Latest News

Maternal Diagnostics Meets Machine Learning

January 28, 2022 | Elaine Gee is Principal Data Scientist at Mirvie, a company dedicated to using technology to improve maternal health and identify pregnancy complications early using blood-based RNA profiling. Gee will be speaking at the upcoming Sensors Summit 2022 March 2-3, which will be held both virtually and in person in La Jolla, CA. Diagnostics World chatted with Gee to learn more about her experience and the philosophy of Mirvie ahead of her upcoming presentation.

Diagnostics World: How did you progress into the field of data science? Given your Ph.D. focused on fibronectin and involving various wet lab techniques coupled with computational and simulation work, did you always know you wanted to work as a data scientist, or did your career just evolve that way?

Elaine Gee: I am passionate about building new technologies that can improve people’s lives in a way never possible before. As a data scientist in healthcare, I’ve learned how vital it is to combine computational theory, laboratory testing, and clinical validation to successfully develop solutions that truly impact people’s lives. I have been attracted to opportunities where it is essential for these aspects to intersect throughout my career. Along the way, I’ve played key roles in developing multiple next-generation sequencing-based tests and smarter continuous glucose monitors. I find it rewarding that the technologies I’ve helped develop have improved many people’s health and wellbeing. 


Once you established yourself as a data scientist, what led you to join Mirvie, your current employer?

I was impressed by Mirvie’s focus on pregnancy health, which like many areas of women’s health, has seen little advancement relative to other areas of healthcare. For the first time in my career, I am able to leverage both my technical expertise as a data scientist and my unique perspective as a woman to help bridge a wide gap with enormous unmet needs. I’m excited to be part of this world-class team of researchers, experts, and innovators working to shape a new future of pregnancy health by predicting unexpected complications before they happen.


How common are pregnancy complications, and what impact can they cause to moms and babies?

One in five pregnancies is impacted by complications such as preeclampsia, preterm birth, and gestational diabetes—impacting more women each year than all forms of cancer. They lead to large economic costs and lifelong health consequences for moms and babies. They result in 2-4 times increased lifetime risk of cardiovascular disease for moms and lifelong neurological and behavioral issues for babies. Black women in the USA experience a 50-100% higher complication rate. More than $50 billion in healthcare costs are attributed to pregnancy complications in the United States and European Union alone. On top of all of this, they leave an emotional toll that lasts a lifetime for both mom and baby.


And how are pregnancy complications traditionally diagnosed? Why is it so challenging?

Today, women and doctors lack a reliable way to detect complications before there are symptoms. Most complications are detected late in pregnancy when symptoms occur and often in a state of crisis. Only 20% of pregnancy complications can be detected using today’s generalized risk assessments based on pregnancy history, race, ethnicity, body mass index, and medical history. These generalized factors don’t provide a fundamental understanding of the underlying biology for each unique pregnancy, preventing personalized care.


What is Mirvie doing to address these challenges?

At Mirvie, we are developing the first RNA platform to predict unexpected complications before they happen by revealing the underlying biology of each pregnancy. The platform reveals vital information about the future health and disease of a pregnancy, allowing women to act and their doctors to intervene before unexpected complications become a crisis. This enables proactive, preventative, and personalized care for the well-being of moms and babies that is not possible today.


How does Mirvie’s platform work? 

Using a simple blood sample from the mom, the Mirvie RNA platform combines revolutionary analysis of tens of thousands of RNA messages from the baby, the placenta, and the mom, with machine learning. The platform identifies patterns in RNA that reveal the underlying biology of pregnancy health. These RNA profiles allow the detection of pregnancy complications well before they occur. 


Our recent study published in Nature shows that the Mirvie RNA platform detects 75% of women who will develop preeclampsia, including those with premature births—months before symptoms. This is the first study to demonstrate RNA can reveal the underlying biology of pregnancy and uses preeclampsia to show its enormous potential. We are conducting ongoing clinical research to further validate these results, enhance the Mirvie RNA platform performance, and improve the understanding of other pregnancy complications. (Read more in Monitoring Pregnancy Progress With RNA Profiling)


Where does your work fit into the platform and Mirvie’s overall research goals?

My role at Mirvie is to advance the machine learning models that leverage the patterns within the tens of thousands of RNA messages to predict pregnancy complications. I apply my data science expertise in healthcare to ensure we have robust signatures to create high-performance predictive models suitable for clinical application in the market.


What about your experience working at a startup vs. a larger company, what’s it like working for a startup, and what do you enjoy about it?

I enjoy how working at a startup presents daily opportunities to flex my creativity to solve difficult challenges in meaningful ways. It feels like we are continually breaking new ground with our statistically rigorous machine learning techniques, laboratory processes, global research collaborations, and our understanding of the underlying biology of pregnancy. It is inspiring to be part of a world-class team that is so strongly dedicated to bringing breakthrough innovation to pregnancy health because it is an area that has lagged behind other medical fields for far too long.


As a data scientist, I’m sure you must keep a pulse on the latest and greatest developments in algorithms and computing technology, so what new ML/AI developments or latest technologies are particularly interesting to you right now?

With more machine learning models deployed to real-world applications, the field recognizes that unbiased model performance across a diverse population is critically important. Machine learning requires large datasets to ensure that models are trained, tested, and validated on independent datasets to create a robust model, so it is critical to ensure that this data is diverse and represents the demographics of the wider population for optimal performance across many subpopulations. At Mirvie we are ensuring our clinical studies are large and diverse to verify the Mirvie RNA platform works for all women. Our breakthrough results published in Nature show that this is possible. By analyzing the largest and most diverse dataset of maternal transcriptomes to date (2,539 blood samples from 1,840 women in the United States, Europe, and Africa), the research proves the platform directly detects the underlying biology independent of demographics or clinical factors.


Editor’s Note: To hear more from Elaine, register now for the Sensors Global Summit. Use discount code SGS300 to save $300 on your registration.

Load more comments