Initial feasibility and challenges of hyperpolarized 129 Xe MRI in neonates with bronchopulmonary dysplasia

Abstract

Purpose: The underlying functional and microstructural lung disease in neonates who are born preterm (bronchopulmonary dysplasia, BPD) remains poorly characterized. Moreover, there is a lack of suitable techniques to reliably assess lung function in this population. Here, we report our preliminary experience with hyperpolarized ¹²⁹ Xe MRI in neonates with BPD.

Methods: Neonatal intensive care patients with established BPD were recruited (N = 9) and imaged at a corrected gestational age of median:40.7 (range:37.1, 44.4) wk using a 1.5T neonatal scanner. 2D ¹²⁹ Xe ventilation and diffusion-weighted images and dissolved phase spectroscopy were acquired, alongside ¹ H 3D radial UTE. ¹²⁹ Xe images were acquired during a series of short apneic breath-holds (˜3 s). ¹ H UTE images were acquired during tidal breathing. Ventilation defects were manually identified and qualitatively compared to lung structures on UTE. ADCs were calculated on a voxel-wise basis. The signal ratio of the ¹²⁹ Xe red blood cell (RBC) and tissue membrane (M) resonances from spectroscopy was determined.

Results: Spiral-based ¹²⁹ Xe ventilation imaging showed good image quality and sufficient sensitivity to detect mild ventilation abnormalities in patients with BPD. ¹²⁹ Xe ADC values were elevated above that expected given healthy data in older children and adults (median:0.046 [range:0.041, 0.064] cm² s^-1 ); the highest value obtained from an extremely pre-term patient. ¹²⁹ Xe spectroscopy revealed a low RBC/M ratio (0.14 [0.06, 0.21]).

Conclusion: We have demonstrated initial feasibility of ¹²⁹ Xe lung MRI in neonates. With further data, the technique may help guide management of infant lung diseases in the neonatal period and beyond.

Keywords: bronchopulmonary dysplasia; hyperpolarized 129Xe; lung MRI; neonatal MRI.

Figure 1:

MRI equipment set-up: a) neonatal MRI scanner and patient table; b) ¹H-¹²⁹Xe RF coil, with frequency switching via insertion/withdrawal of mechanical rods (arrows); and c) breath-hold gas delivery apparatus with Tedlar bag, pressure gauge and clinically available neonatal face mask.

Figure 2:

Examples of coronal ¹²⁹Xe spiral ventilation images (single series selected from several dynamic acquisitions) from four subjects. Qualitatively, a range of ventilation defects (regions of no/reduced signal within the lung cavity) can be observed; varying in both number and size (examples indicated by blue arrows). Subject numbers are shown (refer to Table 1). P: posterior, A: anterior.

Figure 3:

Example comparisons of approximately slice-matched coronal images from ¹²⁹Xe spiral ventilation and ¹H structural UTE MRI acquisitions. In most subjects, large ventilation defects were found to correspond to structural abnormalities (blue arrows), whilst occasionally, ventilation defects did not appear to be associated with structural abnormalities (green arrow; see also Table 2). The structural defects in subjects #4 and #8 are thought to be due to atelectasis, likely related to inflammation. Anecdotally, these subjects had a total Ochiai score of 6 and 3, respectively. Subject numbers are shown (refer to Table 1).

Figure 4:

Example ¹²⁹Xe diffusion-weighted imaging data; ADC maps acquired in Subject #8 and #4. Note the elevated ADC for Subject #4 who was born extremely preterm (22 wk GA at birth); see also, Tables 1 and 2.

Figure 5:

Example ¹²⁹Xe dynamic spectroscopy data, depicting the time series of signal intensity (left) and the mean spectra (right) obtained from averaging the spectra acquired between 1 and 2 seconds (region indicated by curly braces in the left plot) for two subjects. Red blood cell (RBC), tissue “membrane” (M) and gaseous ¹²⁹Xe resonances are indicated. Subject numbers are shown (refer to Table 1).