Know just by hearing it - AI for Prostate Health

Original Article: https://www.hankyung.com/article/202309271998i

Below is an article published in Hankyung on October 2nd, 2023, translated into English. Great interview of CEO Catherine Song and great story written by See-eun Lee (see@hankyung.com).

==========

“There were three criteria for selecting business items. “It was the conquering of chronic diseases, the use of smartphones, and the introduction of artificial intelligence (AI).”

The prevalence of urinary diseases that occur mainly in the prostate and bladder increases with age. That is why there are so many patients who suffer in everyday life. Catherine Song, CEO of Soundable Health, developed a technology that measures the sound of urine with a smartphone and analyzes it with AI to confirm the severity of the disease. CEO Song explained, “We selected the initial service area by narrowing down from 100 diseases one by one.”

In 2018, one year after starting the business. She moved to the United States and began collecting patient data. It took another four years to commercialize. I held on and held on. Recently, it has started to achieve the desired results. We were able to compete with 2,000 companies at a famous healthcare award in the United States and be named as a major company. CEO Song said, “There are approximately 300 million pieces of patient data collected from across the United States, and the AI accuracy is 97%. Next year, we will finalize related procedures to enable local insurance claims and significantly expand the number of users.”

CEO Song, born in 1978, lived a life far from entrepreneurship. She studied electronic engineering from BS to PhD at KAIST. When CEO Song was in school, it was not as easy to find student startups as she is now. She studied communications networks in a university lab, and started his first job at LG Electronics in 2007. She primarily worked in the Technology Strategy Department. She also worked at LG Management Research Institute and semiconductor company Ram Research Korea.

“The reason I ended my eight-year career was when my father suddenly left my side.” CEO Song said, “I was very shocked when she started coughing little by little and passed away in less than 15 days. I knew that hospitals could not do everything, and I thought we needed a medical system that could know and manage health conditions in advance.” said. Afterwards, I experienced working in a healthcare startup for two years through the arrangement of Future Play CEO Ryu Joong-hee, who was my senior in the same lab during graduate school. And the startup she founded with his colleagues at LG Electronics in 2017 is now Soundable Health.

It took half a year to select the item. We searched various papers and actuarial statistics for keywords such as chronic diseases, smartphones, and AI. CEO Song says, “We found an area with even a little bit of market potential.” Those selected were urinary and respiratory diseases. First, we targeted the urinary system. She created a solution that identifies various properties of urine sounds, such as the size, pitch, and tone, and analyzes the urine speed per unit time. She came up with a solution to help doctors make decisions. In the first year of business, we were able to implement low-level AI through clinical trials with Seoul National University Bundang Hospital.

In 2018, the company ‘flipped’ (relocated its headquarters) to the United States. CEO Song said, “The investment company advised that the created solution would have greater opportunities in the U.S. market, where hospital accessibility is poor,” and added, “The domestic medical AI device licensing system is complicated, so we decided to relocate.” She said, “When I first arrived in Silicon Valley, I had no money and nowhere to go, so I moved around to my friend’s house.”

Right now, we had to focus on data collection rather than paid conversions. It was in 2020 that full-scale deep learning technology was introduced and class 2 medical device registration was achieved from the U.S. Food and Drug Administration (FDA). Last year, we succeeded in making it payable. This is the result of collecting a total of 310,000 urine samples from local patients. Currently, major hospitals such as Cornell University Hospital in the U.S. and Mount Sinai Hospital in New York are customers and partners. CEO Song said, “In the United States, it is common to wait four weeks before taking leave and visiting a urologist, so a solution that allows patients to record their urine history with a device at home was useful for doctor diagnosis.”

Recently, it was selected as a candidate for the top award in the ‘Healthcare AI Utilization Field’ at the ‘Digital Health Awards’ of the U.S. Digital Health Hub Foundation. The U.S. Digital Health Hub Foundation is a non-profit foundation affiliated with the University of California, and has been awarding excellent healthcare startups every year since 2017. CEO Song said, “This is a contest that attracts attention from hospitals and pharmaceutical companies in the U.S., so healthcare startups headquartered in the U.S. submit first and then watch.” She added, “This year, the number of applicants was 2,000, which was the largest ever, and it was among the 14 major companies in the field of AI use. “It was chosen as one,” he said.

Song currently spends half of the year in the United States. Last year, he developed an AI solution that analyzes coughing sounds. She plans to try linking the urine AI solution with the U.S. insurance payment system next year. She said, “The current solution is for doctors to separately recommend use to patients, but we are working on a plan to have prescription codes assigned through the American Medical Association and insurance companies. Once prescriptions and insurance claims become possible, the number of users will expand in earnest.” It was predicted. Target sales are $5 million by 2025.

CEO Song encouraged domestic companies to actively challenge the U.S. market. Domestic medical AI companies are struggling with issues with licensing and insurance premiums, but the door to opportunity is much wider in the United States. She said, “It is by no means that the US FDA approval process or insurance registration process is simple,” but added, “The advantage is that negative regulatory values are incorporated throughout.” It is said that during the preparation process, we often heard that commercialization is allowed first if it is not related to human life, and responsibility is held if problems arise later.

She said, “I have been running a small group to help healthcare entrepreneurs advance into the U.S., but the difficulties faced by the domestic medical AI startup industry still seemed big.” In fact, the approval of AI medical devices in Korea and the method of calculating the insurance premiums associated with them are long-standing challenges. The problem of fee calculation is that if it is treated as non-benefit, the number of customers of high-level general hospitals decreases, and if it is treated as salary, the unit price is drastically lowered according to the doctor's fee standard. The United States has separate fees for medical AI.

First, the Korean network established in Silicon Valley was also cited as an advantage in entering the market. CEO Song said, “I received a lot of help from the beginning from Noom’s CEO Jeong Se-ju and Sendbird CEO Kim Dong-shin,” adding, “There is a strong atmosphere of helping each other as long as the English problem is solved, so I hope that many Korean entrepreneurs will advance.”