Purpose

Hyperpolarized ¹²⁹Xe MRI produces high-resolution, non-invasive 3D imaging of pulmonary function. However, current dose administration approaches typically require dedicated personnel to coach subjects through a breathing maneuver and deliver the dose. This process entails coaching subjects to inhale/exhale several times and then positioning the mouthpiece, manually releasing the contents of the dose bag, and prompting the subject to breathe in rapidly¹. This is further challenged by the confined space and limited visibility in the scanner bore, since the dose mouthpiece must be inserted into the subject's mouth at precisely the time of end exhalation. Although seasoned personnel can perform this technique adeptly, it is inefficient and can lead to imperfect inhalations, delays, and confusion, especially when subjects inadvertently breathe in room air or begin inhaling the dose at the wrong point of the breathing cycle. This last point is particularly crucial as the key to making ¹²⁹Xe images highly repeatable is to ensure that subjects start inhalation from a well-defined lung volume^2,3. Here, we demonstrate and test a pneumatic device that allows subjects to take their preparatory breaths with the mouthpiece and dose already in place, with study personnel actuating release of the dose remotely. We provide an initial comparison of this method vs conventional manual delivery by comparing the repeatability of lung inflation levels in the resulting images, as well as surveying subjects regarding their experience.

Methods

The pneumatic device (Figure 1) was designed to be injection molded from medical grade polymers and takes the place of the typical mouthpiece on the dose bag assembly. The interior pneumatic piston (thermoplastic elastomer) has built in seals and includes a center through-bore that allows the passage of room air prior to actuation. It is housed within a high-density polyethylene body and retainment cap, onto which a mouthpiece is fitted. All injection molded parts (Protolabs) were assembled in-house and involved light greasing of the piston, installing in the body, and gluing on the retainment cap. Actuation is achieved via positive pressure provided by a 10 ml syringe attached to a length of capillary tubing (Figure 2). The device is reusable for each subject and can be reset to the room air setting by applying negative pressure via the same syringe.
To evaluate device impact on lung inflation repeatability and patient experience, healthy subjects were recruited (n=9) to undergo breath hold proton MRI and respond to survey questionnaires (Figure 3). Images were acquired using a breath-hold 3D radial acquisition according to the ¹²⁹Xe consortium protocol¹. Room air dose volumes were tailored based on ~20% of forced vital capacity (FVC), which was rounded to either 500, 750, or 1000 ml. Subjects were randomized to first inhale air by either conventional (manual) dose delivery or by using the remote administration device, after which imaging was repeated with the other method. After a 15 minute break, they then underwent a second imaging session in which the two delivery approaches were repeated in reverse order. After the first scan of each administration method, subjects responded to a brief questionnaire assessing their anxiety level, comfort, and overall experience. During the second of these surveys, they were asked which method (if any) they preferred.
Each of the four proton scans (2 for each administration method) were segmented using an automated deep learning-based algorithm that generated a mask to define the thoracic cavity volume⁴. The resulting volumes were compared to measure inflation repeatability achieved with the two delivery methods. Questionnaire results were analyzed using Fisher’s exact test to determine preference for either delivery method.

Results

All subjects completed all four scans without incident. Lung inflation (Figure 4A+4B) was repeatable to within 0.45±0.37L with remote administration as compared to 0.51±0.50L with manual delivery. A p-value of 0.5 indicated no significant difference between the delivery methods overall. The minimum inflation difference detected using either delivery method was 0.05L.
A preliminary analysis of questionnaire results is shown in Figure 5. None of the responses were significantly different for either anxiety, comfort, or overall quality of the experience. However, there was a trend towards subjects having an overall preference for manual delivery while preferring the relative comfort of the remote delivery method.

Discussion

This study illustrates that remote, pneumatic delivery of ¹²⁹Xe is feasible and well-tolerated. This method can reduce the complexity of the delivery procedure by enabling the dose to be placed in the subject’s mouth and remain there during preparatory breaths, after which the MRI technologist can remotely release the gas contents and start the scan.
Preliminary analysis of imaging lung volumes suggests that remote delivery device provided a ~12% decrease in lung inflation difference between repeat sessions. Given the well-known inflation dependence of ¹²⁹Xe gas exchange metrics such as Membrane:Gas and RBC:Gas, this improvement is encouraging^2,3. (We also note that deep learning-based thoracic segmentations were not manually corrected, which may have led to minor imperfections in segmentation).
Survey results indicate that the new method is at least equally well-tolerated as conventional delivery approaches. Future studies conducted with a larger sample size will permit a comprehensive assessment of the quantitative and patient-experience characteristics of this new dose delivery system relative to conventional dose delivery methods.

Acknowledgements

NSF GRFP DGE-1644868, R01HL105643, R01HL12677