An artificial intelligence (AI) algorithm trained to listen to patients urinate is able to identify abnormal flows and could be a useful and cost-effective way to monitor and manage urology patients at home. It is presented today at the annual congress of the European Association of Urology (EAU22), in Amsterdam.
The deep learning tool, Audioflow, performed almost as well as a specialized machine used in clinics, and achieves results similar to those of urology residents in assessing urine output. The current study focuses on the sound created by urine in a soundproof environment, but the ambition is to create an app that allows patients to monitor themselves at home.
Lower urinary tract symptoms, problems related to the functioning of the bladder and urethra, are common and affect approximately 60% of men and 57% of women.
Uroflowmetry is an important tool for evaluating patients with symptoms, but patients must urinate into a machine during outpatient visits. They are asked to urinate into a funnel connected to the uroflowmeter which records flow rate information. During the COVID-19 pandemic, access to clinics has been restricted, and even where patients can get there, testing can take a long time with queues to use a single machine.
Dr. Lee Han Jie and colleagues at Singapore General Hospital collaborated with colleagues in the engineering department to develop an algorithm and recruited 534 male participants between December 2017 and July 2019 to train and validate it. Participants used the usual uroflowmetry machine in a soundproof room and recorded their urination using a smartphone.
Using 220 recordings, the AI learned to estimate flow, volume and time, which can indicate if there is an obstruction or if the bladder is not functioning well. He was trained to listen and analyze male urinary stream which is different from female and would need a separate sample to learn how to analyze female urination.
The results were compared to a conventional uroflowmetry machine and a panel of six urology residents who separately assessed the data set. The RN agreed with conventional uroflowmetry for more than 80% of recordings, and compared to specialist urologists and outpatients for identifying abnormal flows, it achieved an agreement rate of 84%.
Dr. Lee says, “There is a trend to use machine learning in many areas because clinicians don’t have a lot of time. At the same time, especially since the pandemic, there is a shift towards telemedicine and less hospital care. We wanted to develop a way to monitor our patients to see how they are doing between hospital visits.
“Our AI can outperform some non-experts and approximates senior consultants,” he continues. “But the real benefit is having the equivalent of a consultant in the bathroom with you, every time you go. We are now working on getting the algorithm to work when there is noise from melts in the normal home environment and it will make a real difference to patients.
Audioflow will now be deployed as a smartphone app through primary care physicians so it can be tested in the real world and learn from different datasets in different sound environments.
Christian Gratzke, professor of urology at the University Hospital of Friborg and member of the urology committee of the EAU22 scientific congress declares: “Giving patients the possibility of measuring urine output at home is more comfortable for them and reduces the time of waiting in the clinic. This is a well-executed study with a significant number of patients and represents a promising approach to developing a wearable app that can be used at home. I can’t wait to see the results in the real world.
Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of press releases posted on EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.