Keaan Amin

• MBBS
• Can Machine Learning Identify Patients with iAMP21, a High Risk of Tumour Relapse Subtype of Childhood Acute Lymphoblastic Leukaemia?

Acute Lymphoblastic Leukaemia (ALL) is the most common cancer in children. iAMP21 is a high-risk subgroup of B-cell ALL where patients are three times more at risk of a tumour relapse. Subsequently, reliable detection of this subgroup can enable the intensification of chemotherapy and future targeted treatments.

This project utilised machine learning, a form of artificial intelligence, capable of detecting subtle patterns to classify patients with iAMP21 from non-iAMP21 B-cell ALL patients. We used machine learning on gene expression-based data from 59 bone marrow patient samples to separate our iAMP21 patients. Through gene expression visualisation, gene selection and cross-validation, we identified a seven gene Chromosome 21 group that could detect iAMP21 patients with a perfect level of accuracy and specificity. This supports the role of machine learning and gene-expression based markers as a supportive diagnostic tool in clinical decision-making processes.

Funding source: Newcastle University

Project supervisor: Dr Amir Enshaei