The concept, known as the ‘inverted pyramid of care’, was first articulated in 2011 by fetal medicine pioneer Kypros Nicolaides. Rather than concentrating attention on diagnosing and treating complications after they emerge, the model proposes structured risk assessment in early pregnancy, so that prevention and intensified surveillance can be targeted at women and babies at elevated risk of maternal conditions such as pre-eclampsia, intrinsic fetal problems such as chromosomal abnormalities, and uteroplacental disorders such as placenta accreta spectrum. Writing fifteen years on, Laura Magee, Peter von Dadelszen and Kypros Nicolaides of King’s College London update the concept with a detailed map of what can now be screened, when, and how accurately.

A central refinement since 2011 is the recognition that early screening may be contingent. Not every risk can be resolved at twelve weeks; instead, a first-trimester assessment can identify a sub-population who should be re-screened in mid- or late pregnancy. The authors lay out an updated schedule of optimal screening timepoints—around 12, 20 and 36 weeks—matched to the most accurate available methods for each condition, and they distinguish advances that are ready for implementation from promising technologies that are not yet policy-ready.

The logic of screening rests on well-established principles: a programme is justified only when early detection of a serious condition, using an acceptable and accurate test, produces better outcomes than usual care at acceptable cost and harm. The flagship example remains antenatal syphilis screening, associated with reductions of 97 percent in congenital syphilis, 82 percent in stillbirth, 64 percent in preterm birth and 80 percent in neonatal death. For any new test, the balance hinges on the screen-positive rate, the detection rate, and the clinical, psychological and health-system harms of false positives—a trade-off that shifts as confirmatory strategies become safer, as has happened with cell-free DNA testing.

Nowhere is that shift clearer than in the detection of fetal chromosomal abnormalities, which affect 5 to 7 in every 1,000 pregnancies. Traditional combined screening—nuchal translucency, fetal heart rate, and the blood markers PAPP-A and free beta-hCG—detects roughly 90 percent of trisomy 21 cases at a 5 percent screen-positive rate. Cell-free DNA analysis, which exploits the 10 to 15 percent of placenta-derived DNA fragments circulating in maternal plasma, pushes detection above 99 percent at a screen-positive rate below 0.1 percent. Belgium and the Netherlands now offer it universally; where cost, exceeding £200 per patient in the UK, is prohibitive, contingent screening—offering cfDNA only to those who first screen positive—still achieves 96 percent detection.

Placental dysfunction is the other great target of the inverted pyramid. Preterm pre-eclampsia is rare but carries the highest individual risks, and the widely validated Fetal Medicine Foundation competing-risks model—combining maternal characteristics with mean arterial pressure, uterine artery Doppler and placental growth factor—predicts it far more accurately and cost-effectively than clinical risk factors alone, particularly for women of Black ethnicity. Crucially, prevention follows prediction: among women identified as high risk, low-dose aspirin started before 16 weeks reduces preterm pre-eclampsia before 32 weeks by roughly 90 percent. Being screen-positive also flags half of preterm small-for-gestational-age infants and two-thirds of placental dysfunction-related stillbirths before 32 weeks, because the underlying pathophysiology is shared.

Re-screening at 20 and 36 weeks extends the strategy. At 20 weeks, combining maternal history, blood pressure, uterine artery pulsatility and estimated fetal weight detects 90 percent of pre-eclampsia and 80 percent of preterm small-for-gestational-age babies. At 36 weeks—when about 75 percent of pre-eclampsia occurs at term—maternal history, mean arterial pressure, placental growth factor and sFlt-1 detect roughly 75 percent of term pre-eclampsia, compared with just one-third using clinical risk factors. The PREVENT-PE randomised trial of 8,136 women showed that personalised, risk-stratified timed birth based on 35-to-36-week risk reduced pre-eclampsia incidence by 30 percent, with no increase in emergency Caesarean or neonatal unit admissions, and saved around £500 per case averted.

Artificial intelligence emerges as a recurring theme in the authors’ forward look. Measurement error in fetal biometry currently undermines the detection of both small- and large-for-gestational-age fetuses, and the humble symphysis-fundal height measurement—still widely recommended for its simplicity—detects poor fetal growth poorly and has not improved outcomes over clinical palpation. If the spread of measurement error were halved, almost 90 percent of fetuses at risk of being small could be detected at a screen-positive rate of just 2 percent, an improvement the authors argue AI-assisted ultrasound could deliver through automated plane recognition, standardised measurements and real-time analysis of cine loops. The International Society for Ultrasound in Obstetrics and Gynaecology frames AI as an adjunct to improve quality and access, not a replacement for expert diagnosis.

The review also identifies conditions where screening is ready now: ultrasound diagnosis of placenta accreta spectrum at the 20-week scan using at least three of five markers, contingent transvaginal ultrasound for vasa praevia—a condition whose associated stillbirths are largely preventable by elective preterm Caesarean—and mid-pregnancy cervical length screening for preterm birth risk, which matters because 85 percent of preterm births occur among women without the strongest clinical risk factors. Gestational diabetes, complicating 10 to 30 percent of pregnancies, can be predicted in the first trimester with multivariable models detecting nearly 90 percent of women who will need insulin, though the authors caution that no model yet reliably identifies who will remain disease-free.

The authors close with a broader vision: pregnancy as a window onto lifelong health. Pre-eclampsia and gestational diabetes mark elevated cardio-renal-metabolic risk, but that risk is likely driven by shared underlying factors—such as chronic hypertension—identifiable in early pregnancy. As prediction improves and the gestational ‘signal of risk’ fades, they argue, antenatal care should become a launch point for long-term maternal disease prevention. Some advances, they conclude, are ready for implementation today; others, from AI-assisted biometry to metabolomic diabetes screening, will require validation. Either way, the pyramid has been turned—and momentum, they write, continues to build for prediction, prevention, personalisation and participation in maternal and fetal care.

Subject of Research: Risk-stratified antenatal care and early pregnancy screening for maternal and fetal complications

Article Title: The inverted pyramid of care in modern-day obstetrics

Article References: Magee, L. A., von Dadelszen, P., & Nicolaides, K. H. (2026). The inverted pyramid of care in modern-day obstetrics. PLOS Medicine, 23(9), e1005255. https://doi.org/10.1371/journal.pmed.1005255

Image Credits: AI Generated

DOI: 10.1371/journal.pmed.1005255

Keywords: antenatal care, pre-eclampsia, inverted pyramid of care, cell-free DNA screening, placental dysfunction, fetal growth restriction, preterm birth, gestational diabetes, placenta accreta spectrum, artificial intelligence, ultrasound screening, maternal health

The stakes are considerable. Hypertensive disorders of pregnancy, which include gestational hypertension, preeclampsia, chronic hypertension with superimposed preeclampsia and eclampsia, affect between 2 and 15 percent of pregnancies and rank as the second leading cause of maternal mortality worldwide. They sharply raise the risk of acute maternal complications such as renal dysfunction, pulmonary edema, stroke and death, and they contribute to adverse fetal outcomes including small-for-gestational-age infants, preterm delivery and stillbirth. The consequences reach far beyond delivery: women who experience these disorders carry elevated lifetime risks of hypertension, coronary artery disease, stroke and vascular dementia, and their children face higher risks of cardiovascular and neurodevelopmental problems. In the United States alone, hypertensive disorders of pregnancy were estimated to cost the healthcare system 2.18 billion dollars per year as of 2012, largely driven by complications of preterm birth.

What makes the new work compelling is its biological rationale. Preeclampsia is believed to arise in part from maternal vascular inflammation and thrombosis that disrupt placental development, impairing the deep invasion of placental tissue into the maternal decidua between roughly 10 and 16 weeks of gestation. These early deviations from normal placentation precede clinical symptoms by weeks or months, and they share features with the vascular pathology seen in cardiovascular disease. Because the retina is the only place in the body where capillaries, arterioles and venules can be directly and noninvasively visualized, the researchers reasoned that the maternal eye could serve as a living biosensor of the systemic vascular stress that accompanies early placental dysfunction. Prior studies had already found increased retinal arterial tortuosity in women with severe preeclampsia, narrower vessels and reduced capillary density in women who later developed placental insufficiency, and lower vessel densities in both the superficial and deep capillary plexuses of women with the disease.

Current screening, however, remains blunt. Preeclampsia is typically diagnosed only after 20 weeks of gestation, when hypertension, proteinuria or organ dysfunction become clinically apparent. Known risk factors such as prior hypertensive disorders, diabetes, obesity and advanced maternal age provide some predictive value but lack individual-level accuracy. The most validated biomarker, the ratio between the antiangiogenic factor sFlt-1 and placental growth factor, is primarily useful later in pregnancy and in women already suspected of having the disease. First-trimester tools such as the Fetal Medicine Foundation risk calculator combine clinical variables with biomarkers or ultrasound measurements, but these protocols are expensive, logistically complex and inconsistently implemented. In routine practice, prevention-oriented risk stratification often reduces to deciding who qualifies for low-dose aspirin, even though randomized trials show that starting aspirin before 16 weeks can reduce the risk of preterm delivery with preeclampsia by up to 62 percent. Accurate, early identification is the bottleneck that prevents that intervention from reaching the women who need it most.

Visionary AI attacks the problem with an architecture designed for interpretability rather than brute-force pattern recognition. The pipeline begins with ultrawidefield retinal images, captured by scanning laser ophthalmoscopes that cover roughly 80 percent of the retinal surface. A dedicated vessel segmentation model, inspired by generative adversarial network architectures and fine-tuned on hand-annotated ultrawidefield images, converts each photograph into a binary map of the vascular tree. When multiple images of the same eye exist, the segmentations are merged using keypoint matching and deformable registration to recover regions obscured by eyelashes or shadows. The resulting vessel maps are then transformed into three complementary families of features: graph-based measures of branching and connectivity, geometric measures of vessel length, caliber, curvature and tortuosity, and higher-order topological descriptors, including box counting, fractal organization and topological data analysis, that quantify loop structures and hierarchical network organization across scales.

These feature sets feed a stacked ensemble of machine learning models. Each semantically grouped feature set trains independent base learners built on logistic regression, random forests and extreme gradient boosting, and their out-of-fold predictions are integrated by a metalearner. The team evaluated the framework across increasingly stringent settings. In a high-contrast analysis comparing 54 preeclampsia cases against 82 carefully screened healthy controls, the model achieved an AUC of 0.94 and a positive predictive value of 0.85. When the comparison was broadened to a clinically heterogeneous population-wide control pool that deliberately retained women with obesity, diabetes and chronic hypertension, performance remained strong at an AUC of 0.91 plus or minus 0.05. Notably, adding the Fetal Medicine Foundation calculator’s probabilities to the retinal predictions did not materially improve performance, suggesting that the retinal vasculature captures most of the relevant risk signal on its own.

The most demanding test was external. The researchers built a stability-optimized model that prioritized reproducibility over peak performance, using stability-based hyperparameter selection, recursive feature elimination and feature-set deduplication to shrink an initial search space of 96 candidate base learners down to just eight nonredundant learners per fold. This deliberately constrained model was then applied, frozen and untouched, to an independent cohort of 79 pregnant individuals recruited at NYU Langone Health, a population that differed from the development cohort in race and ethnicity distribution, maternal age, in vitro fertilization prevalence and imaging hardware. Restricted to first-trimester images acquired between 6 and 13 weeks of gestation, the model achieved an AUC of 0.81 and an average precision of 0.68, outperforming the official Fetal Medicine Foundation calculator, which reached an AUC of 0.73 and an average precision of 0.50 on the same cohort.

The benchmark against routine clinical practice is perhaps the most striking result. At NYU, aspirin eligibility based on standard clinical criteria identified preeclampsia cases with a sensitivity of 0.62 but at a false positive rate of 0.47 and a prevalence-adjusted positive predictive value of just 0.1. Visionary AI achieved a higher sensitivity of 0.69 at a false positive rate of only 0.13, with a prevalence-adjusted positive predictive value of 0.31. Even within the subgroup of women already recommended aspirin, a clinically enriched high-risk group, the retinal model maintained an AUC of 0.95. The framework also outperformed two direct image-level deep-learning approaches, the retinal foundation model RETFound and a convolutional network, both of which performed near chance in external validation, a finding the authors interpret cautiously but which supports the idea that biologically informed, lower-dimensional vascular representations are better suited to clinical cohorts with limited numbers of outcome events.

Beyond prediction, the features themselves tell a coherent biological story. Recurrently selected features involved retinal vascular topology, geometry, network complexity and hierarchical organization, and model-independent analyses confirmed preeclampsia-associated differences in branch count, topological length, tortuosity, vessel thickness, fractal organization, loop count and vascular asymmetry. The signatures point to vascular simplification through two complementary processes: capillary rarefaction, the loss of small vessels and reduced branching complexity, and arteriolar wall thickening that narrows luminal diameter. Intriguingly, gestational and chronic hypertension produced distinct retinal fingerprints, with localized simplification of hierarchical vascular organization, whereas preeclampsia showed more globally extensive sparsity and disorganization. The authors stress that these associations are hypothesis generating rather than causal, and that the validation cohort, while genuinely independent, remains modest in size with a higher observed preeclampsia prevalence than the institutional average.

The practical appeal is hard to overstate. Retinal fundus imaging requires no laboratory infrastructure, takes less than two minutes, and can be performed by trained nonspecialists, making the approach potentially viable in both high-resource and low-resource settings. The computational pipeline is fully automated, eliminating subjective image grading. Before Visionary AI reaches the clinic, larger and more geographically diverse prospective studies will be needed to establish calibration, optimal thresholds and real-world screening performance, and ongoing work aims to integrate the retinal signal with circulating biomarkers and placental pathology. But if the findings hold, the humble eye photograph, already a staple of ophthalmology and increasingly a window into diabetes and Alzheimer’s disease, may become one of the most consequential tools in prenatal medicine: a two-minute scan that flags silent vascular danger early enough for aspirin, surveillance and time to change the outcome.

Subject of Research: Retinal imaging and interpretable AI for predicting hypertensive disorders of pregnancy

Article Title: Decoding systemic vascular health and hypertensive disorders in pregnancy through retinal imaging and Visionary AI

Article References: Bearelly, S., Areff, C., Hoffman, M. K., Sunko, D., Laine, A. F., Choi, B., Coleman, H. R., Cordazzo Vargas, B., Li, C. Y., Dawkins, J. J., Amir, E., Paz, S., Hark, L. A., Booker, W. A., Jauhal, A. A., Brandt, J. S., Wapner, R. J., & Shenhav, L. (2026). Decoding systemic vascular health and hypertensive disorders in pregnancy through retinal imaging and Visionary AI. Nature Biotechnology. https://doi.org/10.1038/s41587-026-03303-0

Image Credits: AI Generated

DOI: 10.1038/s41587-026-03303-0

Keywords: preeclampsia, hypertensive disorders of pregnancy, retinal imaging, artificial intelligence, Visionary AI, vascular topology, predictive biomarkers, maternal health, machine learning, prenatal screening, placental dysfunction, Nature Biotechnology

The review, led by neonatologist Yunus E. Dogan of the University of Florida together with colleagues spanning neonatology, cardiac surgery, and radiology, integrated eleven prospective and retrospective cohort studies published between 2016 and 2025. Its central numbers are sobering. Preoperative brain injury, detected on magnetic resonance imaging performed before any surgical intervention, is found in between 12 and 54 percent of neonates with congenital heart disease. That enormous range reflects differences in lesion definitions, imaging protocols, and cardiac diagnoses across studies, but even the most conservative estimates mean that at least one in eight babies arrives at the cardiac operating table with a brain that has already been injured. After surgery, the burden persists, with new lesions appearing on postoperative scans in a substantial fraction of survivors.

The patterns of injury are distinctive and diagnostically revealing. Before surgery, the most common findings are white matter injury, arterial ischemic stroke, and cerebral hemorrhage. White matter injury, sometimes described as periventricular leukomalacia-like damage, strikes the fatty sheathed fibers that connect distant brain regions and are still being actively myelinated in the third trimester. After surgery, the imaging signature shifts: white matter injury remains dominant, but clinicians also see focal infarctions that are often thromboembolic in origin, tiny microhemorrhages scattered through the parenchyma, and cerebral venous thrombosis. This temporal fingerprint, captured by serial MRI, allows researchers to distinguish lesions that were present at birth from those acquired in the operating room or intensive care unit, and it has fundamentally changed the framing of the problem from a surgical complication to a continuum of vulnerability that begins in utero.

Why is the fetal brain so fragile when the heart is malformed? The answer lies in the peculiar plumbing of fetal circulation and the metabolic demands of the rapidly growing brain. In a normal fetus, the placenta delivers oxygenated blood, and the brain enjoys a privileged, high-flow supply. In fetuses with critical heart defects, however, the anatomy reroutes this flow. In conditions such as hypoplastic left heart syndrome or transposition of the great arteries, oxygen-rich and oxygen-poor blood mix abnormally, and the cerebral circulation may be perfused at lower pressure, lower oxygen content, or both. Advanced fetal MRI studies have measured this directly, showing reduced fetal cerebral oxygen consumption associated with smaller brain size, delayed cortical development correlated with reduced oxygen delivery, and altered hemodynamic responses when mothers are given supplemental oxygen. The result is a brain that reaches birth with globally reduced volumes, microstructural abnormalities in structures like the corpus callosum, and in some cases aberrant functional connectivity detectable before any operation.

The maternal-fetal environment adds another layer of risk. The review highlights evidence linking placental dysfunction and preeclampsia to preoperative focal white matter injury in neonates with complex heart disease, an emerging field some researchers call neuroplacentology. Three-dimensional volumetric MRI of the placenta in fetuses with complex congenital heart disease has revealed measurable differences compared with controls, and placental pathology correlates with neuroimaging abnormalities after birth. In other words, the same compromised maternal-fetal conditions that stress the heart may simultaneously starve the developing brain of oxygen and nutrients. Perinatal and neonatal characteristics compound this vulnerability: male sex, prematurity, and low birth weight all emerge as risk factors, echoing patterns seen in preterm brain injury more broadly. Notably, the immature oligodendrocytes that build myelin in the third trimester are exquisitely sensitive to hypoxia-ischemia and inflammation, which helps explain why white matter bears the brunt of the damage in this population.

Cerebral autoregulation, the brain’s ability to maintain steady blood flow despite swings in blood pressure, appears to be impaired in many of these newborns even before surgery. Studies of preoperative infants with congenital heart disease have documented blunted autoregulatory capacity, and novel noninvasive measures of cerebrovascular stability are being tested against invasive monitoring to bring this physiology to the bedside. The clinical implication is profound: a blood pressure that is perfectly adequate for one baby’s brain may be dangerously low for another’s, and the transition from fetal to neonatal circulation, with its shifting shunts and falling pulmonary resistance, is a period of particular hazard. Preoperative procedures such as balloon atrial septostomy have been scrutinized as potential stroke triggers, though large studies suggest that oxygenation and time to surgery are more important determinants of injury than the procedure itself.

Then comes the operation, and with it a new set of insults. Cardiopulmonary bypass, the heart-lung machine that sustains circulation during repair, exposes the brain to nonphysiological flow, embolic debris, inflammatory activation, and ischemia-reperfusion injury when the heart is restarted. Neonatal blood is uniquely challenging in this setting: standard heparin doses achieve less consistent thrombin inhibition in newborns than in older patients, and the coagulation system swings between thrombosis and bleeding with little margin for error. Anticoagulation strategy, surgical timing, hemodynamic instability, and the use of extracorporeal membrane oxygenation all emerge in the review as procedural risk factors for postoperative brain injury. Focal infarctions after surgery are frequently thromboembolic, and cerebral sinovenous thrombosis has been documented in infants undergoing cardiac surgery, sometimes producing thalamic hemorrhage and requiring difficult anticoagulation decisions in a fragile patient.

There is, however, genuine cause for optimism embedded in the data. The review notes that advances in blood pressure management, tailored perfusion strategies, and optimized surgical timing have already reduced postoperative injury rates, and recent multicenter work has documented a declining incidence of postoperative neonatal brain injury over time. Studies comparing cohorts with different practice approaches show that institutional choices matter, and European collaborative data have begun to map which perioperative factors predict long-term neurodevelopmental outcomes better than conventional imaging alone. The stakes of these improvements are high: perioperative brain injury on neonatal MRI is associated with worse school-age neurodevelopment, reduced white matter volume predicts language difficulties, and hippocampal volume reductions in adolescents with congenital heart disease track with intellectual function. Survivors face elevated rates of motor delay, cognitive and executive function deficits, and abnormalities in brain regions regulating anxiety and mood.

The path forward, as the authors frame it, is precision neuroprotection across the entire continuum of care. Antenatal interventions, including maternal hyperoxia strategies being tested with blood oxygenation level dependent fetal MRI, could potentially bolster the fetal brain before birth. Individualized hemodynamic management targets, guided by continuous cerebral autoregulation monitoring rather than population-wide blood pressure thresholds, could protect the brain during bypass and in the intensive care unit. Multimodal neuromonitoring, incorporating continuous electroencephalography and amplitude-integrated EEG, which has been shown to predict preoperative MRI findings, could catch injury in real time. Risk-stratified anticoagulation could balance clot and bleeding risks, and neuroprotective drugs such as allopurinol are already in phase III randomized trials for babies requiring bypass surgery. Finally, long-term neurodevelopmental surveillance programs would ensure that early injuries, however subtle, are met with early intervention.

For the roughly one percent of newborns whose hearts form differently, the message of this synthesis is both a warning and a promise. The brain injury that shadows congenital heart disease is not an inevitable consequence of a defective heart, nor simply the price of a heroic operation. It is a mechanistically traceable process, visible on MRI from fetal life onward, driven by hypoxemia, impaired autoregulation, inflammation, and thrombosis, and shaped by the placenta, the delivery room, the bypass circuit, and the intensive care unit. Each of those checkpoints is now a target. If the trajectory of declining postoperative injury continues, and if antenatal and precision perioperative strategies mature from promising trials into standard care, the generation of children born with heart defects may one day keep their hearts repaired without paying a hidden neurological price.

Subject of Research: Brain injury mechanisms, MRI patterns, and perioperative risk factors in neonates with congenital heart disease

Article Title: Brain injury in congenital heart disease: pathophysiology, MRI patterns, and perioperative risk factors

Article References: Dogan, Y. E., Vargas-Boggiano, D. A., Bleiweis, M., Philip, J., Narasimhulu, S. S., Tuna, I., & Weiss, M. D. (2026). Brain injury in congenital heart disease: pathophysiology, MRI patterns, and perioperative risk factors. Pediatric Research. https://doi.org/10.1038/s41390-026-05454-x

Image Credits: AI Generated

DOI: 10.1038/s41390-026-05454-x

Keywords: congenital heart disease, brain injury, white matter injury, neonatal stroke, MRI, cardiopulmonary bypass, cerebral autoregulation, placental dysfunction, neuroprotection, neurodevelopmental outcomes, fetal circulation, thromboembolism