Validation of a shortened MR imaging protocol for pediatric spinal pathology

Conventional pediatric spine MRI protocols have multiple sequences resulting in long acquisition times. Sedation is consequently required. This study evaluates the diagnostic capability of a limited MRI spine protocol for selected common pediatric indications. Spine MRIs at CHEO between 2017 and 2020 were reviewed across pediatric patients younger than four years old. Two blinded neuroradiologists reviewed limited scan sequences, and results were independently compared to previously reported findings from the complete imaging series. T2 sagittal sequences from the craniocervical junction to sacrum and T1 axial sequence of the lumbar spine constitute the short protocol, with the outcomes of interest being cerebellar ectopia, syrinx, level of conus, filum < 2 mm, fatty filum, and spinal dysraphism. A total of 105 studies were evaluated in 54 male and 51 female patients (mean age 19.2 months). The average combined scan time of the limited sequences was 15 min compared to 35 min for conventional protocols (delta = 20 min). The average percent agreement between full and limited sequences was > 95% in all but identifying a filum < 2 mm, where the percent agreement was 87%. Using limited MR sequences had high sensitivity (> 0.91) and specificity (> 0.99) for the detection of cerebellar ectopia, syrinx, fatty filum, and spinal dysraphism. This study demonstrates that selected spinal imaging sequences allow for consistent and accurate diagnosis of specific clinical conditions. A limited spine imaging protocol has potential as a screening test to reduce the need for full-sequence MRI scans. Further work is needed to determine utility of selected imaging for other clinical indications.


Introduction
MRI is the imaging modality of choice for many pediatric nervous system pathologies.The acquisition of highquality sequences often requires long study times and that patients remain immobile.Young children and patients with developmental delay or mental health concerns may struggle with evaluation and routinely require anesthesia to perform traditional MRI scans [1][2][3][4].While sedating techniques in pediatric patients have greatly improved over time, they remain not without risk [1][2][3][4][5].In addition, the need for general anesthetic and airway support further risks aerosolizing pathogens that may endanger healthcare providers and has recently become a point of much greater concern.[6] Simplified imaging protocols have the potential to reduce the overall duration of studies and potentially obviates the need and inherent risks of sedation.Fast-brain MR imaging has been thoroughly described, and it is actively being used alongside conventional brain MRI protocols to monitor abnormalities such as hydrocephalus, macrocephaly, Chiari malformation, intracranial cysts, and ventricular shunt malfunction [7][8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23].The current literature on rapid spine MRI protocols is very limited and has not been validated in pediatric populations [15][16][17][18][19][20][21][22][23].The aim of this study was to

Imaging acquisition and image selection
Studies were acquired on 1.5 T (Sigma HDxt, General Electric Healthcare Technologies, Waukesha, WI, USA 16.0) and 3 T MRI scanners (Magnetom Skyra, Siemens Healthcare, Erlangen, Germany) using routine departmental protocols.The standard full spine MRI protocol included upper sagittal T1, upper sagittal T2, lower sagittal T1, lower sagittal T2, upper axial T1, upper axial T2, lower axial T1, and lower axial T2 to image the entire spine.The limited spine MRI protocol was developed by consensus opinion of the lead authors (EM, JHM, and AT) and included a sagittal T2-weighted image of the entire spine and an axial T1-weighted image of the lumbosacral spine (Table 1).

Limited protocol determination
Consensus opinion (AT, EM, JHM, and DM) was used to determine findings with potentially greatest yield for evaluation with limited spinal MRI.Features including patient age range most likely to be intolerant of long studies, conditions where screening could potentially circumvent further evaluation, need for repeat imaging, and commonality of clinical conditions were all considered.Features for specific review on limited sequence were as follows: the presence or absence of Chiari I malformation (tonsillar ectopia > 5 mm), syrinx, spinal level of conus medullaris (n), presence of fatty infiltration of the filum, normal filum (< 2 mm width), and the presence or absence of any type of spinal dysraphism.Features were scored for presence on a binary basis (y/n) with the exception of the presence of spinal dysraphism which was scored for presence (y/n) as well as qualitatively (i.e., description).

Imaging analysis
The limited MRI sequences were independently reviewed by two senior pediatric neuroradiologists (EM and JH each with > 20 years each of experience).Reviewers were blinded to the pre-existing report on the complete study, as well as any sequences other than those included in the limited protocol.Ten studies were reviewed by both imaging reviewers to ensure inter-rater reliability.Each reader then reviewed 50 MRI studies.The pre-existing radiology reports were reviewed independently by authors (WW and MT) who were blinded to the findings of the imaging reviewers.Ten patient records were reviewed by both the reviewers to ensure reasonable inter-rater reliability.Further information in the form of patient demographics (age at time of study and gender), sedation status (none vs. general anesthesia), and clinical indication for investigation were also recorded.Study data were collected and managed using REDCap electronic data capture tools hosted at CHEO.Discrepancies between the findings of the limited protocol with full MRI spine reports were formally reviewed by the authors and reconciled by group consensus.Discrepancies were classified as resulting from misdiagnosis or "true misses' on limited protocol, or if they resulted from misinterpretation of the written report.Errors resulting from report misinterpretation were reviewed by lead authors (EM and AT) and recategorized appropriately.

Statistical methods
Sample size was predetermined assuming a 10% clinical difference between limited and full-scan sequences.For the percent agreement metric, a sample size of 105 studies was adequate to determine a difference, at a 95% CI level.PRISM and Microsoft Excel were used to carry out statistical analysis as appropriate.

Results
A total of 128 patient MRI records met our search criteria.Twenty patients were excluded as a contrast agent (gadolinium) was administered, and the last 3 studies were excluded as we had met our target sample size.The 105 MRI studies that fit the inclusion criteria were performed in 54 male and 51 female patients.Patients scanned with anesthesia accounted for 67% (70/105; 95% CI: 57-75), and patients scanned without accounted for 33% (35/105; 95% CI: 25-43) of this cohort.The mean age at time of MRI scan was 19.2 months (range: 0.13-46.8months) with an average age of 27.8 and 1.44 months for the with and without anesthesia groups.The most common clinical indications for spinal imaging were workup for sacral dimple, lipomyelomeningocele, low lying conus, lipoma, spinal dysraphism, constipation/urinary incontinence, and neuroblastoma (Table 2).

Discussion
Precedence for utilizing limited series imaging in pediatric populations already currently exists in the form of "fast" or "limited" imaging of the brain.Prior studies have investigated the diagnostic equivalency of T2, FLAIR, diffusion-weighted imaging (DWI), echo planar imaging, T1-weighted 3D spoiled gradient echo, and fast imaging employing steady-state acquisition compared to conventional MRI imaging [7][8][9][10][11][12][13][14].While these studies are only validated in the evaluation of patients with shunted hydrocephalus, they have found a vital role in avoiding repeated doses of ionizing radiation or prolonged periods of time inside a MRI machine [20].The use of shortened imaging protocols for MRI of the spine is not a novel concept and has been previously reported in limited populations.Several studies have accepted the use of less information in order to answer a specific clinical question.In 1996, Robertson et al. utilized two single-effective long TR fast echo acquisition sequences in the sagittal and axial planes for imaging of lumbar spondylosis and achieved a reduction in total scan time from 28 to 2.5 min.[24] More recently, Gewirtz et al. developed an abbreviated pediatric spine MRI protocol to reduce acquisition time from 60 to 2 min with rapid T1 and T2 imaging.These authors reported that in 43 of the 47 rapid imaging studies, the fast spine protocol sufficiently diagnosed syrinx and spinal dysraphism and ruled out other asymptomatic pathologies.However, this study was not paired and thus could  Moreover, only gross pathologies such as syringomyelia and spinal dysraphism were assessed [25].In this study, we showed that in specific clinical conditions, we can use fewer sequences to make the same diagnosis and evaluation of common pediatric spine pathologies compared to the gold standard of a full MRI study.This high diagnostic accuracy of using a limited series is encouraging for use in the screening for common spine pathologies.Currently available screening techniques such as spinal ultrasound are limited by anatomical restrictions (i.e., extent of ossification with age) as well as inconsistent visualization of spine structures.While definitive management for many pathologies may require more detailed studies of greater resolution, limited imaging assists in the detection of those patients who would require no further intervention and avoid more involved investigation altogether.Even after an initial diagnosis has been made, there are many instances when routine follow up imaging may be required (i.e., monitoring for syrinx progression in patients with Chiari I malformation) and sufficient information could be obtained using only limited imaging.In addition to the absolute reduction in time of the scan itself, there is potential circumvention of the need for anesthetic which further reduces the duration a given patient may spend in the radiology department.These consequences may further improve departmental throughput and reduce overall patient wait times.
One challenge that was observed in this study was the effect of expectation bias during the evaluation of the limited spine protocol.We found several discrepancies owing to missed identification during the reporting of the full imaging studies whereas they were detected during review of the limited studies.These discrepancies do not detract from the efficacy of limited imaging, but rather speak to the importance of anticipation and knowledge of clinical indication when reviewing sequences.Furthermore, we identified that identified a thickened filum (i.e., filum > 2 mm) may not be as consistent using limited sequences and, thus, may specifically require more dedicated imaging.A major strength of this study is the reliance on an independent, blind comparison of a novel protocol with conventional spine imaging.This comparison allows for true validation of the use of this limited protocol in a number of common pediatric pathologies.Further, we found that magnet strength did not significantly affect diagnostic capability, suggesting that a limited imaging would be feasible in centers where only lower field strength MRI were available.While we selected pathologies based on common indications for imaging, it is very likely that limited protocol would also be of utility for other conditions.

Conclusion
This study demonstrates that the use of selected spine imaging sequences provides consistent and accurate diagnosis for several common pediatric conditions.This limited protocol may be very effective as a screening tool or in routine follow-up after an initial diagnosis has been made.Further prospective study is required to determine the utility of selected imaging for other potential indications.

Fig. 1
Fig. 1 Average % agreement of full-sequence MRI vs. limited spine imaging based on magnet strength.This figure shows the average percentage agreement of the limited spine series compared to full-sequence MRI across 6 different clinical indications based on magnet strength: 1.5 T (white bars) and 3 T (gray bars)

Table 1
Sequences and duration of sequence per study

Table 2
Clinical indications for spinal imagingThese were the possible clinical diagnoses for workup with imaging but did not necessarily denote the presence of the indication

Table 5
Discrepancy rate and type between full and limited protocol