Machine Learning Models Identify Gut Microbiota That Predict Chronicity in Reactive Arthritis

Poster Presenter: Prakashini MV, MD, Fellow in Training, Kalinga Institute of Medical Sciences, Kalinga Institute of Industrial Technology (KIIT University), India

Poster Title: Machine Learning Models Identify Gut Microbiota That Predict Chronicity in Reactive Arthritis

Poster Session C: Tuesday, Nov. 14

What is your poster about?
“My poster follows a cohort of patients with reactive arthritis (ReA) over three years, comparing the gut microbiota in those who developed chronic arthritis with the ones who did not.

“ReA is a unique infection-triggered sterile arthritis that occurs at a distant site and is mostly described to have a lower-limb predominant, oligoarticular phenotype. Though gastrointestinal (GI) and genitourinary infections have been described as potential putative triggers, in this part of the world, GI infections predominate and hence, our research focuses on the gut microbiota, its diversity, and potential pathogenicity in ReA. We also employed supervised machine learning models adopting different clinical and metagenomics data to identify the factors that accurately predicted chronicity in these patients.

“Our team had previously undertaken a 16S metagenomics study in patients with ReA to explore the diversity of gut microbiota and identify the potential pathobionts contributing to disease. We followed up with patients from this cohort over three years and captured data on who progressed to chronicity and attempted to ascertain the factors that predisposed them to it.”

“Though the gut-organ axis is being described for potentially every disease, the possible direct causation in ReA is intriguing.”

— Prakashini MV, MD

Why did you decide to investigate this topic?
“ReA is unique in the sense that it is triggered by an infection, but sterile, and its incidence is limited to pockets across the world, thus limiting the availability of clinical material for research. The state of Odisha in the eastern part of India is one such geographically and possibly genetically predisposed site. ReA is traditionally described to be self-limiting by 12 weeks; however, our clinical practice suggests that nearly 50% of patients in this part of the world progress to chronicity.

“Though the gut-organ axis is being described for potentially every disease, the possible direct causation in ReA is intriguing. Our team members in the past have explored cytokine profiles, metabolomics, proteomics, and recently metagenomics. Building on this, we sought to explore the clinical factors and the microbiome, and employ machine learning models to predict chronicity in this set of patients, which would eventually translate to early intervention and overall reduced damage accrual.”

What are you working on next related to this research?
“After 16S metagenomics and predictive analysis with machine learning, we are planning to study the transcriptome to identify distinct pathways contributing to the pathogenesis of this unique arthritis. We ultimately plan to employ a systems biology approach to integrate all our findings to propose a comprehensive pathogenic model for reactive arthritis.”

What excites you most about your work?
“The thought of simple infective diarrhea causing severe arthritis that can become chronic and potentially accrue damage resulting in loss of functionality is a cause for concern. By identifying the potential organisms or the right genetic mechanisms contributing to this, we may be able to accurately predict chronicity and possibly intervene to prevent long-term damage.”

What are you most looking forward to at ACR Convergence 2023 in San Diego?
“I look forward to learning from the giants of the field, sharing our research with the vibrant global rheumatology community, networking with like-minded young rheumatologists, brainstorming ideas with each other and finding potential collaborators to ultimately benefit science and humanity, and of course, making memories to last a lifetime!”