Predicting pregnancy risks earlier

Montclair State University Mathematics Professor Diana Thomas and a team of researchers from universities, medical schools and the private sector, are conducting research that could revolutionize obstetric care and make at-risk pregnancies safer. Using geometric equations to measure placental growth, Thomas hopes to be able to predict problem pregnancies much earlier, which could reduce risks to both mother and fetus.

“Currently, complicated pregnancies are treated when symptoms are visible from physical measurements or biomarkers. Often the ball is already rolling and despite best efforts, changing the direction of the pregnancy is impossible,” says Thomas, co-principal investigator. “Mathematical models magnify deviations not otherwise readily detectable earlier, offering a chance to change the course of pregnancy to a healthier and positive outcome. That’s a pretty exciting use of mathematics.”

A normal, healthy placenta grows in a predictable pattern, and the team has learned that abnormal growth can be detected because it falls off the predictive curve. Even in the first trimester, this could indicate problems with the pregnancy such as gestational diabetes or preeclampsia, the later of which, if not caught early, can lead to a sometimes-fatal condition.

In both cases, earlier detection would be a breakthrough, Thomas says. “Identifying women at risk for both preeclampsia and gestational diabetes during early pregnancy using non-invasive, routine ultrasound techniques provides a very real opportunity to intervene and change the course of pregnancy toward more positive outcomes,” she says.

Photo of faculty researchers observing a doctor and nurse demonstrating an ultrasound exam.
Seated from left: Amir Golnabi, math professor; OB-GYN Richard Miller of St. Barnabas Medical Center. Standing from left: Philip Yecko, physics professor, The Cooper Union; engineer Ruchit Shah of Placental Analytics; computer science Professor Emeritus Carl Bredlau; David Trubatch, math professor; principal investigator and perinatal pathologist Carolyn Salafia; Diana Thomas, co-PI and math professor; Andrada Ivanescu, statistics professor. Laying down: nurse volunteer.

Furthermore, the team, comprised of leading experts in placental morphology, obstetrics, mathematics and reproductive physiology, is working to develop the first class of differential equation models that combine 19 placental measures to identify and classify risk for problems in pregnancy. “It’s amazing what we can do with mathematics, but we can’t do it with math alone,” says Thomas, adding that a interdisciplinary team is crucial.

That team includes engineers who extract placental features from images to use in the algorithm; a perinatal pathologist who evaluates the accuracy of model predictions; an OB-GYN who evaluates the course of the pregnancy and the results for use in clinical practice; an applied mathematician and a physicist who simulate through lab experience what can’t be tested on humans; a biostatistician who combines the predictors into a formula to classify pregnancy risk; and a computer scientist who delivers the results through user-friendly software.

“As a perinatal pathologist, I routinely see cases where problems are only recognized for a few days or even a few hours before delivery but the placenta clearly demonstrates pathology of weeks or even months of age,” says Principal Investigator Carolyn Salafia, a perinatal pathologist with Placental Analytics in New York. “This work may allow early identification of at-risk pregnancies and improved fetal childhood and lifelong health.”

The study has been flagged for funding by the Eunice Kennedy Shriver National Institute of Child Health and Human Development, the National Institutes of Health and for an award from the federal Small Business Innovation Research program.