Assessment of clinical discriminatory performance and related uncertainty of colorectal cancer pathological staging

Background

Our focus is on colorectal cancer (CRC), the second most common cause of cancer death accounting for > 16,000 deaths in UK. This disorder is of high prevalence, high risk to individuals and high cost. CRC develops slowly and the risk of recurrence and death is related to the stage of the disease at diagnosis. Whilst pathological staging stratifies prognostic groups, it does not perform well to categorise poor/good prognosis tumours at the individual level.

Thirty-five low penetrance risk single nucleotide polymorphisms (SNPs) have been linked to CRC risk in a number of Genome Wide Association Studies1. However, the genetic predisposition to CRC prognosis has not been extensively investigated2,3. It is anticipated that GWAS with sufficient sample size and genetic coverage may lead to novel insights into common inherited genetic variants. These genetic variants along with blood biomarkers proved to influence CRC prognosis4 could potentially help decide individual disease management.

We propose to combine data from the UK Biobank and Scottish Colorectal Cancer Study (SOCCS) to investigate the association of genotype with all-cause and CRC-specific mortality and to address the clinical variation within pathological staging groups with respect to prognostic categorization and chemotherapy administration at individual level.

Aims

Aim 1: To fit robust logistic-type statistical models to assess the association of genotype with all-cause and CRC-specific mortality, adjusting for age, gender and stage of cancer and accounting for the tumour MSI status. We will include 6,700 CRC patients with GWAS data: 1,000 incident CRC SOCCS cases, 2,700 incident UK Biobank CRC cases and 3,000 prevalent UK Biobank CRC cases (taking into account the bias of underrepresentation of patients with aggressive and/or metastatic disease). The top most statistically significant loci will then be genotyped in 3,000 additional CRC SOCCS cases for replication analysis.

Aim 2: The growing repertoire of available treatments, including new chemotherapy approaches, combined with challenging benefit:toxicity ratios and cost, means that it is crucial to target interventions to patients most likely to benefit. In particular patients with stage 2 CRC may be offered adjuvant chemotherapy if their cancer is classified as high risk and similarly patients with stage 3 disease may only need surgery if they have a low risk tumour. Furthermore any prognosis genetic predisposition may indicate druggable targets in the tumours arising in such individuals, enabling targeted chemotherapy. Hence, improving the clinical discriminatory performance of pathological staging offers much potential for clinical benefit.

We aim to appraise the clinical usefulness of identified genetic, socio-demographic and biomarker risk factors to discriminate high and low risk tumours within a given stage and help with the decision of chemotherapy administration given the side effects and cost implications. We will assess various statistical measures related to predictive ability and prognostic accuracy (e.g. discrimination, calibration, reclassification, global measures of model fit) using methodology relying on posterior predictive checking, under a Bayesian statistical approach that allows for the uncertainty associated with predictions to be quantified. The application of such methodology in the context of this project is novel and at the forefront of current practice.

Experience in epidemiology and statistical methods and strong IT skills will be highly beneficial.

Training outcomes

The student will be trained in the core disciplines of informatics, data science, mathematics and statistics placed in the CRC context and will develop skills related to novel analytical and quantitative methods underpinning modern data-driven predictive modelling. They will also get further cross-disciplinary biomedical science training to determine knowledge value as appropriate.

This MRC programme is joint between the Universities of Edinburgh and Glasgow. You will be registered at the host institution of the primary supervisor detailed in your project selection.

All applications should be made via the University of Edinburgh, irrespective of project location:

http://www.ed.ac.uk/studying/postgraduate/degrees/index.php?r=site/view&id=919

Please note you must apply to one of the projects and you are encouraged to contact the primary supervisor prior to making your application. Additional information on the application process if available from the link above.

For more information about Precision Medicine visit:

http://www.ed.ac.uk/usher/precision-medicine