Multiple Myeloma (MM) remains an incurable plasma cell malignancy despite improvements in treatment over the last two decades. There are 250 patients newly diagnosed with MM in Ireland annually. Patients undergo a CT Skeletal Survey to assess for the presence of bony lesions associated with MM. Each CT Skeletal Survey generates over a thousand images which are assessed by a Consultant Radiologist to identify bony lesions. Identifying these lesions leads to commencement of treatment to prevent further disease complications and improve overall survival.
There is evidence that identifying patients at high risk of progression to Multiple Myeloma and commencing treatment without delay improves the overall survival of these patients.
Machine learning is a branch of artificial intelligence that provides the ability to a computer programme to automatically learn from experience and apply this learning to detect patterns efficiently. In this project machine learning will be used to teach a computer programme based at the University of Limerick to detect focal bony lesions diagnostic of Multiple Myeloma requiring urgent treatment, and we hypothesise that the diagnostic process will be improved for patients at our centre in terms of faster reporting of results and identification of subtle bony lesions that are hidden within thousands of CT images but detected automatically by the machine learner.