Matt Fillingim, Christophe Tanguay-Sabourin, Marc Parisien, Azin Zare, Gianluca V. Guglietti1, Jax Norman, Bogdan Petre, Andrey Bortsov, Mark Ware, Jordi Perez, Mathieu Roy, Luda Diatchenko1 & Etienne Vachon-Presseau
This paper leverages large-scale biomedical datasets to derive, validate, and optimize biomarkers for chronic pain and its associated conditions.
Study highlights:
- We applied machine learning to biological data from blood, brain, bone, and genetic systems alongside psychosocial measures of mental, physical, and sociodemographic health to study underscore the diversity of biological and psychosocial factors relevant to commonly assessed painful medical conditions and chronic pain phenotypes.
- From these models we developed a blood-based biomarker that predicted the development of inflammatory, neuropathic, and nociceptive disorders, including rheumatoid arthritis, Crohn's disease, and angina. This blood-based biomarker predicted 13 distinct conditions up to nine years in advance and tracked their progression in individuals who were initially healthy.
- A second biomarker developed based on brain functional connectivity accurately distinguished nociplastic conditions like fibromyalgia, chronic fatigue syndrome, and widespread pain.
- We characterize the synergistic relationship of biological and psychosocial risk factors on the prevalence and emergence of pain conditions. We showed that biological risk factors for pain are amplified when coupled with psychosocial risk factors suggesting that the significance and interpretation of chronic pain biomarkers are critically dependent on the psychosocial context in which they are assessed.
If you have any questions, please email Matt