Shanghai Tuoxiao Intelligent Technology (Tuoxiao) is working to improve the health of mothers and children across China by creating a potential solution that aims to detect congenital heart disease and lung problems using AI-enabled technology. We caught up with Tuoxiao’s founders, Joey Zhou and Fei Qu, to learn more about the company’s origins and future.
|The Problem||Our Solution|
Many congenital heart disease and lung defects in infants are easily treated if identified early enough[i], but if left untreated, they can lead to permanent and debilitating diseases like pulmonary hypertension[ii]. Across China, hospitals lack the facilities needed to screen for these conditions and in rural areas in particular, doctors may lack relevant training and experience[iii].
Screening is essential for congenital heart disease and lung defects and this is traditionally undertaken through auscultatory methods, such as a stethoscope. Tuoxiao has taken this idea and created AI software that can listen and detect auscultatory irregularities, with the aim to enable more people to be screened faster, more consistently, and even remotely, with the goal of helping more infants across China get a timely diagnosis.
Welcome, Joey and Fei, it’s great to have you with us today! Can you tell us more about your professional backgrounds and how the idea to found Tuoxiao was born?
Joey: Well, my background is an education in science and engineering, and I specialized in medical devices R&D. But it was around 2015, amid the boom of mobile medicine, when I started looking into how traditional medical devices such as the stethoscope could be combined with artificial intelligence and developed into portable mobile and home medical devices to improve the patient experience.
Fei: My background is a little different; I worked in the charity sector prior to Tuoxiao. My work focused specifically on congenital diseases in infants in remote areas. In fact, congenital heart disease can be easily treated if diagnosed early. Unfortunately, for many of the children I met during my work, the disease had progressed significantly by the time it was detected. The idea at that time was to train rural doctors to master the auscultation technique to screen local children in a timely manner. However, such training programs were later proved difficult to implement due to various factors. This is what led me to think about how sound recognition technology and artificial intelligence can be used to facilitate and improve the initial diagnosis of congenital heart disease.
This is how Tuoxiao was born. Joey brought the deep insight into medical devices, and I brought the focus on congenital diseases. We used this outlook to begin making digital stethoscopes that could record sound which was then sent to a doctor for analysis. With the rapid advance of technology behind us, our system has expanded to become so much more than just a digital stethoscope.
Now, everyone is familiar with the concept of a stethoscope, but how do you make it digital and combine it with AI? What does it look like in practice?
Fei: At the simplest level, a digital stethoscope is a bit like an ordinary one but with an exceptionally advanced audio recording device in it. This worked well for helping people in remote areas get screened and diagnosed, but it was still quite slow, which is why we quickly started exploring new cloud technologies and AI. We saw an opportunity to develop a set of algorithmic screening and management software equipped with AI on top of the digital stethoscope. The goal was to put the auscultation experience of professional doctors into the AI algorithm by training the model to achieve the appropriate skills to detect congenital heart disease.
This is significant because it means that we could potentially use the AI diagnosis technology through the cloud to provide the service to every village in China at the highest speed and the lowest cost, to achieve comprehensive congenital heart disease screening.
That sounds remarkable, but also like hard work. What are some of the biggest challenges you’ve faced so far?
Fei: The biggest challenge is the legality of the algorithm. Since the algorithm can be used to diagnose, if there is a mistake, it’s not clear where the accountability lies. Because we’re pioneers in this field, there’s no legal precedent and we’re totally committed to making sure it is covered legally. We are currently in the process of applying for the CFDA certification, which is a difficult but fascinating process. It can be frustrating because it’s slow, but we want to make sure we’re doing things right.
Joey: Like any start-up, the other big challenge is finance. The public benefits of our work are irrefutable, but investors are sometimes worried it won’t be lucrative enough and so are hesitant to join us. We’re hoping to overcome this through better promotion – the more ubiquitous we are, the more confidence investors will have. To achieve this, we’re looking into ways to promote our brand within hospitals.
Many start-ups can relate with your challenges. What’s been your experience as a resident company of Johnson & Johnson Innovation – JLABS?
Fei: Firstly, Johnson & Johnson Innovation – JLABS have really helped us reach a position where we can face our challenges and thrive. The Family of Companies of Johnson & Johnson is a leader in healthcare, so being based at JLABS @ Shanghai has really broadened our horizons. I cannot understate the value of having Johnson & Johnson Innovation – JLABS’ recognition and backing. Not only does it help when we’re introducing Tuoxiao to new people, but it’s also personally reassuring to have that stamp of confidence.
Secondly, we’ve been able to learn a lot from experts within The Family of Companies of Johnson & Johnson, particularly regarding how we communicate our company vision. The resources provided to us as a resident company have helped us hone our investor pitch and refine Tuoxiao’s business model and development strategy.
What could the future hold for Tuoxiao?
Fei: The next goal is to create an integrated diagnostic software and hardware that can make the entire process more seamless, thereby delivering better quality of care for patients whilst simultaneously easing the workload for doctors. To achieve this, we need finance which we’re considering raising by applying for overseas certification, allowing us to increase our potential market. We also want to collect more data to consistently improve our technology’s accuracy. This can be done by building partnerships with other organizations and promoting public awareness of congenital conditions so that they come and get screened. It’s going to be a busy but exciting year!
Apply today to join Johnson & Johnson Innovation – JLABS: jlabs.tv/apply or email us for more information,
Disclaimer: The information in this interview is for informational purposes only. The information is either proprietary or licensed to Johnson & Johnson Innovation LLC, its parent company or any affiliate companies ("JJI") or is provided by or proprietary to third parties. In case of third-party information, please note that only those third parties that provided or own this information are responsible for it. By providing this information, JJI is not affiliated with nor endorses these third parties or any of the information provided in this interview. This content is not intended to influence the use, sale, recommendation, or promotion of any products or services of JJI. JJI is not responsible or liable for any actions you do or do not take in relation to this information and are not responsible for any use of this information by you. By providing the information in this interview, JJI does not make any representations, warranties, or assurances, express or implied, as to the content or the information presented in this interview and is not responsible for any use by or alterations of this information by you. JJI hereby expressly disclaims any and all guarantees, representations, warranties, or assurances and assumes no responsibility or liability, direct or indirect, in relation to the information in this interview.