Prevalence and Characteristics of Nonpolypoid Colorectal Neoplasm in an Asymptomatic and Average-Risk Chinese Population

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2009;7: Prevalence and Characteristics of Nonpolypoid Colorectal Neoplasm in an Asymptomatic and Average-Risk Chinese Population HAN MO CHIU,*,, JAW TOWN LIN,*, CHIEN CHUAN CHEN,*, YI CHIA LEE,*,, WEI CHIH LIAO,*, JIN TUNG LIANG, # CHIA TUNG SHUN,** HSIU PO WANG,* and MING SHIANG WU*,, *Department of Internal Medicine, # Department of Surgery, **Department of Pathology, Health Management Center, and Department of Primary Care Medicine, National Taiwan University Hospital, National Taiwan University College of Medicine, Taipei, Taiwan; Institute of Preventive Medicine and Institute of Epidemiology, College of Public Health, National Taiwan University, Taipei, Taiwan; and Department of Internal Medicine, E-DA Hospital and I-Shou University, Kaohsiung County, Taiwan See related article, Cleary SP et al, on page 1251 in Gastroenterology. Background and Aims: Evidence from Japanese studies suggests that nonpolypoid colorectal neoplasia (NP-CRN) tends to be more pathologically advanced than polypoid neoplasia. However, data are limited regarding the prevalence of NP-CRN in an average-risk population. In addition, the diagnostic yield of the fecal occult blood test (FOBT) in relation to different types of colorectal neoplasms remains unclear. We prospectively investigated the prevalence and characteristics of polypoid and nonpolypoid colorectal lesions in an asymptomatic and averagerisk Chinese population. Methods: The study included 12,731 asymptomatic Chinese subjects (8372 of whom were average-risk subjects) who underwent screening colonoscopy. The prevalence, histopathologic findings, and topographic distribution of polypoid and nonpolypoid colorectal lesions were determined and analyzed. The diagnostic yield of FOBT, in relation to lesion morphology, also was assessed. Results: NP-CRN was detected in 552 (4.3%) asymptomatic and 348 (4.2%) average-risk subjects. The prevalence of depressed NP-CRN was 0.18% in both asymptomatic and average-risk subjects. A higher proportion of smaller-sized but high-grade dysplasia and invasive carcinoma beyond the submucosal layer was noted for depressed NP-CRN compared with flat NP- CRN or polypoid neoplasia. The diagnostic yield of FOBT was comparable in depressed lesions and their polypoid counterparts. Conclusions: The prevalence of NP-CRN is substantial in both asymptomatic and averagerisk Chinese individuals. Some subcategories of NP-CRN in this population tend to have more advanced pathologic characteristics. These findings may lead to modification of screening and prevention strategies for colorectal cancer. The majority of colorectal cancers (CRCs) have been known to arise from preexisting polypoid adenomas. This wellknown adenoma-carcinoma sequence provides an opportunity for screening and prevention of CRC. 1 Accumulating evidence indicates that early detection and removal of colorectal adenomas may greatly reduce the mortality and incidence of CRC. 2 This principle dominated CRC prevention over the past 2 decades until the recognition of nonpolypoid colorectal neoplasia (NP-CRN), which are slightly elevated, flat, or depressed lesions that were discovered and studied by Japanese gastroenterologists. 3,4 Compared with their exophytic counterpart, NP-CRN is more likely to contain advanced pathologic components, such as high-grade dysplasia, and cancer invasion into the submucosa. 5 9 An alternative or de novo pathway has been proposed by which an adenoma rapidly progresses to invasive cancer in the absence of any precursor adenomatous polyp. 3,4 Given their small size and unique morphology, these neoplasms may be more difficult to detect by conventional colonoscopy, and it is almost impossible to identify whether they arise from flat or depressed precursor lesions once they reach an advanced stage. Because the behavior of nonpolypoid adenomas may influence a screening program and recommendations for surveillance interval, elucidation of the clinicopathologic features and prevalence of NP-CRN is of utmost importance. Many Western gastroenterologists have been skeptical of claims that flat or depressed-type colorectal adenomas are associated with a higher risk of high-grade dysplasia and invasive cancers. One study analyzed and reclassified 933 surgically removed sessile adenomas described in the National Polyp Study and found no difference between polypoid and flat adenomas with respect to high-grade dysplasia or invasiveness. 10 However, Soetikno et al 9 recently reported on the prevalence and clinicopathologic characteristics of NP-CRN in US male veterans. 9 Their findings are in accordance with previous studies from Japan, and support the clinical significance of these unique colorectal neoplasms. Rapid transition towards industrialization as well as adoption of a Western lifestyle has resulted in a dramatic increase of CRC in the Asian-Pacific region, particularly among Chinese people. 11 Regardless of previous Japanese studies in symptomatic cohorts, data are limited regarding the prevalence of NP- CRN in an average-risk population. Moreover, the diagnostic yield of fecal occult blood test (FOBT) in relation to colorectal neoplasm morphology remains unclear. In this large prospective study of a series of asymptomatic Chinese people, we Abbreviations used in this paper: CI, confidence interval; CRC, colorectal cancer; FOBT, fecal occult blood test; LST, laterally spreading tumor; NP-CRN, nonpolypoid colorectal neoplasia by the AGA Institute /09/$36.00 doi: /j.cgh

2 464 CHIU ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 4 investigated the prevalence of NP-CRN and its relationship with the result of FOBT using primary screening colonoscopy. Materials and Methods Study Participants We studied a series of 12,828 consecutive subjects who voluntarily underwent screening colonoscopy as part of an annual medical health check-up at the Health Management Center of National Taiwan University Hospital between January 2006 and December The details of this health check-up program have been described elsewhere. 12 Only ethnic Chinese people who underwent total colonoscopy were included in the study. Participants completed a self-administered questionnaire providing details regarding current symptoms, personal medical history, present medications, daily habits, and family history. Participants also provided the following: (1) a personal history of CRC, adenomas, or a polypectomy; (2) a family history of cancer, including CRC; (3) use of medications; and (4) past history of receiving any screening test for CRC. This study was approved by the institutional review board and ethical committee of our hospital. Procedures Fecal occult blood test. All subjects collected their stool samples at home and submitted them during their health check-up. The immune FOBT method (OC-LIGHT V-PC50 and V-PH80; Eiken Chemical Co, Ltd, Tokyo, Japan) was used on a single-spot sample for testing. All colonoscopists were blinded to the results of stool testing. Colonoscopy and treatment of detected lesions. The procedures for performing colonoscopy and recording results and histologic evaluation of biopsy specimens have been described previously. 13 The colonoscopies were performed by 6 experienced colonoscopists who were familiar with chromoendoscopy and morphologic classification of colorectal neoplasms, and had performed a minimum of 3000 colonoscopies. All colonoscopists were asked to spend more than 6 minutes during withdrawing the colonoscopy and inspection. For chromoendoscopy, indigo-carmine dye spraying was used routinely for each identified lesion, and crystal-violet dye staining was used to evaluate the depth of suspected cancerous invasion. Noninvasive neoplasia was resected endoscopically, and invasive lesions or advanced cancers were referred for surgery. Definitions Subjects without the following 6 criteria were considered as average risk: (1) history of CRC, adenoma, or inflammatory bowel disease; (2) criteria for hereditary nonpolyposis colorectal cancer, familial adenomatous polyposis, or other polyposis syndrome; (3) a first-degree relative with CRC; (4) symptoms of colorectal malignancy, such as bloody stool, abdominal pain, change in body weight, or documented irondeficiency anemia; (5) a history of screening tests, including FOBT, endoscopy, or barium studies within 5 years; or (6) long-term use of aspirin, a nonsteroidal anti-inflammatory drug, or cyclooxygenase 2 inhibitor. Subjects who could not be irrigated and suctioned for proper colon preparation also were excluded from analysis. The endoscopic morphology of colorectal neoplasias was categorized according to Paris and Japanese Research Society Figure 1. Macroscopic classification of colorectal neoplasm according to Paris and modified Japanese classifications. classifications with some modifications of subtypes (Figure 1). 14 The location of colorectal neoplasia was defined according to anatomic distribution. The colon above the level of splenic flexure was defined as the proximal colon. All specimens were reviewed by the same pathologist (C.-T.S.), who was unaware of the colonoscopic findings and medical history. Colon neoplasia were classified according to World Health Organization criteria. 15 Advanced colon neoplasia were defined as those lesions with one of the following criteria: lesions larger than 10 mm in diameter, lesions with a villous component, high-grade dysplastic lesions or carcinoma in situ, and lesions with invasive features. Statistical Analysis Comparison of numeric variables was performed with an independent sample t test or the Mann Whitney U test. The Pearson chi-square test or the Fisher exact test was used for categoric variables. All analyses were performed using the SPSS statistical program, version 11.0 (SPSS Inc, Chicago, IL). The reported P values were for 2-sided statistical tests and a P value of less than.05 was considered statistically significant. Results A total of 12,828 asymptomatic ethnic Chinese subjects were enrolled in this study. After excluding subjects who did not meet the inclusion criteria, there were 12,731 asymptomatic subjects eligible for final analysis (55.8% men; mean age, 51.2 y),

3 April 2009 PREVALENCE OF NONPOLYPOID COLORECTAL NEOPLASM 465 Figure 2. Study flow and demographics. IBD, inflammatory bowel disease; FDR-CRC, first-degree relative with colorectal cancer; FAP, familial adenomatous polyposis; HNPCC, hereditary nonpolyposis colorectal cancer of whom were average-risk subjects (53.5% men; mean age, 49.5 y). Of these 12,731 asymptomatic subjects, 12,021 (94.4%) had submitted a stool sample for FOBT (Figure 2). Colorectal neoplasm was detected in 2385 (18.7%) asymptomatic subjects. Among asymptomatic subjects, 418 (3.3%) had advanced colorectal neoplasm. Among average-risk subjects, 1552 (18.2%) had neoplasm and 301 (3.6%) had advanced neoplasm (Table 1). Invasive cancers were noted in 36 (0.28%) asymptomatic subjects and 29 (0.35%) average-risk subjects, respectively. Morphologically, 1833 (14.4%) asymptomatic subjects and 1174 (14.0%) average-risk subjects had protruded neoplasm only. Nonpolypoid neoplasms were detected in 552 (4.3%) asymptomatic and 348 (4.2%) average-risk subjects, respectively. The prevalence of neoplasm was significantly higher Table 1. Prevalence of Neoplasm With Different Morphology and Its Association With Age All asymptomatic subjects (n 12,731) Average-risk subjects (n 8372) 50 (n 4211) P value a 50 (n 4161) 50 Overall (n 7114) P value a (n 8372) 50 (n 5617) Overall (n 12,731) Subjects with Parisian Japanese Neoplasm (%) 2385 (18.7) 591 (10.5) 1794 (25.2) (18.2) 434 (10.4) 1088 (25.8).0001 Advanced neoplasm (%) 418 (3.3) 79 (1.4) 339 (4.8) (3.6) 61 (1.5) 240 (5.7).0001 Protruded neoplasm only (%) 1833 (14.4) 472 (8.4) 1361 (19.1) (14.0) 347 (8.3) 827 (19.6).0001 Protruded neoplasm (%) 0-I Ip, Isp, Is 1991 (15.6) 493 (8.8) 1498 (21.1) (15.2) 361 (8.7) 914 (21.7).0001 Flat neoplasm (%) 0-IIa 523 (4.1) 113 (2.1) 410 (5.8) (3.8) 82 (2.0) 236 (5.6).0001 IIa 466 (3.7) 109 (1.9) 357 (5.0) (3.4) 79 (1.9) 209 (5.0).0001 LST-G 29 (0.23) 2 (0.04) 27 (0.38) (0.22) 1 (0.02) 17 (0.40).0007 LST-F 24 (0.19) 2 (0.04) 22 (0.31) (0.16) 2 (0.05) 11 (0.26).036 Depressed neoplasm (%) 22 (0.17) 5 (0.09) 17 (0.24) (0.17) 4 (0.10) 10 (0.24) IIa IIc IIa IIc 15 (0.12) 1 (0.02) 14 (0.20) (0.12) 1 (0.02) 9 (0.21) IIc or 0-IIa IIc IIc or IIc IIa 8 (0.06) 4 (0.07) 4 (0.06).74 5 (0.06) 3 (0.07) 2 (0.05).94 Cancer (%) 96 (0.75) 16 (0.28) 80 (1.1) (0.84) 10 (0.24) 60 (1.42).0001 Invasive (%) 36 (0.28) 5 (0.09) 31 (0.44) (0.35) 5 (0.12) 24 (0.57).0015 CIS/HGD (%) 61 (0.48) 11 (0.20) 50 (0.70) (0.50) 5 (0.12) 37 (0.88).0001 HGD, high-grade dysplasia; CIS, carcinoma in situ. a P value, comparison of neoplasia prevalence between subjects 50 and those 50 years in each morphologic category.

4 Table 2. Clinicopathologic Characteristics of Colorectal Neoplasm in Study Subjects Morphology Lesion no. (%) Proximally located (%) Lesions in asymptomatic subjects (n 3494) Lesions in average-risk subjects (n 2247) Advanced neoplasia (n 492) (%) 10 mm Villous HGD/CIS Advanced neoplasia (n 359) (%) Invasive Lesion Proximally cancer Overall Proximal no. (%) located (%) 10 mm Villous HGD/CIS Invasive cancer Overall Proximal All lesions a (52.0) (11.0) (6.0) (1.9) (1.0) (14.1) (5.9) (49.4) (12.6) (7.1) (2.1) (1.3) (16.0) (6.1) Protruded (80.9) (50.4) (10.3) (6.6) (1.5) (0.11) (13.6) (4.9) (81.3) (47.9) (12.3) (7.8) (1.8) (0.16) (15.9) (5.4) Flat (17.6) (59.8) (11.2) (3.6) (3.3) (0.81) (12.4) (8.8) (16.7) (57.9) (10.7) (6.4) (2.9) (1.1) (11.5) (7.5) 0-IIa (16.1) (58.0) (2.9) (1.3) (1.6) (0) (4.6) (2.9) (15.3) (56.4) (2.6) (1.2) (1.5) (0) (4.1) (2.0) LST-G (0.83) (82.8) (100) (44.8) (20.7) (6.9) (100) (82.8) (0.8) (77.8) (100) (55.6) (16.7) (11.1) (100) (77.8) LST-F (0.69) (70.8) (100) (8.3) (20.8) (12.5) (100) (70.8) (0.58) (69.2) (100) (0) (23.1) (15.4) (100) (69.2) Depressed (0.69) (54.2) (20.8) (4.2) (29.2) (16.7) (41.7) (8.3) (0.67) (46.7) (26.7) (6.7) (33.3) (26.7) (46.7) (6.7) 0-IIa IIc (0.46) (68.8) (25.0) (6.3) (31.3) (18.8) (43.8) (12.5) (0.45) (50.0) (40.0) (10.0) (40.0) (30.0) (60.0) (10.0) 0-IIc, 0-IIc IIa (0.23) (25.0) (12.5) (0) (25.0) (12.5) (37.5) (0) (0.22) (40.0) (0) (0) (20.0) (20.0) (20.0) (0) NOTE. Percentages in the first column of each population represents the proportion of the total number of lesions (total lesion number 3494 in asymptomatic subjects and 2247 in average-risk subjects), and the percentage in other columns represents the proportion of the total lesion number of each correspondent morphology. HGD, high-grade dysplasia; CIS, carcinoma in situ. a Cancers in this row are all invasive cancers, including those that could not be specified morphologically. 466 CHIU ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 4

5 April 2009 PREVALENCE OF NONPOLYPOID COLORECTAL NEOPLASM 467 Table 3. Comparison of Mean Size (mm) of Malignant Lesions With Different Morphology CIS/HGD (A) (n 59) Invasive cancers (B) (n 12) A B 0-I (n 44) LST (n 16) LST-G (n 8) LST-F (n 8) Depressed lesions (n 11) 0-IIa IIc (n 8) IIc or 0-IIc IIa (n 3) HGD, high-grade dysplasia; CIS, carcinoma in situ. Figure 3. The mean size of the group of invasive cancers with depressed morphology tended to be smaller than either 0-I ( mm; P.093) or LST ( mm; P.057) type lesions. in subjects aged older than 50 in each morphologic category except for 0-IIc lesions. The difference in prevalence between asymptomatic subjects and average-risk subjects for each morphologic category was not statistically significant. The prevalence of colorectal neoplasms with different morphology is listed in Table 1. A total of 3494 neoplastic lesions were detected in the asymptomatic group, whereas 2247 lesions were found in the average-risk group. A higher proportion of nonpolypoid lesions (59.5% [380 of 639] in asymptomatic subjects; 57.4% [224 of 390] in average-risk subjects) was located in the proximal colon compared with their polypoid counterparts (50.4% [1423 of 2826] in the asymptomatic group, P.0001; 47.9% [875 of 1827] in the average-risk group, P.0006). The proportion of advanced lesions was similar among polypoid and nonpolypoid neoplasms (13.6% [383 of 2826] vs 13.5% [86 of 639], P.10 in the asymptomatic group; 15.9% [290 of 1827] vs 12.8% [50 of 390], P.13 in the average-risk group). When nonpolypoid lesions such as laterally spreading tumors (LSTs) and depressed lesions were analyzed separately and compared with protruded lesions (Table 2), a significantly higher proportion of advanced lesions was observed (81.8% [63 of 77] vs 13.6% [383 of 2826], P.0001 in the asymptomatic group; 82.6% [38 of 46] vs 15.9% [290 of 1827], P.0001 in the average-risk group). When the same comparison was made regardless of size, the result was similar. (67.5% [52 of 77] vs 7.4% [208 of 2826], P.0001 in the asymptomatic group; 76.1% [35 of 46] vs 8.8% [160 of 1827], P.0001 in the average-risk group). We compared the size of all lesions with malignant transformation after excluding advanced CRCs that were difficult to classify morphologically. The mean size of the group of invasive cancers with depressed morphology was mm, which tended to be smaller than either 0 to I ( mm; P.093) or LST ( mm; P.057) type lesions (Figure 3 and Table 3). The result of FOBT in relation to neoplasms with different morphologic features is summarized in Table 4. Of 1119 (9.3%) Table 4. Results of FOBT in Subjects With Polypoid Colorectal Neoplasm and NP-CRN Subjects All subjects Subjects with FOBT data FOBT positive FOBT negative Sensitivity, % (95% CI) Specificity, % (95% CI) Neoplasia ( ) 93.1 ( ) Advanced neoplasia ( ) 91.8 ( ) Invasive cancer ( ) 90.9 ( ) 0-I ( ) 92.7 ( ) 0-IIa other than LST ( ) 91.1 ( ) LST-G ( ) 90.8 ( ) LST-F ( ) 90.7 ( ) 0-IIa IIc ( ) 90.7 ( ) 0-IIc or 0-IIc IIa ( ) 90.7 ( ) Depressed lesion ( ) 90.7 ( ) Only 0-I ( ) 92.3 ( ) Only 0-IIa other than LST ( ) 90.9 ( ) and IIa IIc Only LST-G ( ) 90.7 ( ) Only LST-NG ( ) 90.7 ( ) Only 0-IIa IIc ( ) 90.7 ( ) Only 0-IIc or 0-IIc IIa (0 36.9) 90.7 ( ) Only depressed lesion ( ) 90.7 ( )

6 468 CHIU ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 4 (26.9%; 95% CI, 13.8% 46.3%), but the difference was not significant when compared with subjects with only protruded lesions (P.20). Discussion NP-CRN possesses distinct characteristics and has become increasingly important because of both an increase in its detection rate and unresolved controversy regarding its clinical significance. Accurate classification is essential to understanding the disease pathway of both polypoid colorectal neoplasia and NP-CRN, and also the mechanism of the malignant transformation that takes place. Our data reveal that both depressed and flat lesions occur in a substantial proportion of asymptomatic and average-risk populations and are associated with higher invasiveness compared with polypoid lesions. This study s results support the theory that NP-CRN represents a unique entity in colorectal carcinogenesis and the clinical significance of endoscopic morphology. Whether NP-CRN lesions are more prevalent in certain populations remains unclear. Any difference in prevalence in different ethnic groups or countries may be owing, in part, to under-identification of these lesions in Western countries. 7,9,10 As shown in Figures 4 and 5, there was great difficulty in identifying some lesions, even with excellent bowel preparation. Furthermore, delineation of depressed areas was impossible Figure 4. (A) With conventional colonoscopy, only blurred blood vessels are observable. (B) With low-magnification narrow-band imaging, a faint, darkly stained area was observed. (C) With high-magnification narrow-band imaging, a scarce vascular mesh was observed. (D) Lowmagnification chromoendoscopy with indigo-carmine dye spraying reveals a clear delineation of a 0-IIc lesion, sized 0.6 cm. (E) High-magnification chromoendoscopy. (F) High-magnification chromoendoscopy with crystal-violet staining showed Kudo s type IIIs pit pattern in the central depressed area completely distinct from the surrounding colonic mucosa. subjects positive for FOBT, the sensitivity and specificity of FOBT for all neoplasia was 19.8% (95% confidence interval [CI], 18.2% 21.5%) and 93.1% (95% CI, 92.6% 93.6%), respectively. For advanced neoplasia, the sensitivity and specificity was 41.1% (95% CI, 36.3% 46.0%) and 91.8% (95% CI, 91.3% 92.3%), respectively. The sensitivity of FOBT varied among subjects who had a protruded lesion (19.8%; 95% CI, 18.1% 21.7%), laterallyspreading tumor, granular type (LST-G) (37.0%; 95% CI, 21.5% 55.9%), laterally-spreading tumor, flat type (LST-F) (38.1%; 95% CI, 20.7% 59.3%), and depressed lesion (27.3%; 95% CI, 13.2% 48.4%). Subjects with LST had a higher FOBT sensitivity (37.5%; 95% CI, 25.2% 51.7%) than those with protruded lesions (LST vs 0 I, P.005; LST vs depressed lesions, P.57). There was no significant difference between subjects with 0 to I and subjects with depressed lesions (P.42). Similar results were seen when the FOBT sensitivity of subjects with only protruded lesions (18.7%; 95% CI, 16.9% 20.6%) and subjects with only depressed lesions (14.3%; 95% CI, 4.3% 40.5%) were compared (P.94). Subjects with only LST had a higher sensitivity Figure 5. (A and B) The visibility and morphology varied with the amount of air insufflated during colonoscopy. (C) The 0-IIc lesion and the depressed area is clearly visible with indigo-carmine dye spraying. (D) Magnification showing Kudo s type IIIs pits are observable at the depressed area after crystal-violet dye staining. (E and F) Resection of the lesion.

7 April 2009 PREVALENCE OF NONPOLYPOID COLORECTAL NEOPLASM 469 Figure 6. (A) A 10-mm lesion with a depressed area that was almost impossible to observe with conventional colonoscopy. (B) Chromoendoscopy showed a well-demarcated central depressed area. (C) NBI also showed with clarity the depressed area with distinct capillary morphology and distribution. (D) Magnified image after crystal violet dye staining; Kudo s type V I pits can be observed and malignant transformation was diagnosed. (E) Endoscopic Mucosal Resection specimen. (F) Pathology showed a well-differentiated adenocarcinoma with submucosal invasion of 500 m without vessel or lymphatic invasion. without dye spraying. Such lesions were prone to being missed or misclassified as protruded types, particularly if dye spraying was not used (Figure 6). Thus, underdiagnosis and misclassification may happen and may contribute to underestimation of such lesions. In our study, the proportion of lesions with dysplasia or malignant transformation did not differ significantly between 0-IIa and protruded types. This indicated that the clinicopathologic features of 0-IIa, and 0-Is and 0-Ip are very similar. We concluded that distinguishing these lesions on the basis of morphologic category was of little clinical value. In contrast, it is crucial to distinguish depressed lesions from other CRN because the former carry greater risk of high-grade dysplasia and malignant transformation. Without chromoendoscopy, the well-demarcated depressed area of depressed lesions is difficult to identify and could be misclassified into 0-IIa lesions, even with the aid of narrow-band imaging (Figure 4). When analyzing the characteristics of those lesions, we counted depressed lesions separately from other NP-CRN and found that the proportion of lesions carrying invasive cancer or severe dysplasia was significantly higher in depressed lesions compared with protruded ones or other 0-II lesions. Evidence in support of the clinical importance of NP-CRN comes from several clinical and simulation studies. Evidence from the National Polyp Study showed reduction of CRC incidence after polypectomy, but the protective effect is not complete. 2 Goto et al 16 estimated that de novo cancer might comprise up to 18.6% of early CRCs in men and 27.4% in women in a cohort of 14,817 Japanese people. Another casecohort study from Taiwan found that at least 30% of CRC arises de novo. 17 In the current study, we found that about one third of morphologically classifiable invasive cancers, regardless of presence or absence of symptoms or risk category, were composed of depressed lesions. Our data concur with these 2 studies. In short, these studies highlight the importance of recognizing these unique lesions to prevent CRC. We also investigated the diagnostic yield of FOBT in subjects with NP-CRN. As shown in Table 4, the sensitivity of FOBT was not statistically different between subjects with protruded lesions, 0-IIa lesions other than LST, and depressed lesions. Lesions of the LST type constituted an exception and were associated with a significantly higher sensitivity, probably owing to larger size. When subjects with protruded, LST, or depressed lesions alone were analyzed separately, the positive rate of FOBT in subjects with only depressed lesions was similar to that in subjects with only protruded neoplasm. Whether such lesions more easily become FOBT positive as they progress to invasive cancer could not be answered by this study because of the small number of depressed lesions. Several critically important issues that may impact greatly on the detection of NP-CRN during screening and surveillance colonoscopy should be addressed. Our previous study has shown that better colon cleansing conditions substantially improve the adenoma detection rate. 18 Previous studies also have indicated that the adenoma detection rate was higher if more time was spent on inspection during colonoscopy. 19 Therefore, it is very possible that the better the colon cleansing conditions and the longer the time spent on colonoscopic inspection, the higher the detection rate for NP-CRN. Moreover, a recent study also indicated that endoscopists who are more familiar with the knowledge of NP-CRN have a higher overall adenoma detection rate and NP-CRN detection rate. 20 Routine practice of chromoendoscopy also can help delineate and characterize NP-CRN and probably makes detection of such lesions easier. 21,22 Collectively, endoscopists should pay much more attention to the quality indicators of colonoscopy and learn more about the clinical characteristics of NP-CRN to improve the detection of NP-CRN. In conclusion, this study provides insight into the prevalence and clinicopathologic characteristics of NP-CRN in an asymptomatic and average-risk Chinese population and characterizes the prevalence of NP-CRN and the results of FOBT among average-risk Chinese people. It was designed with enough power to provide findings that are accurately representative of asymptomatic subjects undergoing primary colonoscopic screening. Whether newly developed endoscopic technologies such as narrow-band imaging or highdefinition endoscopy can improve the detection rate and diagnosis of NP-CRN, especially depressed lesions, remains unclear and requires further investigation Further studies among different populations and clinical settings, as well

8 470 CHIU ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 7, No. 4 as molecular studies to elucidate tumorigenesis/carcinogenesis, likely will improve our understanding of the clinical significance of NP-CRN. References 1. Vogelstein B, Fearon ER, Hamilton SR, et al. Genetic alterations during colorectal-tumor development. N Engl J Med 1988;319: Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 1993;329: Muto T, Kamiya J, Sawada T, et al. Small flat adenoma of the large bowel with special reference to its clinicopathologic features. Dis Colon Rectum 1985;28: Kudo S, Tamura S, Hirota S, et al. The problem of de novo colorectal carcinoma. Eur J Cancer 1995;31A: Jaramillo E, Watanabe M, Slezak P, et al. Flat neoplastic lesions of the colon and rectum detected by high-resolution video endoscopy and chromoscopy. Gastrointest Endosc 1995;42: Rembacken BJ, Fujii T, Cairns A, et al. Flat and depressed colonic neoplasms: a prospective study of 1000 colonoscopies in the UK. Lancet 2000;355: Saitoh Y, Waxman I, West AB, et al. Prevalence and distinctive biologic features of flat colorectal adenomas in a North American population. Gastroenterology 2001;120: Tsuda S, Veress B, Toth E, et al. Flat and depressed colorectal tumours in a southern Swedish population: a prospective chromoendoscopic and histopathological study. Gut 2002;51: Soetikno RM, Kaltenbach T, Rouse RV, et al. Prevalence of nonpolypoid (flat and depressed) colorectal neoplasms in asymptomatic and symptomatic adults. JAMA 2008;299: O Brien MJ, Winawer SJ, Zauber AG, 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;2: Sung JJ, Lau JY, Goh KL, et al. Increasing incidence of colorectal cancer in Asia: implications for screening. Lancet Oncol 2005;6: Chiu HM, Lin JT, Shun CT, et al. Association of metabolic syndrome with proximal and synchronous colorectal neoplasm. Clin Gastroenterol Hepatol 2007;5: ; quiz Chiu HM, Wang HP, Lee YC, et al. A prospective study of the frequency and the topographical distribution of colon neoplasia in asymptomatic average-risk Chinese adults as determined by colonoscopic screening. Gastrointest Endosc 2005;61: Participants in the Paris Workshop. The Paris endoscopic classification of superficial neoplastic lesions: esophagus, stomach, and colon: November 30 to December 1, Gastrointest Endosc 2003;58:S3 S Hamilton SR, Vogelstein B, Kudo S, et al. Tumours of the colon and rectum. In: Hamilton SR, Aaltonen LA, eds. World Health Organization classifications of tumors: pathology and genetics of tumors of the digestive system. Lyon, France: IARC, 2000: Goto H, Oda Y, Murakami Y, et al. Proportion of de novo cancers among colorectal cancers in Japan. Gastroenterology 2006;131: Chen CD, Yen MF, Wang WM, et al. A case-cohort study for the disease natural history of adenoma-carcinoma and de novo carcinoma and surveillance of colon and rectum after polypectomy: implication for efficacy of colonoscopy. Br J Cancer 2003;88: Chiu HM, Lin JT, Wang HP, et al. The impact of colon preparation timing on colonoscopic detection of colorectal neoplasms a prospective endoscopist-blinded randomized trial. Am J Gastroenterol 2006;101: Barclay RL, Vicari JJ, Doughty AS, et al. Colonoscopic withdrawal times and adenoma detection during screening colonoscopy. N Engl J Med 2006;355: Kalidindi V, Kaltenbach TR, Friedland S, et al. Endoscopists trained in the diagnosis of nonpolypoid (flat and depressed) colorectal neoplasms detect more adenomatous lesions. Gastrointest Endosc 2008;67:AB Fujii T, Hasegawa RT, Saitoh Y, et al. Chromoscopy during colonoscopy. Endoscopy 2001;33: Sano Y, Tanaka S, Teixeira CR, et al. Endoscopic detection and diagnosis of 0-IIc neoplastic colorectal lesions. Endoscopy 2005; 37: Chiu HM, Chang CY, Chen CC, et al. A prospective comparative study of narrow-band imaging, chromoendoscopy, and conventional colonoscopy in the diagnosis of colorectal neoplasia. Gut 2007;56: Adler A, Pohl H, Papanikolaou IS, et al. A prospective randomised study on narrow-band imaging versus conventional colonoscopy for adenoma detection: does narrow-band imaging induce a learning effect? Gut 2008;57: 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;133: Reprint requests Address requests for reprints to: Ming-Shiang Wu, MD, PhD, Department of Internal Medicine, National Taiwan University Hospital, 7 Chung-Shan South Road, Taipei, Taiwan. mingshiang@ntu. edu.tw; fax: (886) Acknowledgments The authors would like to express their special thanks to Dr Takahisa Matsuda of the National Cancer Center Hospital (Japan) and Dr Takahiro Fujii of the TF Clinic (Tokyo, Japan) for providing valuable comments on this study. Conflicts of interest The authors disclose no conflicts. Funding This study was supported by research grant NSC B from the National Science Council of Taiwan.