In-vivo and ex-vivo detection of colorectal cancer at ultra-low field using fast field-cycling methods

Abstract

Colorectal cancer (CRC) remains a major global health burden, with over 1.9 million new cases and approximately 904,000 deaths reported worldwide in 2022. CRC includes cancers of both the colon and rectum, with rectal cancer affecting the last 12-15 cm of the large intestine. Staging this disease relies heavily on CT and MRI of the pelvis. To date, these imaging modalities provide limited information on the microstructural and biological changes associated with tumour progression or treatment response. This thesis aims to investigate the potential of field-cycling magnetic resonance methods, specifically Fast Field-Cycling Nuclear Magnetic Resonance (FFC-NMR) and Field-Cycling Imaging (FCI), as innovative approaches for the detection and characterisation of CRC and LARC. Specifically, the work explores whether intrinsic tissue relaxation properties can differentiate tumours from surrounding tissues and provide early indicators of therapeutic response. To address these questions, two pilot studies were conducted. The ex-vivo study analysed fresh CRC tissues using FFC-NMR relaxometry to assess ¹H relaxation dispersion profiles in tumour, peritumour, and healthy samples. To support interpretation of the signal differences, proteomics analysis was performed on the same samples to explore protein expression contributions to the observed relaxation behaviour. In the in-vivo study, the whole body FCI scanner was used to scan a cohort of LARC patients, acquiring pre- and post-treatment scans to evaluate whether changes in relaxation biomarkers could reflect response to neoadjuvant chemoradiotherapy (nCRT). Findings from the ex-vivo study indicated noticeable differences in the dispersion profiles between tissue types, and proteomic analysis identified different pathways potentially linked to these contrasts, including enrichment in ATP production, protein synthesis, and ribosomal activity in tumour tissues. In the in-vivo study, the FCI imaging protocol was successfully implemented, and preliminary data showed significantly higher R1 values in tumour regions compared to healthy tissue at 2 mT and 0.2 mT (p < 0.0001), with slope and intercept biomarkers also differing significantly (p < 0.005), with changes observed after treatment in patients showing complete or partial response. In conclusion, this thesis provides preliminary evidence supporting the feasibility of FFC-NMR and FCI as non-invasive tools for CRC and LARC assessment. While findings are promising, further research with larger cohorts is essential to validate these findings, improve statistical power and enable patient stratification to better capture tumour heterogeneity and treatment response.