This prospective study shows that unsedated advanced neuroimaging incorporating measurement of CVR by administering exogenous CO2 was overall well tolerated in children as young as 6 years of age, although the youngest with usable data was 8 years of age. All children were able to complete the MRI session. As compared to the MRI only group, children undergoing measured CVR did report more symptoms, although none were serious or long-lasting. Interestingly, a few children undergoing only MRI also reported symptoms of headache, shortness of breath or lightheadedness (n=6/25). Children with chronic disease, including sickle cell disease and reactive airway disease, did not have a difference in reported comfort during the scan. This finding is important for both clinical and research investigations as it gives evidence that changing CO2 for brain imaging is not contraindicated in developmentally normal children and can be completed safely in common childhood conditions such as reactive airway disease. However, this study did not include children with known intracranial disease so does not address all childhood conditions. In our study, MRI scan duration did not contribute to the tolerability of the procedure itself, with all scans being close to an hour.

Our results are congruent with symptoms of similar imaging methodology in a cohort of 294 healthy individuals across the lifespan, which included 59 participants ages 9–19 years, reported by Spano et al. [8]. Symptoms were not specified by age range in this study, but minimal symptoms were reported across the cohort, with the most common being shortness of breath (6.5%), followed by anxiety/claustrophobia (2.5%), headache, and lightheadedness (0.5% each). The type of symptoms experienced by our participants was similar, with shortness of breath being most common (56%), followed by headache and lightheadedness. We did not specifically ask about claustrophobia in our study, but this is unlikely as the participants completed the hour-long MRI scans. The increased incidence of symptoms from our cohort is likely explained not only by the younger age range but also by the prospective nature of our study. Due to the retrospective nature of their study, Spano et al. were only able to detect symptoms that were significant enough to be documented in the research records.

Of our CVR session participants prospectively asked about their comfort level on a scale of 1–6, 45% (32/71) reported experiencing some discomfort (score 4–6). This is higher than the 11.1% in Spano et al. retrospective cohort who had noted discomfort [8]. However, in our cohort, only 9 (13%) selected the uncomfortable or very uncomfortable response, which may be a more accurate comparison. Spano et al. noted that half of those who had reported discomfort had a repeat scan in the database, demonstrating that the discomfort was not severe enough to discourage further participation. Similarly, the majority of our participants who found the scan uncomfortable reported that they would be willing to complete the scan again. Importantly, there were no serious adverse events in either study, supporting the safety of this methodology.

The preference scale has advantages for translating the research experience to common childhood activities and has been used in prior studies [16, 17]. It allows for pediatric age-appropriate ranking of any intervention within a grouping of other activities that are universally unliked (vomiting) and enjoyable (spending time with friends). Thus, investigators can easily compare rankings across a cohort while comparisons are made between groups to evaluate similarities of the population in interests and the intervention or methodology being studied. However, it also has limitations. First, while it lists common experiences, some have a bias based on socioeconomic status. Some children who participated in the study live in an urban setting and required transportation assistance to the scan. They may have different experiences and perception of “long” car rides, possibly not having taken many road trips longer than an hour. Other children may have taken multi-hour road trips, but with the ability to watch movies on devices. Modern handheld devices were not available when the scale was created in the late 1990s. Even with these limitations, it still communicates that the MRI, with or without CVR measurement, was not the most enjoyable activity (as expected) but also that most children did not consider it the worst option on the list.

Importantly, only 3 children undergoing MRI measured CVR reported that they would not be willing to repeat the scan in the future, and 51 would definitely be willing to repeat the scan visit. The option of “maybe” was included to ensure that children who were unsure of repeating the scan or of possible consequences of their response had an option other than “no” to report and thus strengthen our confidence in the sincerity of the “yes” response. This provides context for both the level of discomfort and the ranking of the scan, and it is encouraging data for the feasibility of longitudinal studies.

There are several limitations of this study. First, we include a potentially biased research sample of those who consented to participate in an MRI research study. It is unknown how children who declined to participate may or may not tolerate the imaging without sedation. Second, while children with chronic conditions were included, they were otherwise healthy and feeling well at the time of the research scan. However, this is likely similar to clinical imaging, with the exception of urgent imaging needs. Third, this is a developmentally appropriate cohort. These findings cannot be extrapolated to children who are not at the cognitive level of an elementary school child, either by developmental stage or cognitive impairment. Fourth, we did not include children with known intracranial disease, so it is unknown whether or not that intracranial disease would impact these findings. Fifth,we did not use the common terminology criteria for adverse event (CTCAE) [21]. By this scale, ranging from mild to causing death, all symptoms reported would be categorized as mild, as none required interventions. This would have impeded our primary interest in capturing more subtle variations of children’s perception of their experience, including symptoms that spontaneously resolved. An additional limitation is that while it is one of the largest pediatric cohorts of exogenous CO2 imaging reported to date, it is still a relatively small sample size, and a larger cohort may discover more symptoms or difficulty tolerating the scan.