Assessment of cerebrovascular development and intraventricular hemorrhages in preterm infants with optical measures of the brain arterial pulse wave

Abstract

Preterm infants (born at 24-34 weeks of gestational age) suffer from a high incidence of neurological complications. Cerebrovascular lesions (intraventricular hemorrhages, IVH, and ischemic injury) due to the immaturity of the vascular system and its inability to adapt to the extra-uterine environment are the major causes of adverse neurological outcomes. We investigated the feasibility of assessing cerebrovascular status in preterm infants using a novel non-invasive optical procedure, pulse-DOT, usable within the incubator. Pulse-DOT, validated in adults, provides estimates of cerebral arterial status based on optical measurements of the pulse wave. These measurements are taken with a high-density optode montage and provide accurate spatial and temporal information. We found that two pulse parameters (pulse relaxation function, PReFx, and pulse rise time, PRT) in the investigated frontotemporal region, correlated with infant's age at recording, indexing cerebrovascular development. Moreover, PRT differentiated infants with and without concurrent IVH (sensitivity = 100%, specificity = 70%). These values are at least as high as those of the resistivity index obtained with transcranial Doppler of the middle cerebral artery, the current clinical method of choice for investigating arterial elasticity in preterm infants. This makes pulse-DOT a promising tool for investigating cerebrovascular risk factors and related pathologies in preterm infants.

Keywords: Preterm infants; brain arterial pulse wave (pulse-DOT); cerebrovascular development; diffuse optical tomography; intraventricular brain hemorrhages.

Figure 1.

(a) Experimental setting: two optical patches were placed on each brain hemisphere with the infant remaining within the incubator. (b) Locations of sources (red) and detectors (blue) overlaid onto the right and left hemisphere of the template infant’s brain. (c) Estimates of the region investigated by the optical patches for the youngest (top) and oldest (bottom) infants. Average light sensitivity is displayed on a logarithmic scale up to an attenuation of 100 times its maximum value.

Figure 2.

(a) Average optical pulse waveform from a representative voxel and infant, showing relative changes in NIR light absorption related to the heart pulsation. The four pulse-DOT parameters, pulse amplitude (PA), relative pulse transit time (rPTT), pulse rise time (PRT), and pulse relaxation function (PReFx) are shown on the waveform. Note that computation of PReFx also involved a normalization step not shown in this figure (i.e. dividing the area by both systolic-diastolic peak interval and PA). (b) Time course of the optical pulse waveform averaged across voxels and infants, separately for infants with (blue) and without IVH (red). The spread of the waveforms represents the standard error of the mean for each data point.

Figure 3.

(a) Map of the average pulse amplitude (PA) and (b) Map of the average relative pulse transit time (rPTT) overlaid onto the template infant brain. PA and rPTT were computed only for voxels with data from at least half of the infants.

Figure 4.

Scatter plots of (a) pulse amplitude (averaged across all voxels) vs. gestational age (days/weeks) at the time of optical recording; (b) pulse relaxation function (averaged across all voxels) vs. gestational age (days/weeks) at the time of optical recording; (c) pulse rise time (averaged across all voxels) vs. gestational age (days/weeks) at the time of optical recording; and (d) resistivity index (measured from the middle cerebral artery with TCD) vs. gestational age (days/weeks) at the time of the Doppler measurement.

Figure 5.

(a) Box plot of pulse rise time (averaged across all voxels) for infants with and without IVH. (b) Box plot of the resistivity index (measured from the middle cerebral artery with transcranial Doppler) for infants with and without IVH. (c) ROC curves reporting the ability of PRT and RI to correctly classify infants with IVH as a function of different classification thresholds.

Figure 6.

Maps displaying voxels for which a contrast between the pulse rise time (PRT) values in infants with and without IVH yielded a t-value exceeding 2.5 (p < 0.01, uncorrected), overlaid onto the template brain.