Urogynecology Study Using AI-based Tools Awarded $2.65 Million Through NIDDK

A research team at Duke, led by Principal Investigator Nazema Siddiqui, MD, MHSc, has been awarded a four-year, $2.65 million R01 grant through the National Institute of Diabetes and Digestive Kidney Diseases  (NIDDK) to study Overactive bladder (OAB) and how Artificial Intelligence-based tools and algorithms could potentially be used to help with treatment options. The grant began Feb. 1.

When patients have urinary urgency, frequency and/or nocturia — with or without urgency incontinence — they are considered to have OAB. Male and female patients of all age groups can develop OAB, but this condition becomes extremely common in women over age 40. Post-menopausal women in particular are more likely to suffer from urgency urinary incontinence, which takes a toll on physical and social function, as well as overall health and vitality.

Currently, all patients with OAB are treated with the same algorithm. This “one-size fits all” algorithm often results in repeated medical visits, high health care costs and marginal long-term improvement in symptoms. However, for years clinicians have suspected that there are several underlying factors contributing to OAB symptoms, and specifically several subtypes that could potentially respond differently to treatments, according to Dr. Siddiqui. She notes, for example, in women, there is early evidence of a subtype with dysbiosis, where a shift in the urinary microbiome may contribute to irritative symptoms in the bladder. If dysbiosis is the cause of irritative bladder symptoms, treatments to re-establish a healthy microbiome may be more effective for some patients than what is routinely prescribed for OAB.

The research team’s aim is to determine if OAB subtypes exist, and to further develop an artificial intelligence (AI)-based tool that would allow clinicians to better target the right therapy to the right patient. To do this, data from over 4,000 patients gathered in two large-scale NIH cohort studies will be used. Predictive modeling, which is a type of clinically informed AI, will be utilized to evaluate subtypes and to see if patients from different subtypes respond differently to common therapies using microbiome profiles. The goal is identifying ways for patients with OAB who are seen in clinic to receive more effective, targeted therapy from the start, instead of putting them on a care pathway where they may eventually progress through five or six treatments (with all of the associated office visits).

In addition to Dr. Siddiqui as corresponding PI, co-investigators are urogynecologist Eric Jelovsek, MD, MMEd, MSDS; Li Ma, PhD, Duke Department of Statistical Science; and Charles Page, PhD, Duke Department of Biostatistics & Bioinformatics. Collaborators include co-PI Lisa Karstens, PhD (Oregon Health & Science University); and co-investigator A. Lenore Ackerman, MD, PhD (UCLA).

“For years, non-cancerous gynecologic issues, like OAB, have received far too little attention despite the high impact that they have on people’s daily quality of life. Duke has several investigators that are leading the way in how to sensibly use AI in health care.  I am thrilled to incorporate these data scientists with our microbiome research team so that together we can apply innovative, cutting-edge technologies to improve an area of  women’s health where better therapies are desperately needed.” — Dr. Nazema Siddiqui, urogynecologist