CT Colonography • Negative Screening Colonoscopy

CT Colonography Identifies 90% of Asymptomatic Adults With Polyps > 10 mm

Computed tomography (CT) colonography identifies 90% of asymptomatic adults with polyps > 10 mm (level 1 [likely reliable] evidence), based on a prospective cohort study with 2,600 participants ≥ 50 years old who had colon cancer screening with multidetector CT colonography and same-day colonoscopy. Prevalence of lesions was 5.2% for lesions ≥ 10 mm and 10.2% for lesions 5–9 mm. For adenoma or carcinoma ≥ 10 mm, per-patient predictive performance of CT colonography had 90% sensitivity, 86% specificity, 23% positive predictive value, and 99% negative predictive value. For adenoma or carcinoma ≥ 6 mm, per-patient predictive performance of CT colonography had 78% sensitivity, 88% specificity, 40% positive predictive value, and 98% negative predictive value. Current limitations of CT colonography include serious adverse events (0.24%), lower sensitivity for smaller polyps, positive results require follow-up colonoscopy, and unknown effects of cumulative radiation dose if repeated every five years (N Engl J Med 2008 Sep 18;359(12):1207). These results are consistent with a previously published cohort study with 1,233 asymptomatic adults (N Engl J Med 2003 Dec 4;349(23):2191). For more information, see the virtual colonoscopy topic in DynaMed.

In DynaMed Virtual Colonoscopy Under Computed Tomography Colonography (CTcolonography):

Cohort studies in asymptomatic adults (screening):

• CT colonography identifies 90% asymptomatic adults with polyps > 10 mm (level 1 [likely reliable] evidence)
  
  • based on 2 prospective cohort studies
  • 2,600 participants ≥ 50 years old had colon cancer screening with multidetector CT colonography and same-day colonoscopy
  • 2,531 participants (97%) with complete colonographic and colonoscopic results were analyzed
  • high compliance rates with protocol
    
    • 2,482 (98%) took barium sulfate for fecal tagging
    • 2,390 (94%) took iodinated contrast for fluid tagging
    • 2,328 (92%) had glucagon given
  • reference standard was colonoscopy and histological confirmation
  • prevalence of lesions
    
    • 5.2% for lesions ≥ 10 mm (132 participants)
      
      • 246 adenoma or carcinoma
      • 146 nonadenomatous
    • 10.2% for lesions 5–9 mm (258 participants)
    • 128 adenoma or carcinoma
    • 27 nonadenomatous
  • per-patient predictive performance of CT colonography for adenoma or carcinoma ≥ 10 mm
    
    • sensitivity 90%
    • specificity 86%
    • positive predictive value 23%
    • negative predictive value 99%
  • per-patient predictive performance of CT colonography for adenoma or carcinoma ≥ 6 mm
    
    • sensitivity 78%
    • specificity 88%
    • positive predictive value 40%
    • negative predictive value 98%
  • per-polyp sensitivity of CT colonography
    
    • 84% for lesion ≥ 10 mm
    • 70% for lesion ≥ 6 mm
  • limitations of CT colonography
    
    • serious adverse events reported in 0.24%
    • lower sensitivity for smaller polyps, but clinical value of detecting smaller polyps uncertain
    • positive colonography requires follow-up colonoscopy
    • extracolonic findings sufficient to trigger additional testing or urgent care in 16%, but few of these findings can be treated effectively
    • effects of cumulative radiation dose if CT colonography repeated every five years unknown
    • Reference: N Engl J Med 2008 Sep 18;359(12):1207, editorial can be found in N Engl J Med 2008 Sep 18;359(12):1285
  • 1,233 asymptomatic adults (mean age 58 years) had same-day virtual and conventional colonoscopy
  • • colonoscopists were blinded to virtual colonoscopy findings on initial exam, final unblinded conventional colonoscopy used as reference standard
    • for adenomatous polyps > 10 mm
      
      • virtual colonoscopy had 93.8% sensitivity and 96% specificity
      • conventional colonoscopy had 87.5% sensitivity, specificity not reported
    • for adenomatous polyps > 8 mm
      
      • virtual colonoscopy had 93.9% sensitivity and 92.2% specificity
      • conventional colonoscopy had 91.5% sensitivity, specificity not reported
    • for adenomatous polyps > 6 mm
      
      • virtual colonoscopy had 88.7% sensitivity and 79.6% specificity
      • conventional colonoscopy had 92.3% sensitivity, specificity not reported
    • 2 polyps were malignant, both were seen on virtual colonoscopy, 1 was missed on initial conventional colonoscopy
    • Reference: N Engl J Med 2003 Dec 4;349(23):2191, editorial can be found in N Engl J Med 2003 Dec 4;349(23):2261, commentary can be found in N Engl J Med 2004 Mar 11;350(11):1148, CMAJ 2004 Apr 27;170(9):1392, Am Fam Physician 2004 May 1;69(9):2197, ACP J Club 2004 Jul-Aug;141(1):22

Low Incidence of Advanced Colorectal Neoplasia Five Years After Negative Screening Colonoscopy

The incidence of advanced colorectal neoplasia is low five years after negative screening colonoscopy (level 2 [mid-level] evidence), based on a retrospective cohort study of 1,256 patients with no colorectal adenomas at time of first screening. Rescreening occurred at mean 5.34 years later. The prevalence of adenoma at rescreening was 16% for ≥ 1 adenoma and 1.3% for advanced adenoma. No cancerous lesions were found at rescreening (N Engl J Med 2008 Sep 18;359(12):1218). For more information, see the colonoscopy topic in DynaMed.

In DynaMed Colonoscopy Surveillance Intervals — General Population:

• incidence of advanced colorectal neoplasia low five years after negative screening colonoscopy (level 2 [mid-level] evidence)
  • based on retrospective cohort study
  • 1,256 patients > 50 years old with no colorectal adenomas at first screening (at mean age 56.7 years, 56.7% men) were rescreened at mean 5.34 years later
  • prevalence of adenoma at rescreening
    • one or more adenoma in 16% (201 participants)
    • advanced adenoma in 1.3% (16 patients, total of 19 adenomas, 10 distal to splenic flexure)
  • no cancerous lesions found at rescreening
  • men significantly more likely that women to have any adenoma (relative risk 1.88) and advanced adenoma (relative risk 3.3)
  • Reference: N Engl J Med 2008 Sep 18;359(12):1218, editorial can be found in N Engl J Med 2008 Sep 18;359(12):1285
  • DynaMed commentary — by coincidence, mean age was 56.7 years and proportion of men was 56.7%