Towards implementation of molecular biomarkers for personalized treatment of colorectal cancer

Abstract

This thesis explores the identification and validation of prognostic and predictive biomarkers to inform adjuvant treatment decisions in non-metastatic colorectal cancer (CRC). The first part investigates circulating tumor DNA (ctDNA) as a promising prognostic biomarker close to clinical implementation. The second part examines genomic structural variants (SVs) as candidate predictive biomarkers, still at an early stage of implementation. The path from biomarker discovery to implementation in routine diagnostics is complex, and the steps are not well defined. Chapter 2 reviews the current status of ctDNA testing for two cancer types and four clinical applications, and explores potential implementation scenarios within five years. Experts anticipate gradual clinical adoption of ctDNA testing, and proving clinical utility is defined as both a main facilitator and a major challenge in this process. The implementation challenges need to be addressed per disease type and application, and requires collaboration among the different stakeholders to ensure timely access for patients. Chapters 3 and 4 focus on the potential of ctDNA-based minimal residual disease (MRD) detection to guide adjuvant treatment decisions in stage II and III colon cancer. Chapter 3 presents the protocol of “MEDOCC-CrEATE”, an interventional study evaluating clinical utility of post-surgery ctDNA-based MRD detection to escalate adjuvant chemotherapy (ACT) in stage II colon cancer, currently recruiting. The primary endpoint is to assess the proportion of ctDNA-positive patients who opt for ACT, and the secondary endpoint is to evaluate 2-year recurrence rates in ctDNA-guided vs non-guided ACT. Chapter 4 focuses on stage III colon cancer, where the PLCRC-PROVENC3 observational study evaluates the clinical validity of post-surgery ctDNA detection for risk stratification and treatment (de)escalation. Using a whole genome sequencing (WGS) assay, the study found that post-surgery ctDNA positivity has the highest prognostic value among evaluated factors. CtDNA status complements clinicopathological based risk assessment, and allows to define low-risk ctDNA-negative patients as a group who could avoid unnecessary ACT. The results are being evaluated in an early cost-effectiveness study, to identify knowledge gaps and inform future study designs. The second part of the thesis investigates genomic SVs as predictive biomarkers in CRC. SVs are prevalent in CRC tumors, yet their biological and clinical relevance remains poorly understood due to technical limitations. In Chapter 5, a method for SV detection optimized for formalin-fixed paraffin-embedded (FFPE) tissue using targeted locus capture (FFPE-TLC) was validated. Four genes located in common fragile sites (CFS) of the genome and three source LINE regions were identified both in FFPE tissue and in matched plasma, proving their applicability as highly prevalent markers for ctDNA detection. Chapter 6 focuses on MACROD2, a gene located in a CFS and frequently affected by SVs in CRC. In silico and in vitro analyses revealed that focal deletions in MACROD2 affecting exons 5-6 were associated with increased genomic instability and impaired DNA damage repair and mono-ADP-ribose levels, potentially playing a role in carcinoma progression. This research establishes a foundation for future studies to explore the role of SVs as predictive biomarkers in CRC.