ICR-CTSU Integrative Genomic Analysis in Clinical Trials Group

We are a multidisciplinary group of statistical, computational and translational scientists (in collaboration with the laboratory scientists) in the ICR's Clinical Trials and Statistics Unit (ICR-CTSU) and the Division of Clinical Studies. We develop novel and efficient analytical methods for the application of high-dimensional genomic and proteomic data generated from biospecimens collected in clinical trials in order to:

1. Identify genomic signatures for the selection of patients for specific chemotherapy and biologically targeted agents,
2. Identify biological endpoints for emerging therapeutic targets, and
3. Study the underlying biology of tumours from “exceptional” drug responders in various therapeutic agents across tumour types.

The ICR-CTSU provides a unique and rare opportunity for our translational genomics group to be integral to the trial management and classic statistical analysis environment. Using a systemic approach, we will develop statistical analytic tools for studying the relationship between survival data and high-dimensional genomic data. These include variable selection methods, ensemble learning and cross-data-type prediction to incorporate other data types like DNA copy number, methylation, and proteomics as available into our predictive algorithms, used for stratified medicine.

Integrative analyses of molecular and clinical Data from clinical trials

My research interests are built upon computational analyses of the ever-growing rich datasets of multi-analyte measurements (genomic, genetic, proteomic, metabolites, imaging etc). More specifically, our group will lead the integrated analysis of molecular data with demographic, pathologic and outcome data to identify predictive and prognostic biomarkers for selective therapeutic agents. Using a systemic approach, my research group will develop statistical analytic tools for studying the relationship between survival data and high-dimensional genomic data, such as variable selection methods, ensemble learning and cross-data-type prediction to incorporate other data types like DNA copy number, methylation, and proteomics as available into our predictive algorithms, used for stratified medicine.

Identification of biological endpoints for emerging therapeutic targets

Response-adaptive randomised studies have been advocated as an effective way to allocate subpopulations of patients such that more patients receive better treatments. My other important research priority is to identify intermediate endpoints and develop robust assays for perioperative windows of opportunity and neoadjuvant trials. The development of adaptive trials is an important strategy for the ICR-CTSU. With Professor Judith Bliss, I will concurrently develop clinical trials for rapid evaluation of therapeutic strategies to test lead hypotheses while laying groundwork for future trials.

With Professors Bliss and Mitch Dowsett, we are developing a novel clinical research platform that will embed trials to evaluate the matching of combinations of endocrine plus targeted therapy with biomarkers of specific endocrine resistance pathways (POETIC-2). In collaboration with Professor Andrew Tutt, Dr Sheeba Irshad, and Professor Bliss, we have co-developed the “PHOENIX” trial, a UK platform initiative constituting a post neoadjuvant pre-surgical disease profiling and novel therapy “window of opportunity” biomarker endpoint trial.

Implementation of cognitive computing in clinical trials network

We will be under a deluge of data generated from the research initiatives, including sequencing and gene expression data, protein abundance by mass spectrometry, immunohistochemical images, and clinical outcome.

Leveraging strategic alliance with the clinical trials network, The National Cancer Research Institute Cellular-molecular pathology initiative, and existing collaboration with NanoString® Technologies, my research interest is to collect the genomic and protein expression profiles of tumours (e.g. applying the NanoString® 3-D biology technology) from “exceptional drug responders” to learn the underlying biology mechanisms of these outliers from clinical trials. Our group will assess various applied machine-learning methods for genomics data linked with clinical information. If successful, we may potentially be able to suggest alternative drug options for these “extreme responders”.