AI-based analysis of fetal growth restriction in a prospective obstetric cohort quantifies compound risks for perinatal morbidity and mortality and identifies previously unrecognized high risk clinical scenarios
2024

AI Analysis of Fetal Growth Restriction and Perinatal Risks

Sample size: 9558 publication 10 minutes Evidence: moderate

Author Information

Author(s): Raquel M. Zimmerman, Edgar J. Hernandez, Mark Yandell, Martin Tristani-Firouzi, Robert M. Silver, William Grobman, David Haas, George Saade, Jonathan Steller, Nathan R. Blue

Primary Institution: University of Utah Health

Hypothesis

Can a probabilistic graphical model improve risk estimation for perinatal morbidity in cases of fetal growth restriction?

Conclusion

The study found that probabilistic graphical models can effectively identify and quantify risk relationships in fetal growth restriction scenarios.

Supporting Evidence

  • 8.2% of participants experienced composite perinatal morbidity.
  • The PGM had an AUC of 0.83 in the validation cohort.
  • Identified a nearly 10-fold difference in perinatal morbidity risk based on specific clinical scenarios.
  • Female fetal sex was associated with increased risk in the presence of maternal diabetes.

Takeaway

Researchers used AI to better understand risks for babies that are not growing properly before birth, helping doctors make better decisions.

Methodology

The study used data from 9,558 pregnancies to develop and validate a probabilistic graphical model for risk estimation.

Potential Biases

Potential bias in progesterone use as it reflects clinical concern rather than objective criteria.

Limitations

The study's findings may not be fully generalizable due to the specific timing of ultrasounds and the cohort's demographics.

Participant Demographics

Participants were primarily nulliparous women with a mean age of 27 years, diverse racial backgrounds, and varying income levels.

Statistical Information

P-Value

0.8

Confidence Interval

95% CI 0.79–0.87

Statistical Significance

p=0.8

Digital Object Identifier (DOI)

10.21203/rs.3.rs-5126218

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication