Perfusion and dopamine transporter availability in parkinsonian disorders - machine-learning-based differential diagnosis using positron emission tomography

Parkinson’s disease and related disorders are the second most common neurodegenerative diseases. Parkinsonian disorders present with similar symptoms at their onset. The diagnosis, which is based on clinical criteria and follow-up, is challenging.

An uncertain or wrong diagnosis may lead to inappropriate treatment and unwanted side effects for the individual patient and  increased costs for the healthcare system.

Early, precise diagnosis is crucial for the development of treatments targeting molecular mechanisms of underlying disease.The aim of the present work is to identify characteristic patterns of cerebral blood flow (CBF) and their association with dopamine transporter (DAT) availability in parkinsonian disorders.

This will be used to develop methods for precise differential diagnosis with PE2I positron emission tomography (PET) and supervised machine learning (ML). 18F-FE-PE2I PET/CT scans providing both CBF and DAT availability from a single scan will be acquired in a cohort of at least 2500 patients from five Swedish university hospitals and will, together with comprehensive clinical diagnosis, be used to train ML-based classification.The project will result in a more precise method for differential diagnosis of parkinsonian disorders that will be available beyond specialized academic centres.

It will further increase understanding of parkinsonism, support appropriate treatment in individual patients, and provide improved imaging biomarkers for drug development.