How is artificial intelligence applied during pregnancy? How can we calculate gestational age from AI?

Gestational age is the indicator that allows us to carry out a more exhaustive follow-up of the pregnancy. This is calculated from the biparietal perimeter of the skull and the cranial circumference in the first trimester, unlike the second and third trimesters in which, in addition to these parameters, the length of the femur is added.

The application of Artificial Intelligence for the evaluation of fetal diagnosis based on ultrasound allows us to achieve the same results or with fewer errors unlike the traditional method. This is achieved by taking the transthalamic axial plane of the brain through 2D ultrasound and comparing it with current formulas based on standard fetal biometric parameters in a large cohort of previous pregnancies.

The new calculation method for gestational age from the transthalamic axial plane was developed from the use of DeepLearning. The study was conducted based on a total of 1,394 images of previous pregnant patients and the help of maternal-fetal specialists, these being manually labeled with orientation and specific brain landmarks using computer software. This was necessary for him to learn to locate the brain in the image known today as quantusGA.

In the following image, we can see in detail how it works quantusGA:

Once the ultrasound image of the fetal brain is received in DICOM format as impute, in a matter of days a output with gestational age (GA/GA).

The quantus GA program allows us, first of all, to detect the position and orientation of the fetal skull. Next, from the key points you can cut the image or rotate it allowing you to view the desired position. Lastly, information about the texture and size of the brain can also be extracted. All this with the purpose of generating the EG.

The study that was subsequently carried out using python (Python Software Foundation, Wilmington, DE) in order to affirm whether Artificial Intelligence improves the prediction of gestational age in the second and third months of pregnancy concluded with equal results or even fewer errors compared to the biometric fetal parameters traditionally used. Likewise, with the use of the entire image and its resolution, the method allows us the relevant changes for the estimation of GA that, unlike the human eye, are probably associated with the growth and maturation of the brain.

Thus, AI mostly helps us predict the gestational age of the baby, contributing to optimal prenatal care. Being able to access this is essential to act on multiple perinatal interventions, among which are: induction of labor, fetal lung maturation or even pulmonary resuscitation. On the other hand, this method could be implemented in an automated program applicable to any ultrasound machine. Therefore, this could be applied in numerous areas of the planet and have a greater impact in places where obtaining an ultrasound in the first trimester of pregnancy is difficult.

In addition, it can offer many other advantages, among which it is worth highlighting:

• Better care for pregnant women
• Calmer patients
• Reduction in maternal-fetal mortality rates
• Better control of changes in pathology

Bibliography:

1.- Xavier P Burgos-Artizzu, David Coronado-Gutiérrez, Brenda Valenzuela-Alcaraz, KilianVellvé, ElisendaEixarch, FatimaCrispi, Elisenda Bonet-Carne, Mar Bennasar, EduardGratacos. Analysis of maturation features in fetal brain ultrasound via artificial intelligence for the estimation of gestational age. August 14, 2021. Am J ObstetGynecol MFM.

WRITTEN BY

Eva Arnall

Saturdays.AI

Saturdays.AI is an impact-focused organization on a mission to empower diverse individuals to learn Artificial Intelligence in a collaborative and project-based way, beyond the conventional path of traditional education.