Cancer Patients Digital Twins to Investigating disease fragmentation and its impact on drug response in AML trials

点击翻译按钮获取中文摘要

As not all patients respond equally to the same treatments, there may be clues in a patient's genome that can predict how they will respond to different treatments and whether there are common genomic signatures that influence response to chemotherapy. In particular, Cancer Patients Digital Twins (CPDTs) can assist in the digital or virtual representation of a patient, creating a replica of an individual in combination with AI algorithms that take into account multiple factors to aid prediction or diagnosis.This project will use digital twins of cancer patients to investigate disease fragmentation and its impact on drug response in AML trials. Using genomic data available from a number of AML trials for around 2500 patients, including their treatment and survival data, the project will focus on:i. Use machine learning approaches to identify disease fragmentation and patient response to specific drugs. ii. Look for genetic explanations for differences in drug response that are common to different treatment pathways; and iii. Develop models that use genomic data to predict drug response within standard AML treatment pathways, providing ways to tailor interventions for maximum impact. iv. Develop and design Cancer Patient Digital Twins (CPDTs) to help predict drug response and AML treatment. By taking a human-centred AI approach this project aims to explore and unpack the most appropriate data of AML trials to design and deploy a Cancer Patients Digital Twins (CPDTs) to predict drug responses and AML treatment pathways. Through this project we will identify: i. What are the key features, models and elements to better represent and model a digital cancer patient twin for disease fragmentation and drug response in AML. ii. What patient's genome data might be more relevant to predict how AML patients will respond to different treatments. iii. What are the most common genomic signatures that will influence response to chemotherapy. iv. Follow a human-centred AI approach to unpack the data streams, data interactions alongside with technological interactions to design Cancer Patients Digital Twins (CPDTs) for AML With the NHS CYSGODI (CYmru Service for Genomic Oncology Diagnoses) service generating genomic data for various haematological cancers, this work has a direct route to impact on routinely generated datasets.