Innovator Spotlight Q&A Series: Shinichi Goto, PhD
We are pleased to feature Shinichi Goto, PhD, instructor in the Division of Cardiovascular Medicine at Brigham and Women’s Hospital, for developing an AI model to help screen for heart defect.
Q: Tell us about your innovation and the challenge(s) you are trying to solve. Who are the people involved?
SG: We have developed an artificial intelligence (AI) model to detect atrial septum defect (ASD) from a single recording of a 12-lead electrocardiogram(ECG). ASD is the most common form of adult congenital heart disease, affecting up to 0.5% of all newborns. While ASD is asymptomatic in most cases, it increases the risk of irreversible complications such as atrial fibrillation (AF), stroke, and heart failure. ASD can be treated by minimally invasive percutaneous closure. Thus, diagnosing the disease before the onset of complications is essential. However, the most effective screening tool (echocardiograms) is time consuming and cannot be performed on all the asymptomatic subjects, and thus, the disease is largely underdiagnosed.
Our AI model detects ASD excellently from a single recording of ECG. The model was able to greatly improve the sensitivity for detecting ASD compared to conventional approach using ECG abnormalities.
Q: The innovation process can be long and challenging, but also rewarding. What inspired you to begin this journey?
SG: My background training in computer science (though not formal) played an important part in formulating this research. I was willing to work in a research field in the interface of computer science and medicine/biology. In my clinical training, I felt that a significant number of patients could have been treated better if they were found in an earlier phase, and thought that my skill sets could be useful in this area. After starting my work on training AI models to improve medical practice, I realized that the selection of target diseases is an essential part. ASD was selected as an underdiagnosed yet treatable disease that could be detected by a non-invasive modality.
Q: Where are you in the innovation cycle (i.e., early-stage commercialization)?
SG:We are still in a very early phase of the innovation cycle (seed). While the AI model has shown strong promise for detecting ASD from ECG, the model needs to be integrated into a easy to use software/platform in the commercialization process for wide-spread use. While, we strgongly hope that the model will be commercialized, we do not have a specific plan yet.
Q: What internal resources have been most helpful to you?
SG: To build a robust AI model, a large dataset is required. We achieved this by utilizing the large ECG data repository and the enterprise data warehouse. The availability of such a large and detailed data set to researchers is a significant strength of MGB.
Q: If you could give one piece of advice to another innovator in the Mass General Brigham network, what would it be?
SG: MGB is a great place to develop innovations, given the abundant resources from one of the world’s best healthcare systems. My recommendation is to maximally utilize these exceptional resources and pursue any great ideas that could lead to great innovations.