Artificial Intelligence in Functional Urology: How It may Shape the Future

Imad Bentellis; Sonia Guérin; Zine-Eddine Khene; Rose Khavari; Benoit Peyronnet


Curr Opin Urol. 2021;31(4):385-390. 

In This Article

Other Therapeutic Applications of Artificial Intelligence in Functional Urology

Innovative Devices

The literature about AI therapeutic implications in functional urology is scarce. Whangbo et al. developed an inventive device capable of taking the place of the voiding diary. This is a smart band worn on the wrist that recognizes urination time and intervals between each miction from on specific postures of the patient. The voiding diary is the cornerstone of functional urology but it is restrictive and requires patients to stay at home to perform it. This device would potentiate the clinical examination by providing information on real-life VDs.[27] Another interesting field to invest is electrotechnical artificial urinary sphincters (emAUS) designs, implantation, and control. Marziale et al. have shown in their extensive review all the possible emAUS developments, wireless, magnetic, endo, or extraurethral.[28] There is a large possibility of using AI tools to design such devices based on anatomical or radiological data, predict their outcomes, and adapt dynamically their functioning to unconventional activation protocols. Methods explored for sensing,[29] actuating, and controlling[30] the cuff occlusion includes shape memory alloys, mechatronic hydraulic systems, electroactive polymers, piezoelectric, and magnetorheological components.

Outcomes Prediction

OAB has recently been pointed out as a very heterogeneous clinical entity with high rates of poor responses to first-line medical treatments attributed to a uniform therapeutic approach lacking well-defined patients' selection criteria.[31] In a cohort of 559 female patients treated with anticholinergic medications for OAB, Sheyn et al. developed and externally validate a prediction model for anticholinergic response using an RF algorithm. They found excellent performances of the model (sensitivity 80.4%; sensitivity = 77.4%).[32] Despite its encouraging results, we can make the criticism that this study relied on a subjective report of successful treatment.