In very preterm infants, intraventricular haemorrhage (IVH) can have serious consequences. If this leads to a pathological enlargement of the cerebral ventricles (post-haemorrhagic ventricular dilatation, PHVD), measurements of cerebral oxygenation sometimes deteriorate markedly and over the course of weeks. This is shown by a study based on 154 preterm infants (<34 weeks’ gestation), including 65 with PHVD, which was recently published in Pediatric Research (Nature Portfolio). The AIT Center for Health & Bioresources ensured an appropriate analysis of the data.
High clinical dynamics: brain burden is not always evident early
IVH is among the more common brain injuries in preterm infants. If PHVD develops subsequently, this can be associated with increased intracranial pressure and inflammatory processes, thereby placing an additional burden on brain development. In treatment, the question is therefore not only about structural changes (e.g. via ultrasound), but also about how stable the brain’s oxygen supply remains over a longer period of time.
To capture this, 154 preterm infants (<34 weeks’ gestation), including 65 with PHVD, were observed over an extended period, repeatedly or continuously, using near-infrared spectroscopy (NIRS)—i.e. trajectories over weeks were examined. In total, ~29,700 hours of measurement data were collected. This longitudinal observation is important because risks and burdens in preterm infants do not become apparent in a single measurement point, but often only in temporal patterns in the NIRS time-series data. The study, which was conducted within the framework of the WWTF project PIMIENTO (“PrecIsion Medicine in IntravENTricular hemorrhage for Outcome prediction”), produced the following results:
- With increasing severity of the haemorrhage, regional cerebral oxygen saturation decreases.
- Infants with PHVD show lower oxygen saturation, more time below a critical threshold, and higher oxygen extraction—indications of sustained metabolic stress.
- For infants with PHVD, the following applies: the more often temporary neurosurgical decompressions were required, the less favourable the measurements of cerebral oxygenation.
Continuous monitoring and robust data processing
NIRS is a non-invasive monitoring method that provides ongoing information on cerebral oxygenation and can thus complement imaging with a functional perspective. In this work, the explanatory power relies largely on the fact that very large measurement series over weeks are processed and analysed in a methodologically consistent manner (including defined steps for artefact cleaning and trajectory evaluation).
AIT has strong expertise in study and analysis concepts as well as in the statistical modelling and interpretation of complex biomedical datasets. In the PIMIENTO project, AIT is working on data-driven approaches to predict treatment outcomes after IVH, in particular by integrating molecular data with imaging and electrophysiological data, thereby strengthening a longer-term research line at the interface of neonatology, data analysis and precision medicine. “Our strength lies in linking biomedical data from different domains in such a way that a mechanistically explainable and clinically usable prediction of disease trajectories becomes possible,” explains Klemens Vierlinger, who is responsible for bioinformatics in the Competence Unit Molecular Diagnostics at the AIT Center for Health and Bioresources. “The data show that a higher burden due to PHVD is associated with impaired cerebral oxygenation over a longer period. This underlines how important it is to capture these trajectories functionally and to take them into account in clinical assessment,” says Lucia Ciglar, also in the Molecular Diagnostics team.
Better decision-making through integrated trajectories
Based on the findings, the authors conclude that targeted strategies are needed to stabilise cerebral oxygenation in high-risk preterm infants. Looking ahead, the combination of continuous monitoring of electrophysiological and molecular data is likely to help to better substantiate risk profiles and possible intervention time points.