Using functional imaging to improve chemoradiotherapy outcomes in pancreatic and rectal cancer: moving towards a hypoxia targeting approach

Biologically individualised, adaptive radiotherapy requires the integration of information obtained from molecular biomarkers with functional imaging to inform high-precision radiotherapy planning. Functional imaging allows the non-invasive assessment of heterogeneity in a number of tumour physiological characteristics, including the well-documented cause of chemoradioresistance, hypoxia. Functional imaging may be useful in treatment strategy selection and biological target volume (BTV) definition for selective radiotherapy dose escalation. Studies in a number of tumour sites suggest that the response in functional imaging parameters offers more predictive and prognostic information than single time-point assessments, including heterogeneity in tumour reoxygenation in head and neck cancer. In this thesis, the role of functional imaging was investigated in two tumour types. Data from the ARC-II clinical trial in locally advanced pancreatic cancer (LAPC) was used and a new trial in rectal cancer (RHYTHM) was developed and opened. Chemoradiotherapy (CRT) outcomes in LAPC are poor, with a median overall survival of around 15 months. The benefit of definitive CRT over chemotherapy alone in LAPC is controversial despite 30% of patients with pancreatic cancer only ever having locally progressive disease. The utility of pre- and post-chemoradiotherapy 18F-fluorodeoxyglucose (FDG) positron emission tomography (PET)-derived parameters in treatment strategy selection and patient prognosis was assessed. The limitations of using FDG-PET for BTV definition were explored. A method of characterising hypoxia using 18F-fluoromisonidazole (FMISO) PET was developed that demonstrated the hypoxic BTV to remain spatially consistent with time. The ability of nelfinavir to modulate hypoxia and perfusion was then assessed using dynamic FMISO-PET and perfusion computed tomography (pCT). While approximately 75% of patients with locally advanced rectal cancer can be successfully cured, the current treatment paradigm includes CRT followed by total mesorectal excision (TME) with its associated long-term morbidity. If pathological complete response (pCR) could be accurately identified on post-CRT imaging and if rates of pCR could be improved upon, organ-preserving strategies may be adopted. Current anatomical imaging techniques recognise poor treatment responses well, but under-report pCR. The ability of a CT texture feature approach to pCR prediction was investigated. Imaging, tissue and circulating biomarkers of hypoxia (FMISO-PET, pimonidazole and carbonic anhydrase IX [CAIX] immunohistochemistry of rectal cancer tissue and osteopontin) and perfusion (pCT, dynamic contrast-enhanced magnetic resonance imaging [dce-MRI], CD31 immunohistochemistry) as well as RNA sequencing of rectal cancer tissue were assessed in a clinical study to identify the best method of characterising the modulation of hypoxia and perfusion in rectal cancer during fractionated CRT. Early analyses of trial data will be discussed. Trial recruitment is ongoing.