The science of pregnancy is continually improving. Here is a forthcoming new, not yet commercially available, test to determine preterm birth risk.

Breakthrough Blood Test Predicts Preterm Birth with High Accuracy

Imagine being able to predict whether a pregnant woman is at risk of having a preterm birth, allowing for early interventions to prevent complications. A recent study published in PLoS Medicine has made this a reality, with a blood test that uses artificial intelligence (AI) and genomic sequencing to predict preterm birth with a high degree of accuracy.

What is Preterm Birth?

Preterm birth, which occurs in approximately 11% of all births worldwide, is a major cause of morbidity and mortality for both mothers and their babies. It can lead to respiratory distress, infections, and even long-term health problems.

How Does the Blood Test Work?

The blood test analyzes a sample of cell-free (cf) DNA, which is already widely used to check on the genetic health of babies during pregnancy. The researchers used whole genome sequencing on cfDNA samples from 2,590 pregnant women, focusing on gene promoter profiling. They then applied machine learning models and algorithms to develop a predictive genetic signature for preterm birth.

The Results are Promising

The study, led by Jia Tang and colleagues at the Guangdong Provincial Reproductive Science Institute in China, found that a model called PTerm, which uses a type of machine learning algorithm, was able to predict preterm birth with an accuracy of 0.88 (using a statistical measure called the area under the curve, or AUC) [1]. This means that the test can effectively distinguish between women who will have a preterm birth and those who will not.

But What Does This Accuracy Mean?

An AUC score of 0.88 is considered a high level of accuracy. To put this into perspective:

• An AUC score of 0.5 would indicate a test that is no better than chance.
• An AUC score of 0.7-0.8 would indicate a moderately accurate test.
• An AUC score of 0.9 or higher would indicate a highly accurate test.

In this case, the PTerm model achieved an AUC score of 0.88, which suggests that it is a highly accurate test. However, it’s essential to note that no test is 100% accurate, and there may be false positives (women predicted to have a preterm birth who don’t) and false negatives (women predicted not to have a preterm birth who do).

Validation and Future Directions

The researchers tested PTerm in three independent groups of women and found that it maintained a good AUC score of 0.85. While more research is needed, these findings suggest that the test could be a valuable tool for predicting preterm birth in early pregnancy.

What Does This Mean for Pregnant Women?

If rolled out on a larger scale, this blood test could help identify women at risk of preterm birth, allowing for early interventions to prevent complications. This could include closer monitoring, medication, or other treatments to support a healthy pregnancy.

Takeaways

• A new blood test uses AI and genomic sequencing to predict preterm birth with high accuracy.
• The test analyzes cell-free DNA, which is already widely used to check on fetal health.
• The study found a high accuracy of 0.88 in predicting preterm birth.
• The test could be a valuable tool for identifying women at risk of preterm birth and preventing complications.

Limitations and Future Research Directions

While these findings are promising, there are some limitations to consider:

• The study was conducted in a specific population of women, and more research is needed to confirm the results in diverse populations.
• The test may not be suitable for all women, and more research is needed to understand its applicability.

Overall, this breakthrough has the potential to revolutionize prenatal care and improve outcomes for mothers and babies.

Read the full study: https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1003252

References:

[1] Tang, J., et al. (2023). Promoter profiling of plasma cell-free DNA predicts preterm birth. PLoS Medicine, 20(3), e1003252. doi: 10.1371/journal.pmed.1003252