (29) 32 (29) 55 (37) 23 (35) 125 (33) 
 Rectum 13 (25) 20 (18) 21 (14) 11 (17) 65 (17) 
Histopathology [no. (%)] 
 Adenoma (≥10 mm) 28 (54) 56 (50) 80 (54) 33 (50) 197 (52) 
 Intramucosal cancer 13 (25) 46 (41) 59 (39) 25 (38) 143 (38) 
 Submucosal cancer 4 (8) 3 (3) 6 (4) 7 (11) 20 (5) 
 Advanced cancer 7 (13) 6 (6) 5 (3) 1 (1) 19 (5) 

aCecum–transverse colon.

bDescending–sigmoid colon.

Of these ILs, 197 (52%) were large adenoma ≥10 mm, 143 (38%) were M cancer, 20 (5%) were submucosal (SM) invasive cancer and 19 (5%) were advanced (ADV) cancer. Group A revealed a large adenoma in 28 (54%), M cancer in 13 (25%), SM cancer in 4 (8%) and ADV cancer in 7 (13%). Similarly, Groups B, C and D exhibited large adenoma in 56 (50%), 80 (54%) and 33 (50%), M cancer in 46 (41%), 59 (39%) and 25 (38%), SM cancer in 3 (3%), 6 (4%) and 7 (11%) and ADV cancer in 6 (6%), 5 (3%) and 1 (1%), respectively.

Morphologically, the macroscopic types of ILs apart from ADV cancer were 220 (58%) polypoid, 122 (32%) flat and 18 (5%) depressed lesions (Table 4). Furthermore, concerning the occurrence time of IL, there were 69 (18%), 74 (20%), 50 (13%), 89 (23%) and 97 (26%) within 1, 1–2, 2–3, 3–5 and >5 years, respectively. Group A + B revealed within 1 year occurrence in 21 (13%), 1–2 years in 23 (14%), 2–3 years in 21 (13%), 3–5 years in 44 (27%) and >5 years in 54 (33%). Group C + D exhibited within 1 year occurrence in 48 (22%), 1–2 years in 51 (24%), 2–3 years in 29 (13%), 3–5 years in 45 (21%) and >5 years in 43 (20%).

Table 4.

Clinicopathological characteristics of index lesions in each group

 Group A (n = 52) Group B (n = 111) Group C (n = l 50) Group D (n = 66) Total (n = 379) 
Macroscopic type [no. (%)] 
 Adenoma/early cancer 
  Polypoid 26 (50) 52 (47) 94 (63) 48 (73) 220 (58) 
  Flat 18 (35) 46 (42) 44 (29) 14 (21) 122 (32) 
  Depressed 1 (2) 7 (6) 7 (5) 3 (5) 18 (5) 
 Advanced cancer 7 (13) 6 (5) 5 (3) 1 (1) 19 (5) 
Occurrence time [no. (%)] 
 <l (year) 2 (4) 19 (17) 29 (19) 19 (29) 69 (18) 
 1–2 6 (12) 17 (15) 36 (24) 15 (23) 74 (20) 
 2–3 6 (12) 15 (14) 24 (16) 5 (7) 50 (13) 
 3–5 19 (36) 25 (23) 29 (19) 16 (24) 89 (23) 
 >5 19 (36) 35 (31) 32 (22) 11 (17) 97 (26) 
 Group A (n = 52) Group B (n = 111) Group C (n = l 50) Group D (n = 66) Total (n = 379) 
Macroscopic type [no. (%)] 
 Adenoma/early cancer 
  Polypoid 26 (50) 52 (47) 94 (63) 48 (73) 220 (58) 
  Flat 18 (35) 46 (42) 44 (29) 14 (21) 122 (32) 
  Depressed 1 (2) 7 (6) 7 (5) 3 (5) 18 (5) 
 Advanced cancer 7 (13) 6 (5) 5 (3) 1 (1) 19 (5) 
Occurrence time [no. (%)] 
 <l (year) 2 (4) 19 (17) 29 (19) 19 (29) 69 (18) 
 1–2 6 (12) 17 (15) 36 (24) 15 (23) 74 (20) 
 2–3 6 (12) 15 (14) 24 (16) 5 (7) 50 (13) 
 3–5 19 (36) 25 (23) 29 (19) 16 (24) 89 (23) 
 >5 19 (36) 35 (31) 32 (22) 11 (17) 97 (26) 

Association of Baseline Characteristics with ILs

The 379 patients diagnosed with ILs were older than those without such findings (mean age, 65.4 vs. 62.2 years; P = 0.02). Patients who were classified into Group C + D seemed more likely to be diagnosed with an IL than those who were classified into Group A + B (4.5% vs. 13.1%; P = 0.04) and men seemed more likely than women to have an IL (8.5% vs. 4.8%; P < 0.0001) as shown in Table 5.

Table 5.

Association of baseline characteristics with index lesions

Baseline characteristics Number (%) Index lesion
 
P value 
No (n = 4930) Yes (n = 379) 
Mean agea (year)  62.1 ± 9.7 65.4 ± 9.7 0.02 
Age (year) 
 40–49 487 (9.2) 463 (95.1) 24 (4.9)  
 50–59 1640 (30.9) 1557 (94.9) 83 (5.1)  
 60–69 1882 (35.4) 1737 (92.3) 145 (7.7)  
 >70 1300 (24.5) 1173 (90.2) 127 (9.8)  
Sex 
 Male 3328 (62.7) 3045 (91.5) 283 (8.5) <0.0001 
 Female 1981 (37.3) 1885 (95.2) 96 (4.8)  
Category 
 Group A 2006 (37.8) 1954 (97.4) 52 (2.6) 0.04 
 Group B 1655 (31.2) 1544 (93.3) 111 (6.7)  
 Group C 1123 (21.1) 973 (86.6) 150 (13.4)  
 Group D 525 (9.9) 459 (87.4) 66 (12.6)  
Baseline characteristics Number (%) Index lesion
 
P value 
No (n = 4930) Yes (n = 379) 
Mean agea (year)  62.1 ± 9.7 65.4 ± 9.7 0.02 
Age (year) 
 40–49 487 (9.2) 463 (95.1) 24 (4.9)  
 50–59 1640 (30.9) 1557 (94.9) 83 (5.1)  
 60–69 1882 (35.4) 1737 (92.3) 145 (7.7)  
 >70 1300 (24.5) 1173 (90.2) 127 (9.8)  
Sex 
 Male 3328 (62.7) 3045 (91.5) 283 (8.5) <0.0001 
 Female 1981 (37.3) 1885 (95.2) 96 (4.8)  
Category 
 Group A 2006 (37.8) 1954 (97.4) 52 (2.6) 0.04 
 Group B 1655 (31.2) 1544 (93.3) 111 (6.7)  
 Group C 1123 (21.1) 973 (86.6) 150 (13.4)  
 Group D 525 (9.9) 459 (87.4) 66 (12.6)  

aPlus-minus values are mean ± SD.

Description of Patients Diagnosed with Invasive Cancer Within 3 Years

A total of 13 invasive cancers including three ADV cancers were newly diagnosed during the follow-up period within 3 years as shown in Table 6. The cancers were located in different sites; 8 out of the 13 were located at the sigmoid colon or rectum. The mean size was 14.1 ± 5.6 mm (range: 6–20 mm). Macroscopically, of these invasive cancers, six (46%) were sessile/semi-pedunculated, five (39%) were depressed and two (15%) were flat lesions.

Table 6.

Description of 13 patients diagnosed with invasive cancer during the follow-up period within 3 years

Baseline characteristics
 
Months since initial colonoscopy Location Size/macroscopic type Depth of lesion (T-stage) 
Age (year)/sex Category (group) 
41/M Rectum 8 mm/Is (sessile) SM (T1) 
50/M Sigmoid 10 mm/Is (sessile) SM (T1) 
61/M Sigmoid 13 mm/Isp (semi-pedunculated) SM (T1) 
68/M Sigmoid 15 mm/Isp (semi-pedunculated) SM (T1) 
68/F Cecum 20 mm/IIa + IIc (depressed) SM (T1) 
69/F Transverse 15 mm/IIa (LST-NG) (flat) SM (T1) 
71/M 11 Transverse 20 mm/IIa + IIc (depressed) SM (T1) 
67/F 19 Rectum 20 mm/Is (sessile) MP (T2) 
72/F 24 Rectum 10 mm/IIa + IIc (depressed) MP (T2) 
58/M 25 Ascending 6 mm/IIa + IIc (depressed) SM (T1) 
66/F 26 Transverse 6 mm/Is (sessile) SM (T1) 
47/M 30 Sigmoid 20 mm/IIa + IIc (depressed) SS (T3) 
75/M 32 Sigmoid 20 mm/IIa (LST-NG) (flat) SM (T1) 
Baseline characteristics
 
Months since initial colonoscopy Location Size/macroscopic type Depth of lesion (T-stage) 
Age (year)/sex Category (group) 
41/M Rectum 8 mm/Is (sessile) SM (T1) 
50/M Sigmoid 10 mm/Is (sessile) SM (T1) 
61/M Sigmoid 13 mm/Isp (semi-pedunculated) SM (T1) 
68/M Sigmoid 15 mm/Isp (semi-pedunculated) SM (T1) 
68/F Cecum 20 mm/IIa + IIc (depressed) SM (T1) 
69/F Transverse 15 mm/IIa (LST-NG) (flat) SM (T1) 
71/M 11 Transverse 20 mm/IIa + IIc (depressed) SM (T1) 
67/F 19 Rectum 20 mm/Is (sessile) MP (T2) 
72/F 24 Rectum 10 mm/IIa + IIc (depressed) MP (T2) 
58/M 25 Ascending 6 mm/IIa + IIc (depressed) SM (T1) 
66/F 26 Transverse 6 mm/Is (sessile) SM (T1) 
47/M 30 Sigmoid 20 mm/IIa + IIc (depressed) SS (T3) 
75/M 32 Sigmoid 20 mm/IIa (LST-NG) (flat) SM (T1) 

SM, submucosa; LST-NG, laterally spreading tumor, non-granular; MP, muscularis propria; SS, subserosa.

DISCUSSION

This is the first large multicenter retrospective cohort study to analyze the incidence of advanced neoplasia after initial colonoscopy in Japan. From our data, it is thought that patients with any adenomatous polyps of >6 mm or M cancer at the baseline colonoscopy have a higher risk of ILs rather than the other groups. Some authors have reported that patients categorized into a high-risk group, from the findings of initial colonoscopy, had high recurrence rates of colorectal adenomas. Recurrence rates dependent on adenoma characteristics have been reported as 15–60% within 3–4 years after previous endoscopic removal (3,18–21). In Japan, Yamaji et al. reported that recurrence rates of colorectal neoplasia were estimated to be 7.2% per year in those with no initial neoplasia, 19.3% per year in those with small adenomas and 22.9% per year in those with advanced lesions. However, this study was carried out in an asymptomatic patient cohort, unlike our current study, which includes both symptomatic and asymptomatic cases. For advanced colorectal lesions, the incidence rate was 0.21% per year, whereas recurrence rates in those with small adenomas and advanced lesions were 0.64% and 1.88% per year, respectively. From their study, the recurrence rates after polypectomy were elevated; however, the incidence rates in subjects with no neoplastic lesions initially were quite high (22). In contrast, Lieberman et al. (23) reported from the USA that the cumulative result represents the most advanced lesion found on any colonoscopy performed during the 5.5-year study period. Among 298 patients with no neoplasia at baseline who had follow-up evaluation, 67 (22.5%) had small tubular adenomas (<10 mm), and 2.4% had advanced neoplasia, including 1 (0.3%) patient with cancer. Basically, our results were in agreement with this report. The 5-year incidence of ILs in those with no initial neoplasia (Group A) was 2.6%, in those with small adenomas (Group B), large adenomas (Group C) and M cancers (Group D) were 6.7%, 13.4% and 12.6%, respectively. Moreover, the cumulative incidence of ILs at 1 and 3 years was 0.5/1.9% and 2.7/5.6% in Group A + B (low-risk group) and Group C + D (high-risk group), respectively. These results suggested that a surveillance colonoscopy after initial total colonoscopy should be performed at 3-year for patients without any polyps or with polyps <6 mm (low-risk group). In contrast, it should be performed at 1 year for patients with any large polyp (≥6 mm) or intramucosal cancer (high-risk group).

According to the latest guidelines from the USA, the recommendations for the surveillance interval for patients with one or two small (<10 mm) tubular adenomas with no high-grade dysplasia ranged from 5 to 10 years after baseline colonoscopy. On the other hand, patients with three or more adenomas, high-grade dysplasia, villous features or an adenoma ≥10 mm in size should have a 3-year follow-up colonoscopy (24). Lieberman et al. (23) reported that many of the interval cancers and large adenomas were discovered in the first 36 months after initial colonoscopy, raising issues about the quality of the colonoscopy. Among the 379 ILs, a total of 193 (51%) lesions, including 13 invasive cancers, were newly diagnosed within 3 years in our study, especially 7 SM cancers were detected in the first 12 months. A diagnosis of ILs soon after complete colonoscopy shows that the procedure is not 100% sensitive in identifying prevalent neoplasia. It strongly suggests the possibility that prevalent neoplasia were missed at baseline colonoscopy. Significant miss rates of single colonoscopy, especially for small adenomas, have been estimated on the basis of back-to-back tandem colonoscopy. Rex et al. (25) reported that the miss rate for adenomas ≥10 mm was 6%, for adenomas 6–9 mm was 13% and for adenomas ≤5 mm was 27%. Similarly, in a recent study of virtual colonoscopy, conventional colonoscopy failed to detect 12% of lesions ≥10 mm (26).

From our data, among all ILs except ADV cancer, there were 122 (32%) flat and 18 (5%) depressed lesions. Non-polypoid colorectal neoplasms (NP-CRNs) are considered to have a high malignant potential and a high miss rate compared with polypoid ones of similar size (27–30). Soetikno et al. reported that the overall prevalence of NP-CRNs and NP-CRNs with in situ or SM invasive carcinoma was 9.35% and 0.82%, respectively. They also concluded that NP-CRNs were more likely to contain carcinoma (odds ratio: 9.78) than polypoid lesions, regardless of the size (30). In our study, among all 13 invasive cancers diagnosed during the 3-year follow-up period, there were seven (54%) NP-CRNs (five depressed and two flat lesions). Moreover, the mean size of these lesions was <15 mm in diameter. It is quite difficult to recognize such lesions compared with the polypoid ones; therefore, special attention must be paid to NP-CRNs during colonoscopy. Future advances in image-enhanced endoscopy (31), e.g. narrow band imaging (32–35), autofluorescence imaging (36,37) and chromoendoscopy (38,39), may improve the ability to detect flat and depressed lesions during colonoscopy, whereas the effect of such lesions on clinical outcomes still remains to be established.

The incidence of ILs during follow-up colonoscopy was associated with sex and age in our study. The association of advanced lesions with sex and age was not significant in previous studies (22,40,41); however, it can be concluded that ILs are more likely to develop in males and in older patients. Furthermore, we find that patients with polyps of ≥6 mm or with any M cancer at initial colonoscopy have a very high risk of interval advanced neoplasia during surveillance. Few studies have performed systematic follow-up of patients after curative resection of CRC (42,43). Nava and Pagana followed 240 patients for 4 years after curative resection of CRC. They detected 28 (11.7%) patients with cancer during the follow-up (43). In our high-risk group (Group C+D), 216 (13.1%) patients had ILs including 19 (1.2%) invasive cancers during the follow-up period. The chronology of this makes it more likely that these were missed lesions or followed the ‘de novo pathway’ (44,45) rather than progression of the adenoma–carcinoma sequence.

There are several limitations in this study. First, this present study was a multicenter retrospective cohort study. The number of subjects was probably enough, however, a prospective study would help to overcome some of these limitations. Another point worth mentioning is that we could not investigate the main indication for colonoscopy at the time of initial examination. Therefore, subjects were not limited strictly to asymptomatic patients in this study. Actually, the prevalence of Group A, patients without any adenomatous polyp, was lower than the other study subjects (38% vs. 66%, 63%) (22,23). In addition, we could not evaluate the number of adenomas and adenomas with villous histology at initial colonoscopy. Several studies have found that individuals with either 3 or more adenomas, tubular adenoma ≥10 mm, villous adenoma or adenoma with high-grade dysplasia at a baseline screening colonoscopy have a similarly higher risk of advanced neoplasia within 5 years compared with patients with no polyps or 1 or 2 small (<10 mm) tubular adenomas. On the basis of the results of our current study, a prospective evaluation of these factors would seem logical in order to validate other international guidelines in the Japanese context. Regarding the JPS, we started to recruit the eligible patients since 2003 (46). The JPS is a multicenter randomized controlled trial designed to evaluate CRC surveillance strategies in patients who have undergone complete colonoscopies on two occasions, with the removal of all detected neoplasia by high-resolution colonoscope, including the removal of flat and depressed lesions. The JPS is intended to continue until 2011, and the last step of the randomization process and complete histopathological assessment are ongoing.

In conclusion, there is a strong relationship between the results of baseline colonoscopy and the rate of serious incident lesions during 5 years of surveillance. Patients with any adenomatous polyps of ≥6 mm or M cancer at the initial colonoscopy have a higher risk of advanced lesions compared with the lower risk group. Another issue is that important lesions were missed at the initial colonoscopy and detected during follow-up colonoscopy, although all examinations were performed by experts.

Funding

The study is supported by Grants-in Aid for Clinical Cancer Research (13S-8, 16S-33 and 20S-12) from the Ministry of Health, Labour and Welfare, Japan.

Conflict of interest statement

None declared.

References

1
Saito
H
Screening for colorectal cancer: current status in Japan
Dis Colon Rectum
 , 
2000
, vol. 
43
 (pg. 
S78
-
84
)
2
Winawer
SJ
Zauber
AG
Ho
MN
O'Brien
MJ
Gottlieb
LS
Sternberg
SS
, et al.  . 
Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup
N Engl J Med
 , 
1993
, vol. 
329
 (pg. 
1977
-
81
)
3
Winawer
SJ
Zauber
AG
O'Brien
MJ
Ho
MN
Gottlieb
L
Sternberg
SS
, et al.  . 
Randomized comparison of surveillance intervals after colonoscopic removal of newly diagnosed adenomatous polyps. The National Polyp Study Workgroup
N Engl J Med
 , 
1993
, vol. 
328
 (pg. 
901
-
6
)
4
Morson
B
The polyp-cancer sequence in the large bowel
Proc R Soc Med
 , 
1974
, vol. 
67
 (pg. 
451
-
7
)
5
Kudo
S
Endoscopic mucosal resection of flat and depressed type of early colorectal cancer
Endoscopy
 , 
1993
, vol. 
25
 (pg. 
455
-
61
)
6
Fujii
T
Rembacken
BJ
Dixon
MF
Yoshida
S
Axon
AT
Flat adenomas in the United Kingdom: are treatable cancers being missed?
Endoscopy
 , 
1998
, vol. 
30
 (pg. 
437
-
43
)
7
Rembacken
BJ
Fujii
T
Cairns
A
Dixon
MF
Yoshida
S
Chalmers
DM
, et al.  . 
Flat and depressed colonic neoplasms: a prospective study of 1000 colonoscopies in the UK
Lancet
 , 
2000
, vol. 
355
 (pg. 
1211
-
4
)
8
Tsuda
S
Veress
B
Toth
E
Fork
FT
Flat and depressed colorectal tumors in southern Swedish population: a prospective chromoendoscopic and histopathological study
Gut
 , 
2002
, vol. 
51
 (pg. 
550
-
5
)
9
Winawer
S
Fletcher
R
Rex
D
Bond
J
Burt
R
Ferrucci
J
, et al.  . 
Colorectal cancer screening and surveillance: clinical guideline and rationale—update based on new evidence
Gastroenterology
 , 
2003
, vol. 
124
 (pg. 
544
-
60
)
10
Winawer
S
Zauber
A
O'Brien
M
Gottlieb
LS
Sternberg
SS
Stewart
ET
, et al.  . 
The National Polyp Study design, methods, and characteristics of patients with newly diagnosed polyps
Cancer
 , 
1992
, vol. 
70
 (pg. 
1236
-
45
)
11
Barclay
RL
Vicari
JJ
Doughty
AS
Johanson
JF
Greenlaw
RL
Colonoscopic withdrawal times and adenoma detection during screening colonoscopy
N Engl J Med
 , 
2006
, vol. 
355
 (pg. 
2533
-
41
)
12
Parra-Blanco
A
Nicolas-Perez
D
Gimeno-Garcia
A
Grosso
B
Jimenez
A
Ortega
J
, et al.  . 
The timing of bowel preparation before colonoscopy determines the quality of cleansing, and is a significant factor contributing to the detection of flat lesions: a randomized study
World J Gastroenterol
 , 
2006
, vol. 
12
 (pg. 
6161
-
6
)
13
Chiu
HM
Lin
JT
Wang
HP
Lee
YC
Wu
MS
The impact of colon preparation timing on colonoscopic detection of colorectal neoplasms- a prospective endoscopist-blinded randomized trial
Am J Gastroenterol
 , 
2006
, vol. 
101
 (pg. 
2719
-
25
)
14
Zauber
A
O'Brian
M
Winawer
S
On finding flat adenomas: is the search worth the gain?
Gastroenterology
 , 
2002
, vol. 
122
 (pg. 
839
-
40
)
15
O'brien
MJ
Winawer
SJ
Zauber
AG
Bushey
MT
Sternberg
SS
Gottlieb
LS
, et al.  . 
Flat adenomas in the National Polyp Study: is there increased risk for high-grade dysplasia initially or during surveillance?
Clin Gastroenterol Hepatol
 , 
2004
, vol. 
2
 (pg. 
905
-
11
)
16
Japanese Research Society for Cancer of the Colon and Rectum
General rules for clinical and pathological studies on cancer of the colon, rectum and anus
Jpn J Surg
 , 
1983
, vol. 
13
 (pg. 
557
-
73
)
17
Hamilton
SR
Aaltonen
LA
World Health Organization classification of tumours: Pathology and genetics of tumours of the digestive system
2000
Lyon, France
IARC Press
(pg. 
104
-
19
)
18
van Stolk
RU
Beck
GJ
Baron
JA
Haile
R
Summers
R
Adenoma characteristics at first colonoscopy as predictors of adenoma recurrence and characteristics at follow-up. The Polyp Prevention Study Group
Gastroenterology
 , 
1998
, vol. 
115
 (pg. 
13
-
8
)
19
Noshirwani
KC
van Stolk
RU
Rybicki
LA
Beck
GJ
Adenoma size and number are predictive of adenoma recurrence: implications for surveillance colonoscopy
Gastrointest Endosc
 , 
2000
, vol. 
51
 (pg. 
433
-
7
)
20
Martínez
ME
Sampliner
R
Marshall
JR
Bhattacharyya
AK
Reid
ME
Alberts
DS
Adenoma characteristics as risk factors for recurrence of advanced adenomas
Gastroenterology
 , 
2001
, vol. 
120
 (pg. 
1077
-
83
)
21
Nusko
G
Mansmann
U
Kirchner
T
Hahn
EG
Risk related surveillance following colorectal polypectomy
Gut
 , 
2002
, vol. 
51
 (pg. 
424
-
8
)
22
Yamaji
Y
Mitsushima
T
Ikuma
H
Watabe
H
Okamoto
M
Kawabe
T
, et al.  . 
Incidence and recurrence rates of colorectal adenomas estimated by annually repeated colonoscopies on asymptomatic Japanese
Gut
 , 
2004
, vol. 
53
 (pg. 
568
-
72
)
23
Lieberman
DA
Weiss
DG
Harford
WV
Ahnen
DJ
Provenzale
D
Sontag
SJ
, et al.  . 
Five-year colon surveillance after screening colonoscopy
Gastroenterology
 , 
2007
, vol. 
133
 (pg. 
1077
-
85
)
24
Levin
B
Lieberman
DA
McFarland
B
Andrews
KS
Brooks
D
Bond
J
, et al.  . 
Screening and surveillance for the early detection of colorectal cancer and adenomatous polyps, 2008: a joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology
Gastroenterology
 , 
2008
, vol. 
134
 (pg. 
1570
-
95
)
25
Rex
DK
Cutler
CS
Lemmel
GT
Rahmani
EY
Clark
DW
Helper
DJ
, et al.  . 
Colonoscopic miss rates of adenomas determined by back-to-back colonoscopies
Gastroenterology
 , 
1997
, vol. 
112
 (pg. 
24
-
8
)
26
Pickhardt
PJ
Choi
JR
Hwang
I
Butler
JA
Puckett
ML
Hildebrandt
HA
, et al.  . 
Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults
N Engl J Med
 , 
2003
, vol. 
349
 (pg. 
2191
-
200
)
27
Kudo
S
Colonoscopic diagnosis and management of nonpolypoid early colorectal cancer
World J Surg
 , 
2000
, vol. 
24
 (pg. 
1081
-
90
)
28
Saitoh
Y
Waxman
I
West
AB
Popnikolov
NK
Gatalica
Z
Watari
J
, et al.  . 
Prevalence and distinctive biologic features of flat colorectal adenomas in a North American population
Gastroenterology
 , 
2001
, vol. 
120
 (pg. 
1657
-
65
)
29
Soetikno
R
Friedland
S
Kaltenbach
T
Chayama
K
Tanaka
S
Nonpolypoid (flat and depressed) colorectal neoplasms
Gastroenterology
 , 
2006
, vol. 
130
 (pg. 
566
-
76
)
30
Soetikno
RM
Kaltenbach
T
Rouse
RV
Park
W
Maheshwari
A
Sato
T
, et al.  . 
Prevalence of nonpolypoid (flat and depressed) colorectal neoplasms in asymptomatic and symptomatic adults
JAMA
 , 
2008
, vol. 
299
 (pg. 
1027
-
35
)
31
Kaltenbach
T
Sano
Y
Friedland
S
Soetikno
R
American Gastroenterological Association (AGA) Institute technology assessment on image-enhanced endoscopy
Gastroenterology
 , 
2008
, vol. 
134
 (pg. 
327
-
40
)
32
Sano
Y
Muto
M
Tajiri
H
Ohtsu
A
Yoshida
S
Optical/digital chromoendoscopy during colonoscopy using narrow-band image system
Digestive Endoscopy
 , 
2005
, vol. 
17
 (pg. 
43
-
8
)
33
Chiu
HM
Chang
CY
Chen
CC
Lee
YC
Wu
MS
Lin
JT
, et al.  . 
A prospective comparative study of narrow-band imaging, chromoendoscopy, and conventional colonoscopy in the diagnosis of colorectal neoplasia
Gut
 , 
2007
, vol. 
56
 (pg. 
373
-
9
)
34
Rex
DK
Helbig
CC
High yields of small and flat adenomas with high-definition colonoscopes using either white light or narrow band imaging
Gastroenterology
 , 
2007
, vol. 
133
 (pg. 
42
-
7
)
35
Katagiri
A
Fu
KI
Sano
Y
Ikematsu
H
Horimatsu
T
Kaneko
K
, et al.  . 
Narrow band imaging with magnifying colonoscopy as diagnostic tool for predicting histology of early colorectal neoplasia
Aliment Pharmacol Ther
 , 
2008
, vol. 
27
 (pg. 
1269
-
74
)
36
Uedo
N
Higashino
K
Ishihara
R
Takeuchi
Y
Iishi
H
Diagnosis of colonic adenomas by new autofluorescence imaging system: a pilot study
Digestive Endoscopy
 , 
2007
, vol. 
19
 (pg. 
134
-
8
)
37
Matsuda
T
Saito
Y
Fu
KI
Uraoka
T
Kobayashi
N
Nakajima
T
, et al.  . 
Does Autofluorescence imaging videoendoscopy system improve the colonoscopic polyp detection rate?—a pilot study
Am J Gastroenterol
 , 
2008
, vol. 
103
 (pg. 
1926
-
32
)
38
Fujii
T
Hasegawa
RT
Saitoh
Y
Fleischer
D
Saito
Y
Sano
Y
, et al.  . 
Chromoscopy during colonoscopy
Endoscopy
 , 
2001
, vol. 
33
 (pg. 
1036
-
41
)
39
Hurlstone
DP
Cross
SS
Slater
R
Sanders
DS
Brown
S
Detecting diminutive colorectal lesions at colonoscopy: a randomized controlled trial of pan-colonic versus targeted chromoscopy
Gut
 , 
2004
, vol. 
53
 (pg. 
376
-
80
)
40
Neugut
AI
Jacobson
JS
Ahsan
H
Santos
J
Garbowski
GC
Forde
KA
, et al.  . 
Incidence and recurrence rates of colorectal adenomas: a prospective study
Gastroenterology
 , 
1995
, vol. 
108
 (pg. 
402
-
8
)
41
Rex
DK
Cummings
OW
Helper
DJ
Nowak
TV
McGill
JM
Chiao
GZ
, et al.  . 
5-year incidence of adenomas after negative colonoscopy in asymptomatic average-risk persons
Gastroenterology
 , 
1996
, vol. 
111
 (pg. 
1178
-
81
)
42
Cali
RL
Pitsch
RM
Thorson
AG
Watson
P
Tapia
P
Blatchford
GJ
, et al.  . 
Cumulative incidence of metachronous colorectal cancer
Dis Colon Rectum
 , 
1993
, vol. 
36
 (pg. 
388
-
93
)
43
Nava
HR
Pagana
TJ
Postoperative surveillance of colorectal carcinoma
Cancer
 , 
1982
, vol. 
49
 (pg. 
1043
-
7
)
44
Kudo
S
Kashida
H
Tamura
T
Early colorectal cancer: flat or depressed type
J Gastroenterol Hepatol
 , 
2000
, vol. 
15
 (pg. 
66
-
70
)
45
Goto
H
Oda
Y
Murakami
Y
Tanaka
T
Hasuda
K
Goto
S
, et al.  . 
Proportion of de novo cancers among colorectal cancers in Japan
Gastroenterology
 , 
2006
, vol. 
131
 (pg. 
40
-
6
)
46
Sano
Y
Fujii
T
Oda
Y
Matsuda
T
Kozu
T
Kudo
S
, et al.  . 
A multicenter randomized controlled trial designed to evaluate follow-up surveillance strategies for colorectal cancer: the Japan Polyp Study
Digestive Endoscopy
 , 
2004
, vol. 
16
 (pg. 
376
-
8
)

Appendix

In addition to the authors listed in the author field, following are the authors who contributed equally to this study.

Kinichi Hotta: Department of Gastroenterology, Saku Central Hospital, Nagano.

Nozomu Kobayashi: Endoscopy Division, National Cancer Center Hospital, Tokyo.

Yutaka Saito: Endoscopy Division, National Cancer Center Hospital, Tokyo.

Kuang-I Fu: Division of Digestive Endoscopy and Gastrointestinal Oncology, National Cancer Center Hospital East, Kashiwa.