Search

Introduction: Interval cancer (IC) is an important quality indicator in colorectal cancer (CRC) screening. Previously, we found that fecal immunochemical test (FIT) ICs are more common in women, older age, right-sided tumors, and advanced stage. Here, we extended our existing stage IV patient cohort with clinicopathological and molecular characteristics, to identify factors associated with FIT-IC. Methods: Logistic regression models were fit to identify variables associated with the odds of having a stage IV FIT-IC. Multivariate models were corrected for gender, age, and location. Results: A total of 292 screen-detected (SD) CRCs and 215 FIT-IC CRCs were included. FIT-IC CRC had 5 fold higher odds to be a neuroendocrine (NET) tumor and 2.5 fold higher odds to have lymphovascular invasion. Interestingly, some variables lost significance upon accounting for location. Thus, tumor location is a critical covariate that should always be included when evaluating factors related to FIT-IC. Conclusions: We identified NETs and lymphovascular invasion as factors associated with increased odds of having a stage IV FIT-IC. Moreover, we highlight the importance of tumor location as a covariate in evaluating FIT-IC related factors. More research across all stages is needed to clarify how these insights might help to optimize the Flemish CRC screening program. [ABSTRACT FROM AUTHOR]

DNA methylation biomarkers have emerged as promising tools for cancer detection. Common methylation patterns across tumor types allow multi‐cancer detection. Droplet digital PCR (ddPCR) has gained considerable attention for methylation detection. However, multi‐cancer detection using multiple targets in ddPCR has never been performed before. Therefore, we developed a multiplex ddPCR assay for multi‐cancer detection. Based on previous data analyses using The Cancer Genome Atlas (TCGA), we selected differentially methylated targets for eight frequent tumor types (lung, breast, colorectal, prostate, pancreatic, head and neck, liver, and esophageal cancer). Three targets were validated using ddPCR in 103 tumor and 109 normal adjacent fresh frozen samples. Two distinct ddPCR assays were successfully developed. Output data from both assays is combined to obtain a read‐out from the three targets together. Our overall ddPCR assay has a cross‐validated area under the curve (cvAUC) of 0.948. Performance between distinct cancer types varies, with sensitivities ranging from 53.8% to 100% and specificities ranging from 80% to 100%. Compared to previously published single‐target parameters, we show that combining targets can drastically increase sensitivity and specificity, while lowering DNA input. In conclusion, we are the first to report a multi‐cancer methylation ddPCR assay, which allows for highly accurate tumor predictions. [ABSTRACT FROM AUTHOR]