Multi-Target Stool DNA Testing: High Positive Predictive Value Among Average Risk Patients New to Colorectal Cancer Screening

Abstract

Introduction: Multi-target stool DNA (MT-sDNA) testing was approved by the FDA for average-risk colorectal cancer (CRC) screening in 2014, and its use has increased exponentially. MT-sDNA performance was demonstrated in cross-sectional screen-setting studies, and post-approval clinical-use data are now emerging. We aimed to measure compliance to diagnostic colonoscopy and positive predictive value of MT-sDNA in average risk patients, with and without prior screening by colonoscopy. Methods: In a large multi-setting practice, we identified all patients with positive MT-sDNA test results between 10/1/2014 and 12/31/2016. From the medical record, we abstracted patients’ exposure to prior CRC screening by colonoscopy, date of colonoscopy, personal and family history of colorectal neoplasia (CRN), and presence or absence of high risk conditions. Those at increased CRN risk were excluded. Diagnostic colonoscopy and pathology reports were reviewed to enumerate findings of any CRN and advanced CRN. Positive predictive values were calculated for each endpoint and stratified by prior screening colonoscopy exposure. Results: Among the 1035 patients identified with a positive MTsDNA test, 926 (89%) met average risk screening criteria. Of the 926, 410 (44%) did not have prior screening colonoscopy; 238/410 (58%) were >60 years old. Of the 516/926 (56%) that had previous colonoscopy, median time between colonoscopy and MT-sDNA testing was 10.3 (IQR 9.3-11.3) years. Of the 926 with positive MT-sDNA, 821 (89%) had a documented subsequent diagnostic colonoscopy. Diagnostic colonoscopy was completed at a median of 42 (IQR, 25-66) days after positive MT-sDNA. Of these 821, 569 (69%) had CRN identified by colonoscopy, with 247 (30%) of patients harboring advanced CRN; CRC was found in 7 (1%). In previously screened patients (n=473), CRN was identified on diagnostic colonoscopy in 310 (positive predictive value [PPV]=66%)(Table). Among patients without prior screening (n=348), yield was higher with 259 found to have CRN (PPV=74%, p=0.007). Advanced CRN was found in 26% of patients with and 36% of those without prior screening colonoscopy, respectively (p=0.0001). Conclusion: Compliance to diagnostic colonoscopy after a positive MT-sDNA test was high. CRN was identified in more than 2 out of every 3 patients with positive MT-sDNA. Yield of diagnostic colonoscopy was higher in the 44% of patients with positive MT-sDNA who had not received prior CRC screening by colonoscopy.289 Figure 1. Patient characteristics and neoplasms at diagnostic colonoscopy after positive multi-target stool DNA test, stratified by prior screening