EP592 - Outside the Box: Assessing Colonoscopy Results of Patients Presenting with Positive Multitargeted Stool DNA Tests

Purpose/Background: Colorectal cancer (CRC) is the third most prevalent cancer overall in the United States, and is the second leading cause of cancer mortality, behind only lung cancer. Recently there has been a shift in diagnosis of cancer to younger populations, with the proportion of cases diagnosed in individuals younger than 55 rising to 20% in 2019, from 11% in 1995. Early detection is paramount, as 5-year survival in patients with localized disease is 91%, compared to 14% in those with distant disease. Colonoscopy has been shown to reduce incidence of CRC by 40% and mortality by 60%.



While colonoscopy and biopsy is the gold standard for colorectal cancer screening, multitargeted stool DNA testing (mt-sDNA) has emerged as an alternative. This is primarily due to its sensitivity for detecting CRC, which varies widely but found to be as high as 92%. Given the high prevalence of (mt-sDNA) seen in our practice and significant anxiety to the patient that a positive test result can cause, we sought to determine the effectiveness of the test at predicting the ultimate biopsy-proven result.

Methods/Interventions: We performed a retrospective chart review of 350 patients at our multi-institution colorectal practice who presented for colonoscopy after receiving a positive mt-sDNA test between 2024 and 2025. Primary outcomes included gross colonoscopy findings and histopathologic findings. Colonoscopy results were classified as either (1) no findings or hyperplastic polyps only or (2) adenoma or carcinoma. Locations were defined as right (cecum, ascending colon, hepatic flexure, transverse colon) or left (splenic flexure, descending colon, sigmoid, rectum) -sided lesions.

Results/Outcomes: Of the 350 patients who presented to our colonoscopy suites with positive mt-sDNA tests, only 3 patients were eventually diagnosed with CRC. Positive predictive value for adenoma or cancer was 46.3%. Of patients who were positive for adenoma or carcinoma, 45 (27.8%) had a right-sided lesions, 53 (32.1%) had left-sided lesions, and 65 (40.1%) had lesions on both sides. Multivariate logistic regression revealed that right-sided and both-sided lesions had roughly 8x and 35x higher odds, respectively, of a true positive mt-sDNA test than left-sided lesions (p < 0.001). There were no significant differences in PPV when comparing age or gender.

Conclusion/Discussion: Our results show that an adequate non-invasive alternative to colonoscopy is still yet to be found. In a recent study of over 350,000 screening colonoscopies, polyp detection rate was demonstrated to be 25%. That in our study a positive mt-sDNA test correlates with polyps only 46% of the time raises the question of the utility of using it as a screening tool. This study also demonstrates the pitfalls of mt-sDNA as a predictive tool; while there may be a role in ruling out CRC, it has little use in determining what lesion a patient is likely to have. For those who opt for mt-sDNA, our results demonstrate that these patients need follow up colonoscopy and should be assured that a positive test has a very low likelihood of being cancer and is much more likely to be associated with benign or premalignant lesions.