To inform the 2016 United States Preventive Services Task Force (USPSTF) colorectal cancer screening recommendations, the National Cancer Institute’s (NCI’s) Cancer Intervention and Surveillance Modeling Network (CISNET) performed a decision analysis to model the benefits, harms, and burden of various colorectal cancer screening strategies. As part of this analysis, the CISNET used screening test characteristics reported in a USPSTF commissioned systematic review conducted by Lin et al. The test characteristic values used are summarized in the table adapted from the CISNET technical report
|Base-Case Value||Source||Worst-Case Value||Best-Case Value||Source|
|gFOBT (per person)||Zauber, 2008|
|Specificity||0.925||Not varied||Not varied||Not applicable|
|Sensitivity for adenomas ≤5 mm||0.075*||0.075||0.075||Zauber, 2008|
|Sensitivity for adenomas 6-9 mm||0.124||0.1||0.262||Zauber, 2008|
|Sensitivity for adenomas ≥10 mm||0.239||0.177||0.494||Zauber, 2008|
|Sensitivity for colorectal cancer||0.7||0.615||0.794||Levi, 2011
|FIT (per person)||Imperiale, 2014||Imperiale, 2014|
|Specificity||0.964||Not varied||Not varied|
|Sensitivity for adenomas ≤5 mm||0.076†||0.067||0.086|
|Sensitivity for adenomas 6-9 mm||0.067||0.086|
|Sensitivity for adenomas ≥10 mm||0.238‡||0.208||0.27|
|Sensitivity for colorectal cancer||0.738||0.623||0.833|
|FIT-DNA (per person)||Imperiale, 2014||Imperiale, 2014|
|Specificity||0.898||Not varied||Not varied|
|Sensitivity for adenomas ≤5 mm||0.172†||0.159||0.186|
|Sensitivity for adenomas 6-9 mm||0.159||0.186|
|Sensitivity for adenomas ≥10 mm||0.424‡||0.387||0.462|
|Sensitivity for colorectal cancer||0.923||0.84||0.97|
(within reach, per lesion)§
|Specificity||0.86?||Schroy, 2013||Not varied||Not varied||Not applicable|
|Sensitivity for adenomas ≤5 mm||0.75||van Rijn, 2006||0.7||0.79||Zauber, 2008|
|Sensitivity for adenomas 6-9 mm||0.85||van Rijn, 2006||0.8||0.92||Zauber, 2008|
|Sensitivity for adenomas ≥10 mm||0.95||van Rijn, 2006||0.931||0.995||Johnson, 2008|
|Sensitivity for colorectal cancer||0.95||By assumption||0.931||0.995||By assumption|
|Sigmoidoscopy (within reach, per lesion)||By assumption|
|Specificity||0.87?||Weissfeld, 2005||Not varied||Not varied|
|Sensitivity for adenomas ≤5 mm||0.75||By assumption||0.7||0.79|
|Sensitivity for adenomas 6-9 mm||0.85||By assumption||0.8||0.92|
|Sensitivity for adenomas ≥10 mm||0.95||By assumption||0.931||0.995|
|Sensitivity for colorectal cancer||0.95||By assumption||0.931||0.995|
|CT colonography (per lesion)||Johnson, 2008||Johnson, 2008|
|Specificity||0.88^||Not varied||Not varied|
|Sensitivity for adenomas ≤5 mm||--||--||--|
|Sensitivity for adenomas 6-9 mm||0.57||0.489||0.716|
|Sensitivity for adenomas ≥10 mm||0.84||0.756||0.924|
|Sensitivity for colorectal cancer||0.84||0.756||0.924|
FIT – fecal immunochemical test; FIT-DNA – multi-target stool DNA test (fecal immunochemical test with a DNA stool test); gFOBT – sensitive guaiac-based fecal occult blood test; -- indicates sensitivity is not provided because adenoma size is smaller than the referral threshold for a colonoscopy of 6mm, that is, only persons with a ≥6mm lesion visualized at CT colonography are deemed to have a positive screening test.
* We assume that 1-5 mm adenomas do not bleed, and therefore cannot cause a positive stool test. We also assume that gFOBT can be positive due to bleeding from other causes, the probability of which is equal to positivity rate in persons without adenomas (i.e. 1 – 0.925).
† Sensitivity for persons with non-advanced adenomas. For persons with 1-5 mm adenomas, we assume that the sensitivity of the test is equal to the positivity rate in persons without adenomas (i.e., 1 – specificity). The sensitivity for persons with 6-9 mm adenomas is chosen such that the weighted average sensitivity for persons with 1-5 mm and with 6-9 mm adenoma(s) is equal to that of non-advanced adenomas.
‡ Sensitivity for persons with advanced adenomas (i.e., adenomas ≥ 10 mm and/or adenomas with advanced histology). Sensitivity was not reported for the subset of ≥ 10mm adenomas.
§ We assume the same test characteristics for screening colonoscopies as for colonoscopies for diagnostic follow-up or for surveillance. We assume no correlation in findings between CTC or sigmoidoscopy and subsequent diagnostic colonoscopy.
? The lack of specificity with endoscopy reflects the detection of non-adenomatous polyps, which, in the case of sigmoidoscopy, may lead to unnecessary diagnostic colonoscopy, and in the case of colonoscopy screening, leads to unnecessary polypectomy, which is associated with an increased risk of colonoscopy complications.
^ The lack of specificity with CTC reflects the detection of ≥ 6 mm non-adenomatous lesions, artifacts, stool, and adenomas smaller than the 6 mm threshold for referral to colonoscopy.
** table adapted from Table 4 in Zauber, Ann, et al. “Evaluating the Benefits and Harms of Colorectal Cancer Screening Strategies: A Collaborative Modeling Approach. ” Agency for Healthcare Research and Quality Publication No. 14-05203-EF-2 October 2015.