Do the benefits outweigh the side effects of colorectal cancer surveillance? A systematic review

Table 3 Probability of false positive test results (1-specificity) for commonly used colorectal cancer follow-up tests

Test	False positive rate (1-specificity)	Ref.
Serum CEA	10%	[54]
CT-hepatic metastases	5%-28%1	[55-58]
CT-other abdominal metastases	2%	[58]
Contrast enhanced ultrasound-liver	4%-33%2	[56,57,59]
Ultrasound-liver	50%	[59]
CT-lungs	4%	[58]
Colonoscopy	0%	[32]

¹Based on specificity estimates from individual studies of 89%[55] (n = 24), 95%[58] (n = 115), 72%[56] (n = 87), and 91%[57] (n = 100);

²Based on specificity estimates from individual studies of 96%[60] (n = 68), 96%[57] (n = 99), and 67%[59] (n = 56) subjects. The last was the only to employ intraoperative confirmation of hepatic metastases. The annual probability of at least one false positive test for a patient with no actual recurrence would be 41% in each of year one and two, and 28% in each of year three, four, and five. Over the entire five-year period, the probability of at least one false positive would be 87%. CT: Computed tomography; CEA: Carcinoembryonic antigen.