Wearable Sensors + AI for Accessible and Personalized Women’s Health Monitoring
ABSTRACT
As the world population is aging, there is an increased demand for effective healthcare technologies that are accessible, accurate, and personalized. Furthermore, consumers are increasingly interested in taking charge of their own health with increasing demand for comfortable and convenient wearable technologies such as fitness trackers, intelligent patches and armbands. These wearable systems often utilize miniaturized sensors to collect user’s health data and integrate with smart algorithms to track changes over times to detect early signs and progression of diseases. These smart wearable systems can enable remote patient monitoring by physicians and proactive self monitoring by the patients themselves. Even though women compromise more than half of the population and make more than 70% of healthcare decisions for the household, there has been a lag in development of smart wearables for monitoring women’s health. Many women globally don’t have access to any health monitoring. Cost-effective AI-powered wearables will help in democratizing health monitoring for all women everywhere.
1 in 8 women will be diagnosed with breast cancer (BC). Breast cancer is the leading cause of cancer death among women and incidence is on the rise globally, with 1.7M new BC cases annually. Treatment of BC at earlier stages is associated with better survival outcomes: the 5-year relative survival rate is 98.6% if diagnosed at the local stage, 84.9% at the regional stage, and 25.9% at the distant stage. Early detection also results in significant savings for healthcare system, however, today 1 in 3 breast cancers are missed at early stages. Most BC deaths occur in unscreened or under-screened women, therefore access to regular monitoring of breast health would maximize survival rate. However, over a 1 billion women don’t have access to any regular BC screening.
Mammogram as a screening tool has significant shortcomings such as poor sensitivity in dense breasts affecting 47% of women in US, limited access and frequency, resulting in ~56% of BCs discovered in between screenings. Furthermore, the poor patient experience results in low adherence (33% drop-off after first mammogram). Ultrasound (US) is a proven and most scalable technology worldwide with considerable advantages such as no harmful radiation, portability, and lower cost. US typically has better sensitivity than mammography. However, the image quality of handheld US probe is highly dependent on the operator skill often resulting in poor diagnostic specificity. This along with shortage of skilled US technologists has been a key impediment in broad adoption of US for breast cancer screening.
At iSono Health, our mission is to transform health monitoring with a platform that combines AI with automated, quantitative, wearable ultrasound. Our first product, ATUSATM, features a wearable, automated ultrasound with AI-powered, on-premise image processing and analysis to offer consistent quality, lower cost, and unprecedented access. ATUSA system produces repeatable 3D US images, independent of operator skill in just 2mins. Hardware of the scanner has a transducer, electronics, and mechanical components packaged in a compact assembly. The scanner attached to bra-like wearable accessory for accurate and repeatable image acquisition. This disposable accessory accommodates different breast sizes made of inexpensive plastic and silicone materials.
ATUSA data is transferred to our software suite to undergo AI-assisted 3D image processing, visualization, and evaluation. This software application is integrated with an API with our cloud-based machine learning (ML) engine and backend infrastructure utilizing an existing HIPAA-compliant platform. De-identified data and images are synced to cloud regularly and update ML algorithms with ever-expanding datasets. Our ML engine first detects images with lesions and classifies those with 94%+ accuracy. Those images are then triaged and transmitted to experts for final diagnosis. Furthermore, our ML engine extracts acoustic biomarkers from quantitative and repeatable ultrasound data to identify and track functional tissue properties enabling early detection and monitoring of breast cancer. Because of its’ operator-independence and AI-powered image analysis, ATUSA system can be deployed in point-of-care, pharmacy, and low resource settings such as rural and mobile clinics, and finally at home for remote patient monitoring and diagnosis.
BIOGRAPHY
Maryam Ziaei, PhD, CEO of iSono Health has over 15 years of experience in product and business development. He has led iSono Health team over past 5 years from founding, to initial development of automated and wearable ultrasound systems, fundraising ($5M venture funding), hiring, clinical development and strategic partnerships. Also she is the principal investigator on several SBIR grants (over $1.2M in non-dilutive funding). Before founding iSono Health, she had led engineering teams in design and manufacturing of microscale devices and systems. Maryam also completed an entrepreneurship degree at Stanford Business School. Maryam has several peer-reviewed publications and patents in the area of MEMS/sensors. She holds a PhD in Electrical Engineering from Stanford, and a MS in Electrical Engineering from UC Berkeley.