Dr Neil Basu, Senior Clinical Lecturer of Rheumatology, University of Glasgow
Disorders such as fibromyalgia and chronic fatigue syndrome represent some of the most clinically challenging conditions in Medicine. Their excess biological and phenotypic heterogeneity is a primary driver of this. A situation which is further conflated by their prevalent co-existence with other, mechanistically distinct, disorders. Numerous initiatives have sought to characterise and ultimately classify these syndromes with varying degrees of success and controversy. These have almost entirely been ‘top-down’ in approach i.e. derived from either expert consensus or phenotypic data. With historically limited knowledge of pathogenesis, it has been challenging to adopt a biologically based ‘bottom-up’ approach to classification. However, advances in brain imaging, which at long last have begun to deliver mechanistic insights into these syndromes, offer a significant opportunity. The possibility to classify patients into homogenous subgroups will greatly support aetiological research - until now true signals have likely been masked by methodological artefacts generated from studying centralised syndromes as single entities rather than focusing on potentially mechanistically distinct subtypes. Clinically, the ability to firstly parse out co-existing centralised features in peripherally dominated chronic diseases will support the judicious use of existing therapeutics. In the future, mechanism based sub-classification will help triage patients towards optimal interventions, in keeping with the ideals of personalised medicine. In order to meet such ambitions, close collaborations with data scientists will be essential. Evidence is accruing that machine learning methods can successfully integrate rich MRI data streams in order to answer clinically relevant questions. Ultimate implementation of such algorithms into health care services has yet to be realised but is certainly feasible.