An innovative evidence-based laboratory medicine (EBLM) test to help doctors in the basic assessment of colorectal cancer.

Abstract

e15715 Background: According to the World Health Organization (WHO), colorectal cancer (CRC) is the second leading cause of cancer-related deaths, claiming nearly 1.9 million lives each year. CRC is often diagnosed at an advanced stage, when treatment options are limited. The incidence of CRC can be reduced by implementing primary prevention strategies and practicing early detection. Building upon this imperative, the aim of this study was to assess the estimated diagnostic accuracy of a newly developed biomarker algorithm in confirming or ruling out CRC. Methods: This novel diagnostic model combines several independent public algorithms to provide comprehensive diagnostic insights, thus enabling the confirmation of CRC diagnosis. To evaluate the estimated accuracy of this new test, we conducted an extensive literature review to identify studies assessing the diagnostic accuracy of constituent algorithms, calculations, and combinations of analytes included within it, on the basis of a previous work done by Dr. Juan Bayo et al. in 2022. We developed an upgraded version of this panel and created a refined model that relies on the core set of three machine learning algorithms: Multiple Biomarkers Disease Activity Algorithm (MBDAA), Evidence Based Laboratory Medicine Algorithm (EBLMA), and Artificial Intelligence Recursive Algorithm (AIRA). Parallel and serial approximations were also performed to further improve the sensitivity (Se) and the specificity (Sp) of the model. The model is made around the following tumor markers (TM): AFP, CA 19.9, CA 72.4, CA 125, and CEA, as well as the serum calprotectin, which is a new experimental TM. We also added a set of blood and urine analytes into the test, to detect and discard false positive (FP) results due to benign diseases that can increase serum levels of TM, such as liver disease and kidney disease. Thus, meeting the 1994 Barcelona Criteria, proposed by the Spanish Society of Clinical Chemistry (SEQC). To evaluate its accuracy, we performed a clinical trial with a sample size (n) of 221 individuals. Results: We developed a specific model for men and women, that were evaluated separately. Thus, the Se achieved of the model for men was 0.96, while the Sp was 1, the positive predictive value (PPV) was 1, and the negative predictive value was 0.99. And the results of the model for women were Se: 0.95, Sp: 0.97, PPV: 0.86, and NPV: 0.99. Conclusions: This data suggests that the innovative non-invasive blood and urine-based biomarker algorithm holds promise in providing timely and accurate diagnosis of CRC. Moreover, the results have also shown that this test can be very useful for avoiding unnecessary colonoscopies and other invasive techniques to confirm or rule out CRC. These results advocate for further exploration, prompting our intention to conduct a clinical study involving 10,000 participants to validate and enhance our findings and inform clinical practice.