PPV and Detection Rate of mt-sDNA Testing, FIT, and CT Colonography for Advanced Neoplasia: A Hierarchic Bayesian Meta-Analysis of the Noninvasive Colorectal Screening Tests.

Abstract

BACKGROUND. Noninvasive tests for colorectal cancer (CRC) screening and prevention limit the need for invasive colonoscopy to follow up positive test results. However, the relative performance characteristics of available noninvasive tests have not yet been adequately compared. OBJECTIVE. We performed a systematic review and meta-analysis to compare the diagnostic performance of the available noninvasive CRC screening tests, including multitarget stool DNA (mt-sDNA) testing, fecal immunochemical testing (FIT), and CT colonography (CTC), with an emphasis on comparison of PPV and detection rate (DR) for advanced neoplasia (AN; encompassing cases of advanced adenomas and CRC). EVIDENCE ACQUISITION. After systematic searches of MEDLINE and Google Scholar databases, 10 mt-sDNA, 27 CTC, and 88 FIT published screening studies involving 25,132, 33,493, and 2,355,958 asymptomatic adults, respectively, were included. Meta-analysis with hierarchic Bayesian modeling was conducted in accordance with Cochrane Collaboration and PRISMA guidelines to determine test positivity rates (TPRs) leading to optical colonoscopy, as well as PPVs and DRs for both AN and CRC. Different positivity thresholds were considered for FIT and CTC. EVIDENCE SYNTHESIS. Point estimates (with 95% credible intervals) from pooled Bayesian meta-analysis combining all thresholds for FIT and stratifying CTC results by a polyp size threshold of 6 mm or larger (CTC6) and 10 mm or larger (CTC10) were calculated. TPR was 13.5% (10.9-16.6%) for mt-sDNA testing, 6.4% (5.8-7.2%) for FIT, 13.4% (11.4-15.6%) for CTC6, and 6.6% (5.2-7.7%) for CTC10. AN PPV was 26.9% (95% credible interval, 21.8-33.2%) for mt-sDNA testing, 31.8% (29.3-34.5%) for FIT, 34.4% (27.2-41.0%) for CTC6, and 61.0% (54.0-70.0%) for CTC10. CRC PPV was 2.4% (1.5-3.9%) for mt-sDNA testing, 4.9% (4.3-5.3%) for FIT, 3.5% (2.5-4.8%) for CTC6, and 6.0% (4.3-8.0%) for CTC10. The DR for AN was 3.4% (95% credible interval, 2.5-4.8%) for mt-SDNA, 2.0% (1.8-2.3%) for FIT, 4.8% (4.0-6.5%) for CTC6, and 4.0% (3.0-4.6%) for CTC10. When FIT is restricted to a lower threshold (< 10 μg Hb/g feces), its performance profile is similar to that of mt-sDNA testing, although available data are limited. AN PPV odds ratios (relative to CTC10 as the reference) were 0.24 (95% credible interval, 0.17-0.33) for mt-sDNA testing, 0.30 (0.24-0.45) for FIT, and 0.33 (0.25-0.47) for CTC6. CONCLUSION. Among noninvasive CRC screening tests, CTC with a polyp size threshold of 10 mm or larger most effectively targets AN, preserving detection while also decreasing unnecessary colonoscopies compared with mt-sDNA testing and FIT. CLINICAL IMPACT. CTC performed with a polyp size threshold for colonoscopy referral set at 10 mm or larger represents the most effective and efficient noninvasive screening test for CRC prevention and detection.