Imaging of peritoneal metastasis : evaluation of diagnostic performance of DWI-MRI and FDG-PET/CT, correlation of the functional indices and feasibility study based on amide proton transfer MRI

Introduction Peritoneal metastasis is advanced disease and is usually widely disseminated at the time of discovery. It is crucial to detect peritoneal metastasis at an early stage and to allow precise patient selection for the right treatments. Both fluorodeoxyglucose positron emission tomography/computed tomography (FDGPET/CT) and magnetic resonance imaging (MRI) are used in peritoneal metastasis detection. Standardized uptake value (SUV), derived from FDGPET/ CT can evaluate glucose metabolism in cells, whilst apparent diffusion coefficient (ADC) derived from diffusion-weighted MRI (DWI) is used for quantitative analysis of tumour cellularity. Amide proton transfer (APT) MRI is a novel imaging technique based on exchangeable amide protons as endogenous contrast agent and can measure the increased amide proton signal in malignant tissues indirectly through bulk water proton signals. This thesis aims to: 1) compare the diagnostic performance of FDG-PET/CT and MRI in peritoneal metastasis evaluation; 2) study the relationship between SUV and ADC in peritoneal metastasis; 3) investigate the feasibility of APT in peritoneal metastasis evaluation. Materials and Methods Patients with peritoneal metastasis detected on FDG-PET/CT were prospectively recruited for MRI. FDG-PET/CT and MRI images were reviewed independently by two radiologists in separate sessions. Diagnostic characteristics were calculated for both imaging techniques. SUVmax and SUVmean were obtained by placing ROIs on PET, while ADCmin and ADCmean were calculated by contouring lesions on DWI. APT MRI was performed using a single-slice turbo spin echo sequence, with a block of presaturation pulses at 33 frequency offsets. ROIs were placed on peritoneal metastasis and muscle on APT. Mean of integrated asymmetrical magnetic transfer ratio (MTRasym) over 3-4 ppm with respect to water resonance was calculated for peritoneal metastasis and muscle. Results Eight patients were recruited in this study with 128 anatomical sites evaluated. DWI/MRI had good diagnostic performance (sensitivity = 92%, specificity = 99%, accuracy = 98%) compared to that of FDG-PET/CT (sensitivity = 90%, specificity= 100%, accuracy = 98%). Thirty-four peritoneal metastases were selected for quantitative analysis. Significant inverse correlation was found between ADCmean and SUVmax (r = -0.528, p = 0.001) and between ADCmean and SUVmean (r = -0.548, p = 0.001). ADCmin was significantly and negatively correlated with SUVmax (r = -0.508, p = 0.002) and SUVmean (r = -0.513, p = 0.002). In the above study cohort, 6 patients underwent APT imaging with 8 peritoneal metastases evaluated. Seven lesions showed positive APT signal and one had negative APT signal. The mean APT signal for peritoneal metastasis was 2.28%±1.76%, significantly different from that of muscle (-2.79%±0.95%, p < 0.001). Conclusions In conclusion, both DWI/MRI and FDG-PET/CT had good diagnostic performance in peritoneal metastasis evaluation. The negative correlation between SUV and ADC suggested an inverse relationship between tissue metabolism and cellularity. APT MRI is feasible to generate sufficient contrast signal for peritoneal metastasis and has potential to discriminate peritoneal tumours from its surrounding soft tissue using integrated MTRasym as a quantitative marker. These functional indices allow understanding of the biological behaviours of peritoneal tumours and could act as adjuncts in peritoneal metastasis imaging.published_or_final_versionDiagnostic RadiologyMasterMaster of Philosoph