Abstract
Objective:
Screening for cervical spine injury (CSI) after blunt trauma is common, but there remains varied practice patterns and clinical uncertainty regarding adequate radiographic evaluation. An oft-cited downside of MRI is the added risk compared to CT in the pediatric population; however, these specific risks have not yet been reported. This study examines the risks of cervical spine MRI in pediatric trauma patients in the context of what value MRI adds.
Methods:
Retrospective observational study of all pediatric blunt trauma patients who were evaluated with a cervical spine MRI over a four-year period at a level 1 pediatric trauma center. Clinical and radiographic data were abstracted, as well as anesthesia requirements and MRI-related major adverse events. CT and MRI results were compared for their ability to detect clinically unstable injuries—those requiring halo or surgery.
Results:
There was one major adverse event related to MRI among the 269 patients who underwent cervical spine MRI—a rate of 0.37%. While 55% of children had an airway and anesthesia for MRI, only 57% of these airways were newly placed for the MRI. None of the 85 patients newly intubated for MRI developed aspiration pneumonitis or ventilator-associated pneumonia, and no patients had a significant neurologic event while at MRI. Another area of the body was imaged concurrently with the cervical spine MRI in 64% of patients and 83% of MRIs were performed within 48 hours. CT and MRI were both 100% sensitive for injuries requiring halo or operative intervention. Eighty-three patients had an MRI performed after a negative CT, 11% (9/83) of these patients had a clinically stable injury detected on subsequent MRI, and none of these patients presented for delayed cervical spine complications.
Conclusions:
Overall, the safety profile of MRI in this setting is excellent and less than one third of patients need new airway and anesthesia solely for MRI. In this clinical scenario, MRIs can happen relatively quickly and many patients require another body part to be imaged concurrently anyway. MRI and CT were both 100% sensitive for clinically unstable injuries. In the appropriate patients, MRI remains a safe and radiation-free alternative to CT.
Keywords: pediatric, trauma, cervical spine, injury, MRI
Introduction
Cervical spine injury (CSI) occurs in approximately 1-2% of the pediatric blunt trauma population.[1, 2] There remains varied practice patterns, clinical uncertainty, and medicolegal insecurity regarding adequate evaluation of CSI and how to “clear the collar.” Due to the high sensitivity of modern protocols, a negative CT alone is considered adequate for most adult patients without incriminating neurologic symptoms or deficits, even if obtunded.[3] Evaluation requirements in pediatric populations are less clear given considerations of premature anatomy, unreliable communication, poor patient cooperation, and radiation exposure. Despite a stable incidence of pediatric CSI, the use of CT has increased substantially.[2]
The most current consensus statement provides algorithms for cervical spine clearance based on Glasgow Coma Scale (GCS) score in pediatric patients.[4] In short, clinical exam and plain radiographs are recommended for patients GCS ≥9. For patients with GCS ≤8, CT is recommended initially. However, even with negative CT, MRI is still recommended if mental status remains depressed. MRI of the cervical spine is also recommended for all patients who present with abusive head trauma (AHT), for forensic utility in addition to CSI screening. Recent literature in pediatric patients argues for perfect sensitivity of modern high-resolution CT for injuries that eventually require surgical intervention, therefore negating the need for MRI.[5–7] Such a practice is enticing because MRI is more expensive, has a longer acquisition time, often requires anesthesia and airway, and is thought to be riskier and have a higher false positive rate in the pediatric patient population. However, the literature is mixed.[8] MRI has traditionally been considered the most sensitive radiographic screen for CSI and, if done instead of CT, spares a vulnerable population from radiation exposure.
The purpose of this analysis is to explore the risks of cervical spine MRI in a pediatric blunt trauma population, which are not yet reported. We weigh these risks with other considerations of CT versus MRI in order to better assist providers in clinical decision-making, institutions in designing protocols, and expert panels in future guidelines.
Methods
Patient Selection
Patients were selected for inclusion from a prospectively maintained trauma database at our level 1 pediatric trauma center. Inclusion criteria were pediatric (18 years old or younger) blunt trauma patients who presented between January 1, 2012 – December 31, 2015 (four-year period) and underwent a cervical spine MRI during their admission. MRI is available around-the-clock at our institution and has generally been preferred over CT to spare radiation exposure in the pediatric population, but no rigid protocol was followed prospectively.
Data Collection
Most demographic and clinical variables were already prospectively collected for patients in the pediatric trauma database, but some variables were supplemented retrospectively. The final interpretation of the imaging by the attending neuro-radiologist was used as the default result in our analysis. However, patients with any question of abnormality on any modality had the entirety of their imaging reviewed by the study team. Injury mechanism was also recorded to classify patients into AHT and non-AHT groups, and the data from these groups is separated given that recommendations are differentiated in the most recently published consensus statement and algorithm.[4]
Patient charts were reviewed to determine if an airway was present for MRI, including laryngeal mask airway, endotracheal intubation, and nasotracheal intubation. Results were confirmed visually on the actual MRI images. Imaging and clinical notes were reviewed to determine if the airway was newly placed for MRI or was already in place for another reason (e.g., depressed mental status).
To determine the risks of MRI, clinical notes were reviewed to determine occurrence of a significant event related to either MRI itself or, if the patient required, MRI-related anesthesia and airway. Significant events included aspiration pneumonitis and ventilator-associated pneumonia (if the airway was newly placed for MRI), cardiac and respiratory arrests, and neurologic events (e.g., herniation event, ICP crisis). In addition to the clinical notes, chest imaging (x-ray or CT) in the seven days following MRI was reviewed for evidence of pneumonitis or pneumonia. Whether or not the patient had an MRI of another body part imaged with the cervical spine was also recorded. Timing of MRI from presentation was categorized as <24 hours, 24-48 hours, and >48 hours. For patients with abnormal MRI results, their charts were reviewed for long-term follow-up documentation.
Classifying and Comparing Imaging Abnormalities and Outcomes
Acute abnormalities related to the occipitocervical junction and cervical spine were included. Concordant with the recent literature[7], injuries were classified as clinically unstable if they required halo or operative intervention. Sensitivities of CT and MRI were determined for detection of clinically unstable injuries versus clinically stable or no injuries. For the patients who had positive MRI after negative CT, follow up records were reviewed for delayed need for cervical spine surgical intervention or halo.
Statistical Analysis
Statistical analysis in SPSS (IBM, Armonk, NY) was performed in consultation with a biostatistician. Chi square was used to compare proportions of categorical variables and one way ANOVA was used to compare means of continuous variables. A p value of 0.05 was used to denote significance.
Results
Airway and Anesthesia Needs for MRI
There were 269 patients included in the analysis (Table 1). Overall, 55% (149/269) of patients had an airway in place for MRI (Table 2), which correlated with GCS (p<0.001): 97% (36/37) of patients with GCS <9 had an airway in place compared to 42% (80/190) of patients with GCS 14-15. AHT patients were also more likely to have an airway in place for MRI (96%, 46/48; vs 47%, 103/221, p<0.001) than non-AHT patients.
Table 1.
Basic demographic and clinical variables, including injury mechanism
| All GCS | GCS <9 | GCS 9-13 | GCS 14-15 | ||||||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| Number | Percentage | Number | Percentage | Number | Percentage | Number | Percentage | ||
| Age (years) | |||||||||
| <3 | 98 | 36% | 21 | 57% | 20 | 48% | 57 | 30% | p=0.006* |
| 3to6 | 49 | 18% | 3 | 8% | 10 | 24% | 36 | 19% | |
| 7to12 | 75 | 28% | 6 | 16% | 10 | 24% | 59 | 31% | |
| >12 | 47 | 17% | 7 | 19% | 2 | 5% | 38 | 20% | |
| Median | 5 | 2 | 3 | 7 | |||||
| Mean | 6.2 | 4.8 | 4.2 | 6.9 | p=0.002* | ||||
|
| |||||||||
|
| |||||||||
| Sex | |||||||||
| Male | 160 | 59% | 21 | 57% | 30 | 71% | 109 | 57% | p<0.001* |
| Female | 109 | 41% | 16 | 43% | 12 | 29% | 81 | 43% | |
|
| |||||||||
|
| |||||||||
| Race | |||||||||
| White | 138 | 51% | 16 | 43% | 20 | 48% | 102 | 54% | p=0.529 |
| Black | 108 | 40% | 19 | 51% | 17 | 40% | 72 | 38% | |
| Other/Unknown | 23 | 9% | 2 | 5% | 5 | 12% | 16 | 8% | |
|
| |||||||||
|
| |||||||||
| Injury Mechanism | |||||||||
| AHT | 48 | 18% | 12 | 32% | 10 | 24% | 26 | 14% | p<0.001* |
| Fall | 100 | 37% | 4 | 11% | 12 | 29% | 84 | 44% | |
| Crush | 9 | 3% | 3 | 8% | 1 | 2% | 5 | 3% | |
| Occupant MVC | 47 | 17% | 5 | 14% | 10 | 24% | 32 | 17% | |
| Ped Struck | 30 | 11% | 7 | 19% | 6 | 14% | 17 | 9% | |
| Sports | 23 | 9% | 1 | 3% | 1 | 2% | 21 | 11% | |
| Other | 12 | 4% | 5 | 14% | 2 | 5% | 5 | 3% | |
|
| |||||||||
| 269 | 37 | 14% | 42 | 16% | 190 | 71% | |||
Basic demographic and clinical variables. Statistical correlations are assessed between GCS groups and age (first p value represents entire distribution of ages compared to GCS groups, second compares mean ages to GCS groups), sex, race, and injury mechanisms.
AHT = abusive head trauma. MVC = motor vehicle collision. Ped struck = pedestrian struck.
Table 2.
Airway and heavy sedation/anesthesia needs for MRI
| All GCS | GCS<9 | GCS 9-13 | GCS 14-15 | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||||||||||
| Number | Percentage | Number | Percentage | Number | Percentage | Number | Percentage | |||||||||||
| All | 269 | 37 | 14% | 42 | 16% | 190 | 71% | |||||||||||
| Non-AHT | 221 | 82% | 25 | 11% | 32 | 14% | 164 | 74% | ap=0.013* | |||||||||
| AHT | 48 | 18% | 12 | 25% | 10 | 21% | 26 | 54% | ||||||||||
|
| ||||||||||||||||||
|
| ||||||||||||||||||
| Airway for MRI | 149 | 55% | 36 | 97% | 33 | 79% | 80 | 42% | b p<0.001* | |||||||||
| Non-AHT | 103 | 47% | 24 | 96% | 23 | 72% | 56 | 34% | cp=0.934 | |||||||||
| AHT | 46 | 96% | 12 | 100% | 10 | 100% | 24 | 92% | ||||||||||
|
| ||||||||||||||||||
| dp<0.001 | ||||||||||||||||||
|
| ||||||||||||||||||
| New Airway | 85 | 57% | 5 | 14% | 14 | 42% | 66 | 83% | b p<0.001* | |||||||||
| Non-AHT | 61 | 59% | 3 | 13% | 11 | 48% | 47 | 84% | cp=0.715 | |||||||||
| AHT | 24 | 52% | 2 | 17% | 3 | 30% | 19 | 79% | ||||||||||
| dp=0.422 | ||||||||||||||||||
All patients who underwent MRI stratified by GCS group and AHT status. “Airway for MRI”: endotracheal tube, nasotracheal tube, or laryngeal mask airway in place at time of MRI. “New airway”: newly placed for MRI specifically.
Statistical correlations are assessed between GCS group and AHT status,
GCS group and airway needs,
GCS group and AHT status for patients requiring airway
AHT status and airway needs independent of GCS.
Overall, 57% (85/149) of patients with an airway had it newly placed for the MRI (Table 2), similar for AHT (52%, 24/46) and non-AHT (59%, 61/103) patients (p=0.42). As expected, better GCS correlated with needing new airway placement for MRI (p<0.001). For patients with GCS <9, 14% (5/36) had an airway newly placed, compared with 83% (66/80) of patients GCS 14-15.
Major Events Associated with Airway and MRI
There were no cases of aspiration pneumonitis or ventilator-associated pneumonia among patients who required a new airway for MRI (Table 3). There were no clinically significant neurological events while at MRI or related to anesthesia and airway placement for MRI. There was one cardiac arrest that occurred upon induction of anesthesia for airway placement; this event was in a AHT patient in the GCS <9 group who had presented to the hospital with respiratory arrest the day prior. This patient was high risk and may not represent the typical population. That being said, with this patient, the overall rate of major events for all patients was only 0.37% (1/269).
Table 3.
Risks and major events related to MRI and MRI-related anesthesia/airway
| All GCS (n=269) | GCS <9 (n=37) | |||||
|---|---|---|---|---|---|---|
|
| ||||||
| Number | Percentage | Number | Percentage | |||
| All Events | 1 | 0.37% | 1 | 2.70% | a p=0.043* | |
| Non-AHT | 0 | 0% | 0 | 0% | ||
| AHT | 1 | 2.08% | 1 | 8.33% | ||
|
| ||||||
| bp=0.64 | ||||||
|
| ||||||
| Arrests | 1 | 0.37% | 1 | 2.70% | ||
| Non-AHT | 0 | 0% | 0 | 0% | ||
| AHT | 1 | 2.08% | 1 | 8.33% | ||
|
| ||||||
|
| ||||||
| VAP/Aspiration | 0 | 0% | 0 | 0% | ||
|
| ||||||
|
| ||||||
| Neurologic Events | 0 | 0% | 0 | 0% | ||
The clinically significant risks and major events are stratified by AHT status and GCS group. Only the GCS <9 group is shown given no events in the other groups. “Arrests” includes both cardiac and respiratory events. “Neurologic events” includes ICP crises, strokes, and herniation events.
Statistical correlations are assessed between GCS group and event rate
AHT status and event rate independent of GCS.
VAP = Ventilator-associated pneumonia.
Timing of MRI and Concurrent Studies
Sixty-two percent (167/269) of MRIs were obtained within 24 hours of presentation and 83% (224/269) were performed within 48 hours (Table 4). Higher GCS correlated with faster MRI (p<0.001), as did non-AHT status (p=0.01). Sixty-four percent (173/269) of patients had another body part imaged via MRI (e.g., brain, thoracic, and/or lumbar spine) with the cervical spine. Concurrent MRI rates were higher with lower GCS score (p<0.001) and AHT injury mechanism (p<0.001).
Table 4.
Timing of MRI and Concurrent Studies
| Number | Percentage | Number | Percentage | Number | Percentage | Number | Percentage | ||
|---|---|---|---|---|---|---|---|---|---|
|
| |||||||||
| ALL PATIENTS | All GCS (n=269) | GCS <9 (n=37) | GCS 9-13 (n=42) | GCS 14-15 (n=190) | |||||
| Timing of MRI | |||||||||
| <24hr | 167 | 62% | 16 | 43% | 22 | 52% | 129 | 68% | ap<0.001 |
| 24-48hr | 57 | 21% | 6 | 16% | 11 | 26% | 40 | 21% | |
| >48hr | 45 | 17% | 15 | 41% | 9 | 21% | 21 | 11% | |
|
| |||||||||
|
| |||||||||
| Concurrent MRI | |||||||||
| Yes | 173 | 64% | 34 | 92% | 30 | 71% | 109 | 57% | bp<0.001 |
| No | 96 | 36% | 3 | 8% | 12 | 29% | 81 | 43% | |
|
| |||||||||
| NON-AHT | All GCS (n=221) | GCS <9 (n=25) | GCS 9-13 (n=32) | GCS 14-15 (n=164) | |||||
|
| |||||||||
| Timing of MRI | |||||||||
| <24hr | 141 | 74% | 10 | 40% | 17 | 53% | 114 | 70% | cp=0.003 |
| 24-48hr | 50 | 23% | 6 | 24% | 10 | 31% | 34 | 21% | |
| >48hr | 30 | 14% | 9 | 36% | 5 | 16% | 16 | 10% | |
|
| |||||||||
| dp=0.01 | |||||||||
|
| |||||||||
| Concurrent MRI | |||||||||
| Yes | 128 | 58% | 22 | 88% | 21 | 66% | 85 | 52% | cp<0.001 |
| No | 93 | 42% | 3 | 12% | 11 | 34% | 79 | 48% | |
|
| |||||||||
| ep<0.001* | |||||||||
| AHT | All GCS (n=48) | GCS <9 (n=12) | GCS 9-13 (n=10) | GCS 14-15 (n=26) | |||||
|
| |||||||||
| Timing of MRI | |||||||||
| <24hr | 26 | 54% | 6 | 50% | 5 | 50% | 15 | 58% | cp=0.191 |
| 24-48hr | 7 | 15% | 0 | 0% | 1 | 10% | 6 | 23% | |
| >48hr | 15 | 31% | 6 | 50% | 4 | 40% | 5 | 19% | |
|
| |||||||||
|
| |||||||||
| Concurrent MRI | |||||||||
| Yes | 45 | 94% | 12 | 100% | 9 | 90% | 24 | 92% | cp=0.568 |
| No | 3 | 6% | 0 | 0% | 1 | 10% | 2 | 8% | |
The proportions of patients undergoing MRI at different time frames are stratified by GCS group and AHT status. Also shown are the proportions of patients in each group who had another MRI concurrently with their cervical spine study.
Statistical correlations are assessed between GCS group and timing both independent of AHT statusa and among AHTc and non-AHTb patients. Independent of GCS, comparison is made between between AHT status and timingd and AHT status and concurrent MRIe.
Sensitivity of CT and MRI
There were three patients who had clinically unstable injuries, two requiring halo placement and one requiring operative intervention. Both CT and MRI were 100% sensitive in detecting these three cases of clinically unstable injuries (Table 5).
Table 5.
Sensitivity of CT and MRI for Clinically Unstable Injuries
| Clinically Unstable Injury | Clinically Stable or No Injury | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||
| Number | Percentage | Number | Percentage | Sensitivity | Specificity | PPV | NPV | |||
| CT | ||||||||||
| Positive | 3 | 11% | 25 | 89% | CT | 100% | 77% | 11% | 100% | |
| Negative | 0 | 0% | 83 | 100% | ||||||
|
| ||||||||||
| MRI | ||||||||||
| Positive | 3 | 8% | 35 | 92% | MRI | 100% | 87% | 8% | 100% | |
| Negative | 0 | 0% | 231 | 100% | ||||||
Diagnostic characteristics of cervical spine CT and MRI for the detection of clinically unstable injuries (i.e., requiring surgery or halo fixation).
Outcomes After Negative CT
Thirty-one percent of patients (83/269) had a negative CT prior to obtaining a subsequent MRI; 89% (74/83) of these patients then had a negative MRI (Table 6). Commensurate with the perfect sensitivity of CT, no patients with a previous negative CT were later found to have a clinically unstable injury. After negative CT, 11% (9/83) of patients had a clinically stable injury detected on subsequent MRI. This did not correlate with GCS (p=0.597). All were treated with cervical collar. None of these 9 patients underwent delayed surgery or halo placement. Sixty-seven percent (56/83) of patients with a prior negative CT had another body part imaged on MRI besides the cervical spine.
Table 6.
Outcome after Negative CT
| ALL PATIENTS | Number | Percentage | Number | Percentage | Number | Percentage | Number | Percentage | ||
|---|---|---|---|---|---|---|---|---|---|---|
|
| ||||||||||
| All GCS | (n=83) | GCS <9 | (n=17) | GCS 9-13 | (n=12) | GCS 14-15 | (n=54) | |||
| Negative MRI | 74 | 89% | 14 | 82% | 11 | 92% | 49 | 91% | p=0.597 | |
| Clinically Stable Injury | 9 | 11% | 3 | 18% | 1 | 8% | 5 | 9% | ||
| Clinically Unstable Injury (Halo or Surgery) | 0 | 0% | 0 | 0% | 0 | 0% | 0 | 0% | ||
| Airway for MRI | 42 | 51% | 16 | 94% | 8 | 67% | 18 | 33% | p<0.001* | |
| New Airway for MRI | 15 | 35% | 2 | 13% | 2 | 25% | 11 | 61% | p<0.001* | |
| VAP/Aspiration | 0 | 0% | 0 | 0% | 0 | 0% | 0 | 0% | - | |
| Concurrent MRI | 56 | 67% | 15 | 88% | 7 | 58% | 34 | 63% | p=0.117 | |
|
| ||||||||||
| NON-AHT | All GCS | (n=71) | GCS <9 | (n=13) | GCS 9-13 | (n=11) | GCS 14-15 | (n=47) | ||
|
| ||||||||||
| Negative MRI | 62 | 87% | 10 | 77% | 10 | 91% | 42 | 89% | p=0.455 | |
| Clinically Stable Injury | 9 | 13% | 3 | 23% | 1 | 9% | 5 | 11% | ||
| Airway for MRI | 30 | 42% | 12 | 92% | 7 | 64% | 11 | 23% | p<0.001* | |
| New Airway for MRI | 10 | 14% | 1 | 8% | 2 | 18% | 7 | 15% | p=0.018* | |
| Concurrent MRI | 44 | 62% | 11 | 85% | 6 | 55% | 27 | 57% | p=0.174 | |
|
| ||||||||||
| AHT | All GCS | (n=12) | GCS <9 | (n=4) | GCS 9-13 | (n=1) | GCS 14-15 | (n=7) | ||
|
| ||||||||||
| Negative MRI | 12 | 100% | 4 | 100% | 1 | 100% | 7 | 100% | - | |
| Clinically Stable Injury | 0 | 0% | 0 | 0% | 0 | 0% | 0 | 0% | - | |
| Airway for MRI | 12 | 100% | 4 | 100% | 1 | 100% | 7 | 100% | - | |
| New Airway for MRI | 5 | 42% | 1 | 25% | 0 | 0% | 4 | 57% | p=0.394 | |
| Concurrent MRI | 12 | 100% | 4 | 100% | 1 | 100% | 7 | 100% | - | |
A subset analysis of the 83 patients who had a negative CT and subsequent positive or negative MRI, stratified by GCS group and AHT status. Statistical comparisons are made between GCS groups for each variable both independent of AHT status and within each AHT status subgroup. “Airway for MRI”: an airway was in place for any reason. “New airway for MRI”: an airway was placed for obtaining MRI. “Concurrent MRI”: another region (e.g., brain, thoracic or lumbar spine) was imaged during the same acquisition as the cervical spine.
Discussion
Risks of MRI
The risks associated with obtaining a cervical spine MRI in the pediatric blunt trauma patient population are not well known. Previous articles have cited hospital-wide MRI risks and have combined both major events and smaller events, such as an infiltrated IV. One such report produced a total incident rate of only 0.62% among all of pediatric radiology.[9] A similar study for all of pediatric MRI reported a rate of 0.02%.[10] For the cervical spine specifically, Yecies et al. looked at safety of dynamic spine MRI in pediatrics and there were no events among 66 acquisitions.[11]
We sought to identify the rate of major events that would factor into clinical decision-making. The overall rate of 0.37% was lower than expected and represented only one event in the series of 269 patients. Admittedly, risk of MRI is mitigated because patients perceived to be at higher risk would presumably have their imaging delayed until safer. Even so, 62% of MRIs were obtained within 24 hours and 83% within 48 hours. Also of note, 67% of patients had another body part imaged via MRI concurrently and therefore the risk is not solely born by need to evaluate the cervical spine. Around half of patients (55%) had an airway in place for MRI, but only 57% of these were newly placed, meaning that only 31% of patients required a new airway for MRI. Further, only 36% (31/85) of these patients with a new airway had cervical spine MRI alone.
There are also suspected longer term neurodevelopmental effects of many common anesthetic agents (recently reviewed).[12] The current U.S. Food and Drug Administration advisory warns of increased anesthesia risks for repeated procedures and those >3 hours in children <3 years of age.[13] This risk also should certainly be considered in the risk-benefit calculation of obtaining an MRI.
Comparison of CT and MRI
While MRI is critical when there are neurological symptoms or deficits and when an injury on x-ray or CT needs further characterization, the added value is debated in other circumstances. Depending on patient scenario and clinical exam, flexion-extension films and consultation of the spine service may be considered to either clear the spine without MRI or determine who may benefit from imaging. When mental status is depressed, expert consensus recommends cervical spine CT in place of or after x-ray given the excellent sensitivity, speed, and lower cost of CT.
Recent analyses demonstrate that there are no frankly unstable injuries requiring halo orthosis or surgical intervention after a negative modern high-quality CT.[5–7] Indeed, in our analysis, there were no patients with a prior negative CT who went on to have surgery or halo placement. Both CT and MRI were 100% specific for clinically unstable injuries. However, Al-Sarheed et al. reported a 62 patient series wherein 5 children had an abnormal MRI after normal CT: one who underwent occipitocervical instrumented fusion and the rest managed by external orthosis.[8] This raises the question of whether MRI might be preferred when pediatric radiologists are not available, and if CT should only replace MRI in certain centers.
Given that MRI has traditionally been considered the most sensitive modality for detecting CSI, and the burden has been placed on CT to match, one wonders in what circumstances MRI may be still preferred to CT. Radiation exposure is of particular importance in pediatrics, where the secondary effects of radiation are more significant. One recent analysis calculated a median effective radiation dose for CT in this setting at around 5 mSv,[15] which amounts to an estimated 1 additional cancer per 2000 scans.[16] While x-ray does confer radiation exposure, CT is estimated to deliver 90-200 times more, and specifically 200 times more to the thyroid.[17, 18] Banerjee and Thomas reported a mean radiation dose of 20.3 mSv in their recent series of CT for pediatric CSI, resulting in a mean lifetime risk of CT scan-induced malignancy at 0.37%.[19] Connelly et al. was able to decrease rate of CT use for CSI evaluation from 30% to 13% with a performance improvement and patient safety program, reducing estimated lifetime attributable risk of any cancer by 22-38% and of thyroid cancer by 54%. [20]
Despite a movement toward CT as the initial screening tool, cervical spine xray followed by MRI remains a viable alternative to reduce radiation. Several recent reports have demonstrated excellent sensitivity of xray for CSI.[21, 22] Lindholm et al. showed that even a single lateral view xray was equivalent to multiple views and had 100% sensitivity for bony injuries.[23] Another concern of MRI is that of “false positives,” where minor injuries on imaging may prompt maintenance of a cervical collar for prolonged periods while inpatient or on discharge. In our series, 11% (9/83) of patients with a negative CT were deemed to have a clinically stable injury on subsequent MRI. All were treated with cervical collar and none underwent delayed surgery or halo immobilization. One way to ameliorate unnecessary collaring is to expand education of providers about the true need of collar placement, rather than eliminating MRI use for fear of false positives.
Limitations
As a single-institution study, our results are subject to the biases of the providers (e.g., imaging protocols), resources available (e.g., easily available MRI), and patient population at our institution. Other providers and institutions will need to factor in availability of resources into the decision calculation. For example, these findings may not be generalizable to an adult-oriented trauma facility. This study is also retrospective, which could affect the accuracy of data generally, but in particular of rates of major events as they may have not been documented in the chart. However, we believe the chance of a major event not being noted is extremely low.
At our institution, the general protocol is for the trauma service to clear the cervical spine if no clinical concerns exist, as they are the primary service by default. Orthopedic surgery or neurosurgery are consulted when there is concern. In general, due to radiation concerns, MRI is favored over CT in the pediatric population at our institution and as such it is the more standard scan, although patients who came from other institutions may have come with CT scans. Physical exam and patient context was not always recorded in the chart at the time of the imaging decision; therefore, the indication for obtaining an MRI in a patient with a prior negative CT was not always clear. We suspect that many of these patients may have been experiencing neck pain and/or neurological symptoms, prompting further MRI evaluation. We hope that forthcoming prospective studies will control for these important variables. However, we do not believe that this limitation detracts from a novel aspect of this manuscript, which is the reporting of risk associated with MRI.
Conclusion
The purpose of this analysis was to define the rate of major risks of cervical spine MRI in the evaluation of pediatric blunt trauma patients. A single major event related to MRI occurred among 269 patients over a four-year period—a rate of 0.37%. Only about 1/3 of patients required a new airway to be placed for MRI given that many patients were already intubated beforehand for other reasons. In addition, most patients had another part of the body imaged via MRI concurrently, and only 36% of patients were intubated for cervical spine imaging only. None of the newly intubated patients developed aspiration pneumonitis or ventilator-associated pneumonia and no patients had a significant neurological event. Most MRIs were obtained within 48 hours.
The process of clearing the cervical spine can be complex for determining a one-size-fits-all algorithm due institutional differences and relative risks that must be weighed in terms of imaging sensitivity, imaging time, radiation exposure, and risks of anesthesia—both short and long-term. Overall, the safety profile of MRI is good, and it remains highly sensitive for detection of CSI. Both CT and MRI demonstrated 100% sensitivity to detect clinically unstable injuries. After negative CT, injuries that result in placement of a cervical collar are still diagnosed on MRI, although none of these patients re-presented for delayed surgical intervention or halo after discharge. While our data support the existing literature that negative CT likely rules out an injury requiring any intervention other than a collar, these points, in conjunction with attention to sparing the pediatric population from radiation, suggest continued consideration of MRI as an alternative to CT.
Funding Sources
This publication was made possible by the Johns Hopkins Institute for Clinical and Translational Research (ICTR) which is funded in part by Grant Number UL1 TR003098 from the National Center for Advancing Translational Sciences (NCATS) a component of the National Institutes of Health (NIH), and NIH Roadmap for Medical Research. Its contents are solely the responsibility of the authors and do not necessarily represent the official view of the Johns Hopkins ICTR, NCATS or NIH.
Footnotes
Statement of Ethics
Study Approval Statement: This study was reviewed and approved by the Johns Hopkins University Institutional Review Board, approval number IRB00098135.
Consent to participate statement: The need for informed consent by participants was waived by the Johns Hopkins University Institutional Review board given low risk to the subjects and the publication of only aggregate, de-identified data.
Conflict of Interest Statement
The authors report no relevant conflicts of interest other than that Dr. Mari Groves is an Editorial Board Member of Pediatric Neurosurgery.
Data Availability Statement
Data cannot be shared for confidentiality reasons. Queries about the data should be directed to the corresponding author.
References
- 1.Viccellio P, Simon H, Pressman BD, Shah MN, Mower WR, Hoffman JR, et al. A prospective multicenter study of cervical spine injury in children. Pediatrics. 2001. Aug;108(2):E20. [DOI] [PubMed] [Google Scholar]
- 2.Adelgais KM, Browne L, Holsti M, Metzger RR, Murphy SC, Dudley N. Cervical spine computed tomography utilization in pediatric trauma patients. J Pediatr Surg. 2014. Feb;49(2):333–7. [DOI] [PubMed] [Google Scholar]
- 3.Patel MB, Humble SS, Cullinane DC, Day MA, Jawa RS, Devin CJ, et al. Cervical spine collar clearance in the obtunded adult blunt trauma patient: a systematic review and practice management guideline from the Eastern Association for the Surgery of Trauma. J Trauma Acute Care Surg. 2015. Feb;78(2):430–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Herman MJ, Brown KO, Sponseller PD, Phillips JH, Petrucelli PM, Parikh DJ, et al. Pediatric Cervical Spine Clearance: A Consensus Statement and Algorithm from the Pediatric Cervical Spine Clearance Working Group. J Bone Joint Surg Am. 2019. Jan;101(1):e1. [DOI] [PubMed] [Google Scholar]
- 5.Gargas J, Yaszay B, Kruk P, Bastrom T, Shellington D, Khanna S. An analysis of cervical spine magnetic resonance imaging findings after normal computed tomographic imaging findings in pediatric trauma patients: ten-year experience of a level I pediatric trauma center. J Trauma Acute Care Surg. 2013. Apr;74(4):1102–7. [DOI] [PubMed] [Google Scholar]
- 6.Qualls D, Leonard JR, Keller M, Pineda J, Leonard JC. Utility of magnetic resonance imaging in diagnosing cervical spine injury in children with severe traumatic brain injury. J Trauma Acute Care Surg. 2015. Jun;78(6):1122–8. [DOI] [PubMed] [Google Scholar]
- 7.Derderian SC, Greenan K, Mirsky DM, Stence NV, Graber S, Hankinson TC, et al. The utility of magnetic resonance imaging in pediatric trauma patients suspected of having cervical spine injuries. J Trauma Acute Care Surg. 2019. 12;87(6):1328–35. [DOI] [PubMed] [Google Scholar]
- 8.Al-Sarheed S, Alwatban J, Alkhaibary A, Babgi Y, Al-Mohamadi W, Masuadi EM, et al. Cervical spine clearance in unconscious pediatric trauma patients: a level l trauma center experience. Childs Nerv Syst. 2020. 04;36(4):811–17. [DOI] [PubMed] [Google Scholar]
- 9.Snyder EJ, Zhang W, Jasmin KC, Thankachan S, Donnelly LF. Gauging potential risk for patients in pediatric radiology by review of over 2,000 incident reports. Pediatr Radiol. 2018. 12;48(13):1867–74. [DOI] [PubMed] [Google Scholar]
- 10.Jaimes C, Murcia DJ, Miguel K, DeFuria C, Sagar P, Gee MS. Identification of quality improvement areas in pediatric MRI from analysis of patient safety reports. Pediatr Radiol. 2018. 01;48(1):66–73. [DOI] [PubMed] [Google Scholar]
- 11.Yecies D, Fogel N, Edwards M, Grant G, Yeom KW, Cheshier S. Safety of Dynamic Magnetic Resonance Imaging of the Cervical Spine in Children Performed without Neurosurgical Supervision. World Neurosurg. 2018. Aug;116:e1188–e93. [DOI] [PubMed] [Google Scholar]
- 12.Andropoulos DB. Effect of Anesthesia on the Developing Brain: Infant and Fetus. Fetal Diagn Ther. 2018;43(1):1–11. [DOI] [PubMed] [Google Scholar]
- 13.Administration USFaD. FDA approves label changes for use of general anesthetic and sedation drugs in young children. 2017.
- 14.Hassan N, Butler C, DeCou J, Crumb T, Flohr S, Reischman D, et al. Pediatric cervical spine injuries on CT: difference in accuracy of interpretations by pediatric versus non-pediatric radiologists. Emerg Radiol. 2020. Apr;27(2):185–90. [DOI] [PubMed] [Google Scholar]
- 15.Somppi LK, Frenn KA, Kharbanda AB. Examination of Pediatric Radiation Dose Delivered After Cervical Spine Trauma. Pediatr Emerg Care. 2018. Oct;34(10):691–95. [DOI] [PubMed] [Google Scholar]
- 16.Smith-Bindman R, Lipson J, Marcus R, Kim KP, Mahesh M, Gould R, et al. Radiation dose associated with common computed tomography examinations and the associated lifetime attributable risk of cancer. Arch Intern Med. 2009. Dec;169(22):2078–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Fenton SJ, Hansen KW, Meyers RL, Vargo DJ, White KS, Firth SD, et al. CT scan and the pediatric trauma patient--are we overdoing it? J Pediatr Surg. 2004. Dec;39(12):1877–81. [DOI] [PubMed] [Google Scholar]
- 18.Jimenez RR, Deguzman MA, Shiran S, Karrellas A, Lorenzo RL. CT versus plain radiographs for evaluation of c-spine injury in young children: do benefits outweigh risks? Pediatr Radiol. 2008. Jun;38(6):635–44. [DOI] [PubMed] [Google Scholar]
- 19.Banerjee P, Thomas M. CT scans to exclude spine fractures in children after negative radiographs may lead to increase in future cancer risk. Eur J Orthop Surg Traumatol. 2019. Jul;29(5):983–88. [DOI] [PubMed] [Google Scholar]
- 20.Connelly CR, Yonge JD, Eastes LE, Bilyeu PE, Kemp Bohan PM, Schreiber MA, et al. Performance improvement and patient safety program-guided quality improvement initiatives can significantly reduce computed tomography imaging in pediatric trauma patients. J Trauma Acute Care Surg. 2016. 08;81(2):278–84. [DOI] [PubMed] [Google Scholar]
- 21.Brockmeyer DL, Ragel BT, Kestle JR. The pediatric cervical spine instability study. A pilot study assessing the prognostic value of four imaging modalities in clearing the cervical spine for children with severe traumatic injuries. Childs Nerv Syst. 2012. May;28(5):699–705. [DOI] [PubMed] [Google Scholar]
- 22.Cui LW, Probst MA, Hoffman JR, Mower WR. Sensitivity of plain radiography for pediatric cervical spine injury. Emerg Radiol. 2016. Oct;23(5):443–8. [DOI] [PubMed] [Google Scholar]
- 23.Lindholm EB, Malik A, Parikh D, Mamdouhi T, Alper L, Nanassy A, et al. Single-lateral cervical radiograph in pediatric trauma is equivalent to multiple views. J Trauma Acute Care Surg. 2019. 10;87(4):813–17. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data cannot be shared for confidentiality reasons. Queries about the data should be directed to the corresponding author.