Computed Tomographic Colonography and High-Risk Adenomas: A Gap in the Guidelines?

Abstract

Rex DK, Overhiser AJ, Chen SC, et al. (Division of Gastroenterology, Indiana University Medical Center, Indianapolis, Indiana.) Estimation of impact of American College of Radiology Recommendations on CT colonography reporting for resection of high-risk adenoma findings. Am J Gastroenterol 2009;104:149–153.Colorectal cancer (CRC) screening remains underutilized, despite an expanding array of effective test strategies (Ann Intern Med 2008;149:659–669). In March 2008, the American Cancer Society, US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology (ACR) endorsed computed tomographic colonography (CTC) as another acceptable option for CRC screening (Gastroenterology 2008;134:1570–1595). Shortly thereafter, the US Preventive Services Task Force concluded that the evidence base for CTC screening was insufficient, citing incomplete and/or inconsistent sensitivity and specificity estimates for subcentimeter adenomas, among other issues (Ann Intern Med 2008;149:638–658). Although recent advances in CTC hardware, software, and examination techniques have improved test performance, broadly applied reporting criteria may, inadvertently, be contributing to the recognized knowledge gap.Put forward by an ad hoc panel of experts in 2005 (Radiology 2005;236:3–9), the CT Colonography Reporting and Data System (C-RADS) combines colorectal findings and follow-up recommendations into 5 categories (C0–C4). Categories C1 and C2 pertain to subcentimeter polyps: C1 patients with no abnormal findings or no polyp >6 mm in diameter should undergo repeat screening evaluation in 5–10 years; C2 patients with 1 or 2 polyps, 6–9 mm in diameter, and <3 in number should undergo surveillance CTC in no more than 3 years, or be referred to colonoscopy for attempted polypectomy. ACR practice guidelines are generally similar to C-RADS, although specific follow-up recommendations for patients with C2 findings differ somewhat.To estimate the potential impact of C-RADS/ACR reporting criteria on the recognition and management of patients with high-risk (≥1 cm in diameter, villous component, high-grade dysplasia, or ≥3 total number) adenomas, Rex et al conducted a retrospective review of data from consecutive colonoscopies performed between 1999 and 2004 by 9 experienced, staff gastroenterologists at a single, university-based hospital practice (Am J Gastroenterol 2009;104:149–153). After excluding patients with known inflammatory bowel disease, polyposis syndromes, or who were referred for polypectomy, 10,034 colonoscopy patients were identified with a mean age (standard deviation) of 56.0 (± 13.5) years. In total, 5079 patients (51%) had ≥1 polyp and 2907 patients (29%) had ≥1 adenoma, the large majority of which were ≤5 mm in diameter (78%). Villous histology or high-grade dysplasia was detected in 0.87% and 5.3% of adenomas ≤5 and 6–9 mm, respectively; invasive cancers were diagnosed in 0.05% and 0% of neoplasms in these size ranges. High-risk adenoma findings were reported for 1001 patients (10%), with 293 (29%) having no polyps >5 mm and another 184 (18%) having only 1 or 2 adenomas in the 6- to 9-mm range; when limited to the subset of screening colonoscopy patients with high-risk adenoma findings (n = 326/3117), the corresponding percentages were modestly increased at 33% and 23%, respectively. Per the authors, these results suggest that approximately one half of all patients with high-risk adenoma findings would be categorized as normal, or have polypectomy delayed for ≥3 years, if they were initially examined by CTC and managed according to C-RADS/ACR recommendations. Of note, the paucity of existing data referent to possible interdependence between subcentimeter adenoma size and multiplicity on future CRC risk was acknowledged as a limitation to the reported analysis.CommentTo appreciate the potential impact of current CTC reporting guidelines on CRC prevention, some additional factors beyond those highlighted by Rex et al are also worthy of consideration. First, although long-term surveillance has been shown to reduce CRC risk, some experts have proposed that the initial polypectomy contributes approximately 90% of the incidence reduction over the first 5–6 years of follow-up (Gastroenterology 2006;130:1872–1885), highlighting the need for accurate detection of prevalent, clinically significant lesions. Based on data from a recent, prospective CRC screening trial, the sensitivity of state-of-the-art CTC (isotropic imaging using a 64-slice system and submillimeter slice thickness) may be comparable to colonoscopy for detecting advanced neoplasia, irrespective of size (Gut 2009;58:241–248). Given the relatively low prevalence of diminutive and small polyps with advanced histology observed in the Rex study (0.9% and 5.3%, respectively) and others, the major benefit of detecting (and reporting) subcentimeter polyps may relate to their relevance in predicting metachronous neoplasia. Because adenoma size, histology, and number seem to be highly interrelated, evaluating the degree to which each of these factors independently influences CRC risk is complicated (Gastroenterol 2006;130:1872–1885). Thus, as noted in the presently reviewed article, further understanding of (1) the natural history of polyps ≤10 mm in diameter and (2) interactions between adenoma size and number with respect to risk stratification would be informative. Second, it is important to recognize that both C-RADS and ACR guidelines recommend either surveillance CTC or colonoscopy referral for patients with 1 or 2 polyps in the 6- to 9-mm size range. From the ACR Web site: The reporting and recommendations for polyps measuring 6–9 mm may vary, depending on the certainty of the finding and clinical context. When identified with reasonable probability they should be reported. The likelihood that a polyp in this size category will progress to a clinically significant neoplasm diminishes with increasing patient age due to the low likelihood of malignant degeneration in conjunction with the long natural history of this process. In some individuals follow-up CTC at 3–5 years may be acceptable. Recommendations should be based upon consideration of the lesion size, diagnostic confidence, patient's age, and existing comorbid conditions. As the polyp approximates the upper limit of this size threshold, greater emphasis may be placed upon removal if the quality of the colonic preparation is adequate. It might be more appropriate to recommend polypectomy for a high probability polyp measuring 8–9 mm in an individual <70 years of age.The full guidelines are available at: http://www.acr.org/SecondaryMainMenuCategories/quality_safety/guidelines/dx/gastro/ct_colonography.aspx; accessed February 12, 2009). Proper adherence to these recommendations would no doubt diminish the number of patients for whom polypectomy is indicated, but unnecessarily delayed. Third, diagnostic colonoscopy after CTC screening represents an imperfect approach for reidentifying and treating subcentimeter polyps, particularly those ≤5 mm in diameter. Incomplete concordance between examination modalities may be due to false-positive or -negative results by either test. Interestingly, the rate of patients with ≥3 subcentimeter polyps, all <5 mm, reported by Rex et al (27%) is similar to the colonoscopy miss rate for diminutive adenomas (26%) reported from a systematic review of tandem colonoscopy studies, although the miss rate by polyp multiplicity was not provided (Am J Gastroenterol 2006;101:343–350). Fourth, data from a simulation study of 100,000 average-risk subjects suggest that CTC with nonreporting of diminutive lesions is a more cost-effective CRC screening strategy than CTC with no polyp size reporting threshold (Cancer 2007;109:2213–2221). In this model, CTC with nonreporting of diminutive polyps also resulted in substantially fewer serious colonoscopy-related complications (n = 351) compared with CTC with no polyp size reporting threshold (n = 691), although the number of CRC cases prevented was only minimally reduced (n = 1073 and n = 1110, respectively).In light of the fact that only about half of adults ≥50 years of age adhere to current CRC screening guidelines (MMWR 2006;55:308–311), the need for better early detection strategies is undeniable. Improved compliance with complete structural evaluations (ie, colonoscopy or CTC, assuming appropriate quality metrics are achieved), should serve to facilitate a continued decline in CRC mortality rates. Further attempts to integrate radiology and gastroenterology efforts in CRC screening, for example, creating an efficient infrastructure to routinely afford same-day follow-up for abnormal CTCs or incomplete colonoscopies, should also result in more positive patient experiences. With this perspective in mind, moving CRC screening from a situation where “confusion reigns” (Cancer Epidemiol Biomarkers Prev 2008;17:2205–2207) to a point where available strategies are applied in complementary fashion seems imminently possible. Rex DK, Overhiser AJ, Chen SC, et al. (Division of Gastroenterology, Indiana University Medical Center, Indianapolis, Indiana.) Estimation of impact of American College of Radiology Recommendations on CT colonography reporting for resection of high-risk adenoma findings. Am J Gastroenterol 2009;104:149–153. Colorectal cancer (CRC) screening remains underutilized, despite an expanding array of effective test strategies (Ann Intern Med 2008;149:659–669). In March 2008, the American Cancer Society, US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology (ACR) endorsed computed tomographic colonography (CTC) as another acceptable option for CRC screening (Gastroenterology 2008;134:1570–1595). Shortly thereafter, the US Preventive Services Task Force concluded that the evidence base for CTC screening was insufficient, citing incomplete and/or inconsistent sensitivity and specificity estimates for subcentimeter adenomas, among other issues (Ann Intern Med 2008;149:638–658). Although recent advances in CTC hardware, software, and examination techniques have improved test performance, broadly applied reporting criteria may, inadvertently, be contributing to the recognized knowledge gap. Put forward by an ad hoc panel of experts in 2005 (Radiology 2005;236:3–9), the CT Colonography Reporting and Data System (C-RADS) combines colorectal findings and follow-up recommendations into 5 categories (C0–C4). Categories C1 and C2 pertain to subcentimeter polyps: C1 patients with no abnormal findings or no polyp >6 mm in diameter should undergo repeat screening evaluation in 5–10 years; C2 patients with 1 or 2 polyps, 6–9 mm in diameter, and <3 in number should undergo surveillance CTC in no more than 3 years, or be referred to colonoscopy for attempted polypectomy. ACR practice guidelines are generally similar to C-RADS, although specific follow-up recommendations for patients with C2 findings differ somewhat. To estimate the potential impact of C-RADS/ACR reporting criteria on the recognition and management of patients with high-risk (≥1 cm in diameter, villous component, high-grade dysplasia, or ≥3 total number) adenomas, Rex et al conducted a retrospective review of data from consecutive colonoscopies performed between 1999 and 2004 by 9 experienced, staff gastroenterologists at a single, university-based hospital practice (Am J Gastroenterol 2009;104:149–153). After excluding patients with known inflammatory bowel disease, polyposis syndromes, or who were referred for polypectomy, 10,034 colonoscopy patients were identified with a mean age (standard deviation) of 56.0 (± 13.5) years. In total, 5079 patients (51%) had ≥1 polyp and 2907 patients (29%) had ≥1 adenoma, the large majority of which were ≤5 mm in diameter (78%). Villous histology or high-grade dysplasia was detected in 0.87% and 5.3% of adenomas ≤5 and 6–9 mm, respectively; invasive cancers were diagnosed in 0.05% and 0% of neoplasms in these size ranges. High-risk adenoma findings were reported for 1001 patients (10%), with 293 (29%) having no polyps >5 mm and another 184 (18%) having only 1 or 2 adenomas in the 6- to 9-mm range; when limited to the subset of screening colonoscopy patients with high-risk adenoma findings (n = 326/3117), the corresponding percentages were modestly increased at 33% and 23%, respectively. Per the authors, these results suggest that approximately one half of all patients with high-risk adenoma findings would be categorized as normal, or have polypectomy delayed for ≥3 years, if they were initially examined by CTC and managed according to C-RADS/ACR recommendations. Of note, the paucity of existing data referent to possible interdependence between subcentimeter adenoma size and multiplicity on future CRC risk was acknowledged as a limitation to the reported analysis. CommentTo appreciate the potential impact of current CTC reporting guidelines on CRC prevention, some additional factors beyond those highlighted by Rex et al are also worthy of consideration. First, although long-term surveillance has been shown to reduce CRC risk, some experts have proposed that the initial polypectomy contributes approximately 90% of the incidence reduction over the first 5–6 years of follow-up (Gastroenterology 2006;130:1872–1885), highlighting the need for accurate detection of prevalent, clinically significant lesions. Based on data from a recent, prospective CRC screening trial, the sensitivity of state-of-the-art CTC (isotropic imaging using a 64-slice system and submillimeter slice thickness) may be comparable to colonoscopy for detecting advanced neoplasia, irrespective of size (Gut 2009;58:241–248). Given the relatively low prevalence of diminutive and small polyps with advanced histology observed in the Rex study (0.9% and 5.3%, respectively) and others, the major benefit of detecting (and reporting) subcentimeter polyps may relate to their relevance in predicting metachronous neoplasia. Because adenoma size, histology, and number seem to be highly interrelated, evaluating the degree to which each of these factors independently influences CRC risk is complicated (Gastroenterol 2006;130:1872–1885). Thus, as noted in the presently reviewed article, further understanding of (1) the natural history of polyps ≤10 mm in diameter and (2) interactions between adenoma size and number with respect to risk stratification would be informative. Second, it is important to recognize that both C-RADS and ACR guidelines recommend either surveillance CTC or colonoscopy referral for patients with 1 or 2 polyps in the 6- to 9-mm size range. From the ACR Web site: The reporting and recommendations for polyps measuring 6–9 mm may vary, depending on the certainty of the finding and clinical context. When identified with reasonable probability they should be reported. The likelihood that a polyp in this size category will progress to a clinically significant neoplasm diminishes with increasing patient age due to the low likelihood of malignant degeneration in conjunction with the long natural history of this process. In some individuals follow-up CTC at 3–5 years may be acceptable. Recommendations should be based upon consideration of the lesion size, diagnostic confidence, patient's age, and existing comorbid conditions. As the polyp approximates the upper limit of this size threshold, greater emphasis may be placed upon removal if the quality of the colonic preparation is adequate. It might be more appropriate to recommend polypectomy for a high probability polyp measuring 8–9 mm in an individual <70 years of age.The full guidelines are available at: http://www.acr.org/SecondaryMainMenuCategories/quality_safety/guidelines/dx/gastro/ct_colonography.aspx; accessed February 12, 2009). Proper adherence to these recommendations would no doubt diminish the number of patients for whom polypectomy is indicated, but unnecessarily delayed. Third, diagnostic colonoscopy after CTC screening represents an imperfect approach for reidentifying and treating subcentimeter polyps, particularly those ≤5 mm in diameter. Incomplete concordance between examination modalities may be due to false-positive or -negative results by either test. Interestingly, the rate of patients with ≥3 subcentimeter polyps, all <5 mm, reported by Rex et al (27%) is similar to the colonoscopy miss rate for diminutive adenomas (26%) reported from a systematic review of tandem colonoscopy studies, although the miss rate by polyp multiplicity was not provided (Am J Gastroenterol 2006;101:343–350). Fourth, data from a simulation study of 100,000 average-risk subjects suggest that CTC with nonreporting of diminutive lesions is a more cost-effective CRC screening strategy than CTC with no polyp size reporting threshold (Cancer 2007;109:2213–2221). In this model, CTC with nonreporting of diminutive polyps also resulted in substantially fewer serious colonoscopy-related complications (n = 351) compared with CTC with no polyp size reporting threshold (n = 691), although the number of CRC cases prevented was only minimally reduced (n = 1073 and n = 1110, respectively).In light of the fact that only about half of adults ≥50 years of age adhere to current CRC screening guidelines (MMWR 2006;55:308–311), the need for better early detection strategies is undeniable. Improved compliance with complete structural evaluations (ie, colonoscopy or CTC, assuming appropriate quality metrics are achieved), should serve to facilitate a continued decline in CRC mortality rates. Further attempts to integrate radiology and gastroenterology efforts in CRC screening, for example, creating an efficient infrastructure to routinely afford same-day follow-up for abnormal CTCs or incomplete colonoscopies, should also result in more positive patient experiences. With this perspective in mind, moving CRC screening from a situation where “confusion reigns” (Cancer Epidemiol Biomarkers Prev 2008;17:2205–2207) to a point where available strategies are applied in complementary fashion seems imminently possible. To appreciate the potential impact of current CTC reporting guidelines on CRC prevention, some additional factors beyond those highlighted by Rex et al are also worthy of consideration. First, although long-term surveillance has been shown to reduce CRC risk, some experts have proposed that the initial polypectomy contributes approximately 90% of the incidence reduction over the first 5–6 years of follow-up (Gastroenterology 2006;130:1872–1885), highlighting the need for accurate detection of prevalent, clinically significant lesions. Based on data from a recent, prospective CRC screening trial, the sensitivity of state-of-the-art CTC (isotropic imaging using a 64-slice system and submillimeter slice thickness) may be comparable to colonoscopy for detecting advanced neoplasia, irrespective of size (Gut 2009;58:241–248). Given the relatively low prevalence of diminutive and small polyps with advanced histology observed in the Rex study (0.9% and 5.3%, respectively) and others, the major benefit of detecting (and reporting) subcentimeter polyps may relate to their relevance in predicting metachronous neoplasia. Because adenoma size, histology, and number seem to be highly interrelated, evaluating the degree to which each of these factors independently influences CRC risk is complicated (Gastroenterol 2006;130:1872–1885). Thus, as noted in the presently reviewed article, further understanding of (1) the natural history of polyps ≤10 mm in diameter and (2) interactions between adenoma size and number with respect to risk stratification would be informative. Second, it is important to recognize that both C-RADS and ACR guidelines recommend either surveillance CTC or colonoscopy referral for patients with 1 or 2 polyps in the 6- to 9-mm size range. From the ACR Web site: The reporting and recommendations for polyps measuring 6–9 mm may vary, depending on the certainty of the finding and clinical context. When identified with reasonable probability they should be reported. The likelihood that a polyp in this size category will progress to a clinically significant neoplasm diminishes with increasing patient age due to the low likelihood of malignant degeneration in conjunction with the long natural history of this process. In some individuals follow-up CTC at 3–5 years may be acceptable. Recommendations should be based upon consideration of the lesion size, diagnostic confidence, patient's age, and existing comorbid conditions. As the polyp approximates the upper limit of this size threshold, greater emphasis may be placed upon removal if the quality of the colonic preparation is adequate. It might be more appropriate to recommend polypectomy for a high probability polyp measuring 8–9 mm in an individual <70 years of age. The full guidelines are available at: http://www.acr.org/SecondaryMainMenuCategories/quality_safety/guidelines/dx/gastro/ct_colonography.aspx; accessed February 12, 2009). Proper adherence to these recommendations would no doubt diminish the number of patients for whom polypectomy is indicated, but unnecessarily delayed. Third, diagnostic colonoscopy after CTC screening represents an imperfect approach for reidentifying and treating subcentimeter polyps, particularly those ≤5 mm in diameter. Incomplete concordance between examination modalities may be due to false-positive or -negative results by either test. Interestingly, the rate of patients with ≥3 subcentimeter polyps, all <5 mm, reported by Rex et al (27%) is similar to the colonoscopy miss rate for diminutive adenomas (26%) reported from a systematic review of tandem colonoscopy studies, although the miss rate by polyp multiplicity was not provided (Am J Gastroenterol 2006;101:343–350). Fourth, data from a simulation study of 100,000 average-risk subjects suggest that CTC with nonreporting of diminutive lesions is a more cost-effective CRC screening strategy than CTC with no polyp size reporting threshold (Cancer 2007;109:2213–2221). In this model, CTC with nonreporting of diminutive polyps also resulted in substantially fewer serious colonoscopy-related complications (n = 351) compared with CTC with no polyp size reporting threshold (n = 691), although the number of CRC cases prevented was only minimally reduced (n = 1073 and n = 1110, respectively). In light of the fact that only about half of adults ≥50 years of age adhere to current CRC screening guidelines (MMWR 2006;55:308–311), the need for better early detection strategies is undeniable. Improved compliance with complete structural evaluations (ie, colonoscopy or CTC, assuming appropriate quality metrics are achieved), should serve to facilitate a continued decline in CRC mortality rates. Further attempts to integrate radiology and gastroenterology efforts in CRC screening, for example, creating an efficient infrastructure to routinely afford same-day follow-up for abnormal CTCs or incomplete colonoscopies, should also result in more positive patient experiences. With this perspective in mind, moving CRC screening from a situation where “confusion reigns” (Cancer Epidemiol Biomarkers Prev 2008;17:2205–2207) to a point where available strategies are applied in complementary fashion seems imminently possible. ReplyGastroenterologyVol. 136Issue 7PreviewKisiel et al have identified key issues raised by our study and surrounding computed tomographic colonography (CTC) use, and seem to project an optimistic perspective that CTC use will not be associated with problems in these areas. My perspective is simply that neither optimism nor pessimism is justified in the absence of data, and that we should understand the range of potential outcomes with regard to CTC use. Full-Text PDF