Prenatal care (before birth), during pregnancy, and postnatal care (after birth) are crucial for ensuring a healthy pregnancy and newborn. As these aspects of care evolve, the role of Artificial Intelligence (AI) in maternal and neonatal health raises important questions, generating both interest and concern among medical practitioners.
According to the World Health Organization (WHO), India faces a significant burden of birth defects, with over 1.7 million children affected annually, and several of whom succumb to mortality due to delayed detection and inadequate care. Therefore, early diagnosis and intervention are crucial in reducing the impact of these birth defects and improving survival rates.
Diagnosing foetal abnormalities and the use of AI
Foetal abnormality detection is a critical aspect of prenatal care, ensuring early identification of potential complications. Traditional ultrasound methods rely on skilled sonographers and obstetricians, but challenges such as inter-observer variability and limited access to expertise can impact accuracy. AI-powered imaging, according to the specialists can transform feotal diagnostics by enhancing precision, standardising assessments, and improving neonatal outcomes.
However , specialists in maternal and neonatal health in India stress that AI should not replace human expertise but rather assist in time-consuming tests. They emphasise the importance of clinical oversight, ensuring that AI complements trained sonographers in interpreting results and providing holistic care. This perspective aligns with a recent study by King’s College London and Guy’s and St Thomas’ NHS Foundation Trust, which found that AI-assisted scans nearly halve screening time while maintaining diagnostic reliability.
Key areas of AI integration in foetal imaging
According to Deepa Hariharan, senior neonatologist based in Chennai , accurate gestational age assessment is fundamental for managing newborns, particularly premature or sick infants. It is typically determined in the first trimester, but if high-quality scans are unavailable early in pregnancy, errors of up to two to three weeks can occur. In such cases, AI-enhanced ultrasound assessments can help in determining accurate gestational age, allowing better preparation for potential complications.
Monitoring foetal growth involves measuring parameters like head circumference (HC) and abdominal circumference (AC). HC helps assess brain development and hereditary conditions, while AC indicates intrauterine health. AI aids in reducing human error and standardising biometric measurements, improving early detection of anomalies such as intrauterine growth restriction (IUGR) or macrosomia related to maternal diabetes.
The 20-week anomaly scan is a crucial milestone in prenatal screening. Post-28 weeks, Doppler studies or Doppler ultrasonography, evaluate foetal blood flow and amniotic fluid levels, essential for assessing foetal well-being. Here, AI-integrated ultrasound systems assist in real-time monitoring, helping clinicians identify distress signals early and plan timely interventions, including preterm deliveries or cesarean sections.
AI in maternal, neonatal and paediatric care
According to Harini Sreedaran, consultant neonatologist and paediatrician, Narayana Health, Bengaluru, AI can aid in diagnostic accuracy, particularly in paediatric care. “Paediatric emergency departments can use AI to triage patients, determining which cases need urgent attention,” Dr. Sreedaran says. Similarly, AI models can help predict and diagnose conditions such as neonatal sepsis and pneumonia by analysing symptoms alongside imaging and laboratory findings.
“In paediatric endocrinology, AI can aid in managing conditions like diabetes, adjusting treatment plans based on symptom patterns and patient responses,” she says. Moreover, it can also contribute to patient education by alerting caregivers to seek medical help when concerning symptoms arise, helping prevent delayed interventions.
In maternal health, AI plays a vital role in improving prenatal care by predicting and monitoring complications like preeclampsia and gestational diabetes, enabling timely interventions that can significantly improve outcomes for both mothers and babies. It can also support personalised care plans, tailoring treatment based on a woman’s specific health needs.
The existing foetal imaging faces several challenges, including inconsistent scan quality and a shortage of highly trained personnel. Several ultrasound centers lack expertise in conducting detailed anomaly scans, leading to misdiagnoses or the need for repeated imaging. AI has the potential to address these challenges by standardising assessments and enhancing diagnostic accuracy, particularly in underserved regions.
Dr. Hariharan points out AI-powered portable ultrasound devices can aid in situations where clinical decision-making can be through data-driven insights, thereby expanding prenatal diagnostics to remote and resource-limited areas. This, she says, can aid in providing better maternal and foetal health.
Why expert evaluation matters ?
While AI holds immense promise in neonatal and maternal healthcare, Dr. Sreedaran emphasises that it should complement, not replace, medical professionals. “AI can refine diagnostics and streamline processes, but human expertise remains essential in interpreting results and making critical decisions,” she says.
Swati Bhayana, consultant in paediatric hematology and oncology at Fortis, Gurugram, says,” AI can assist in complex diagnoses, acting as a second opinion in cases where there is uncertainty. It provides valuable insights and can help in detecting patterns that might be missed by the human eye.”. However, she stresses the need for human judgment in medical decision-making. “Even though AI can provide precise information, there still needs to be an experienced doctor to determine what to do with that information,” Dr. Bhayana says.
AI can also generate a wide range of possible diagnoses, but without expert evaluation, there is a risk of misinterpretation. “One of the biggest hurdles is establishing normative values that reflect India’s diverse population. Paediatric AI, in particular, faces difficulties due to variations in height, weight, and developmental metrics across different regions,” Dr. Hariharan says. Additionally, she says “AI must be tailored to account for the unique healthcare needs and constraints of both urban and rural settings”.
Dr. Bhayana likens AI’s role in medicine to online search engines. “Whenever you ask Google what causes a fever, it gives you a million answers, ranging from the common flu to cancer. AI works in a similar way; it might suggest both benign and severe conditions, but a doctor is needed to differentiate between them.”
According to specialists AI in sonography serves as an aid rather than a replacement. While it enhances diagnostic accuracy and speeds up the process, the final decision must always be made by a qualified medical professional.
Published – April 01, 2025 03:10 pm IST