Multitarget Stool RNA Testing

To the Editor The study by Dr Barnell and colleagues raises methodological and clinical concerns. First, sessile serrated lesions (SSLs) 1 cm or larger are widely considered to be advanced precancerous lesions comparable to advanced adenomas, and their categorization as a nonfinding and part of the specificity calculation is unconventional. From the study’s Table 2, moving these lesions to the sensitivity calculation reduces sensitivity for advanced precancerous lesions from 46% (278/606) to 39% (311/798) for multitarget stool RNA (mt-sRNA) testing and from 29% (175/606) to 23% (187/792) for fecal immunochemical testing (FIT). Can the authors provide an explanation for their classification and confirm our calculations using the more conventional definition of advanced precancerous lesions? Second, at equal colorectal cancer (CRC) specificity in Figure 2A, it appears that FIT is more sensitive than mt-sRNA testing. Full receiver operating characteristic curves for FIT and mt-sRNA would help clarify this issue. Third, eFigure 3 in Supplement 3 shows area under the curve (AUC) for the RNA component of the mt-sRNA test alone, with CRC and advanced adenoma both compared against all other findings having respective AUCs of 0.62 and 0.58. What were the respective AUCs for the combination of FIT and smoking status? Could smoking status or FIT account for the higher sensitivity of the mt-sRNA test for distal advanced adenomas vs proximal ones? And what was the incremental discrimination of the RNA component to FIT and smoking status? Lastly, the 100% CRC sensitivity for those aged 45 to 49 years is based on 5 cancer diagnoses and has a 95% CI of 48% to 100%, suggesting that caution should be taken when interpreting sensitivity of the mt-sRNA test for this age group. Clarification of these concerns is needed to provide comparison and interpretation of these new data for screening asymptomatic average-risk persons.