AIMS AND BACKGROUND
Thiopurines are a class of drug commonly used in the treatment of autoimmune disorders (e.g. Crohn’s disease (CD), ulcerative colitis (UC), and rheumatoid arthritis), acute lymphoblastic leukaemia and post operatively in organ transplant patients. A substantial minority of people treated with thiopurines will develop poor outcomes including myelosuppression (6.5%) or liver injury (3%). Genetic factors including TPMT and NUDT15 show modulating effects on the risk of developing myleosupresison, while pathology tests including ALT, AST, AP and total bilirubin are related to thiopurine-induced liver injury. This project will use machine learning to develop a predictive marker for poor outcomes from thiopurine treatment.
Develop predictive algorithms to identify people who are likely to develop thiopurine induced liver injury or myelosuppression.
Thiopurines are a class of drug including Azathioprine (AZA) and 6-mercaptopurine (6-MP) which are commonly prescribed in the treatment of acute CD or UC and have been shown to reduce the need for surgical interventions by up to 40%. However, up to 30% of patients administered thiopurines discontinue the drug due to side effects including liver injury and myelosuppression.
Thiopurine induced liver injury is reported in between 3 and 10% of people taking thiopurines for inflammatory bowel disease (IBD). This is typically identified through liver function tests including aspartate transaminase (AS), alanine transaminase (ALT), alkaline phosphatase (AP), or total birubin between 1 and 2 times above the normal range4. However, Bjoernsson identified the ratio ALT to AP as more salient than either in isolation.
Bone marrow is a spongy material embedded within bones which produces most of the blood cells. When the production of various blood cells from the bone marrow is suppressed, it is known as myelosuppression and occurs in approximately 3% of patients.
Diagnostic support systems
Diagnostic support systems use objective, health related data to aid clinicians in diagnosing disease and predicting the outcome of treatment.
This project will leverage a large dataset acquired by Associate Professor Graham Radford-Smith containing real world data on patients who have undergone thiopurine treatment for Inflammatory Bowel Disease at the Royal Brisbane Hospital. We will aim to develop prognostic markers for poor outcomes from thiopurine treatment and will have the potential to impact the clinical decision making process around patients who are at a high risk of failing thiopurine treatment.