An artificial intelligence (AI) algorithm can now listen to patients pass urine in order to successfully and efficiently identify abnormal flows and the corresponding health issues, according to a press release published on July 2nd. The deep learning tool is appropriately called Audioflow, and it has thus far performed almost as well as a specialist machine used in clinics and provided similar results to urology residents.
Evaluating the sound created by urine
“There is a trend towards using machine learning in many fields because clinicians do not have a lot of time. At the same time, particularly since the pandemic, there is a shift towards telemedicine and less hospital-based care. We were keen to develop a way to monitor our patients to see how they are doing between hospital visits,” said Dr. Lee Han Jie from Singapore General Hospital, the study’s lead.
The current algorithm evaluates the noise resulting from urine produced in a soundproof environment, but researchers hope to develop an app that is self-sufficient enough that patients can use it to monitor themselves at home. Current uroflowmetry is effective at assessing urinary-related conditions but requires that patients urinate into a machine during outpatient visits.
However, the COVID-19 pandemic has restricted access to clinics. Han Jie and colleagues wanted to develop a more effective way to assess urine at home without any medical assistance and so they recruited the help of the engineering department to develop a urine-assessing algorithm.
To train and validate this algorithm, they recruited 534 male participants between December 2017 and July 2019. The process was fairly straightforward: participants used the usual uroflowmetry machine in a soundproof room and recorded their urination using a smartphone.
Results that matched professional outcomes
Using a mere 220 recordings, the AI learned to accurately evaluate flow rate, volume, and time, all of which together can point to an obstruction or to issues with the bladder.
“Our AI can outperform some non-experts and comes close to senior consultants,” explained Han Jie. “But the real benefit is having the equivalent of a consultant in the bathroom with you every time you go. We are now working towards the algorithm being able to work when there is background noise in the normal home environment, and this will make the true difference for patients.”
Indeed, Audioflow produced results that could compete with a conventional uroflowmetry machine and a panel of six urology residents. The AI produced conclusions that were in line with conventional uroflowmetry for over 80% of recordings and, compared to the specialist urologists, it achieved an 84% rate of agreement.
Now, the researchers hope that the new AI could soon prove beneficial in home settings.
Audioflow will soon be rolled out as a smartphone app to be tested in the real and very noisy world. It does, however, have one drawback: it has thus far only been tested on male urinary flow which is different than the female one. Could a female-focused version be on the horizon?