Hi, my name is Christina Sayama, and I'll be presenting the Diagnostic Accuracy of an Abbreviated STIR MRI Sequence in Pediatric Spine Trauma. We have no disclosures. Children's spinal anatomy differs from adults, and in trauma, bony X-rays and CTs are often done first line, but suboptimal for imaging and ligamentous concerns. MRI then becomes critical for spine evaluation, but may be difficult to obtain in uncooperative or critically injured children. Spinal MRIs require long scan times, and children often require sedation and need to lay flat, which can be suboptimal in concurrent TBI. Our study proposes to evaluate an Abbreviated Sagittal STIR Sequence, or QuickSTIR MRI, in the setting of pediatric spine trauma, and determine the sensitivity, specificity, and inter-rater reliability compared to the standard STIR MRI. This was a retrospective study looking at patients less than 18-year-old with concern for CT or L-spine trauma over an 18-month period. All had QuickSTIR imaging and regular STIR imaging done as part of a routine study. Exclusion criteria were spine MRIs done for reasons other than trauma. We created an MRI imaging protocol with neuroradiology and tested our Abbreviated STIR Sequence. Two blinded, board-certified neuroradiologists then reviewed de-identified regular STIR images and QuickSTIR images, and findings were tabulated by vertebral level for the following injury types. A consensus review was held to resolve discrepancies. Statistical analysis was done with SPSS. We had 21 patients with 31 MRI studies. Five of those were excluded due to missing data. We ended up with 18 cervical, 5 thoracic, and 3 lumbar spine MRIs. 9 of the 21 patients had clinically significant injuries that required treatment with long-term cervical collar or TLSO bracing. None of our patients required surgery. Table 1 shows that the average QuickSTIR MRI took 2 minutes and 19 seconds, compared to the normal 3 minutes and 54 seconds. Table 2 shows the number of times an injury pattern was identified on our standard STIR sequences. We did the same counts for our QuickSTIR sequences and calculated sensitivity and specificity. As you can see, specificity was very high for QuickSTIR, but sensitivity was variable ranging from 50% to 100% based on the injury type. We therefore further broke down our study to look at vertebral regions. This is a busy slide, but I want to point out from Ossiput C2, we noted a higher sensitivity and specificity that was better than all other regions and for all injury patterns. We had a sensitivity ranging now from 83% to 100%. We also did the same looking at clinically significant injuries and found that our sensitivity and specificity increased to 100% with the exception of ligament disruption slash rupture, which only increases 67%. When looking at the inter-rater reliability, we see that for the QuickSTIR imaging, we had fair agreement in a majority of injury types. And then for everything else, there was moderate to substantial agreement across both studies. Both QuickSTIR and regular STIR had IRRs within normal ranges reported in previous MRI IRR studies. So in conclusion, the QuickSTIR MRI may be used to rapidly assess the spine for significant injuries warranting early intervention and precaution. Sensitivity for detecting injuries was the highest at O2C2 or with clinically significant injuries. Specificity was high for all injury types and all spinal regions. Future studies are obviously warranted with a larger population in order to determine the true utility of the QuickSTIR MRI. Our limitations include a small number of patients and scans. None of the patients required an operation, so we're unable to correlate any imaging to andropedic findings. And there was a large heterogeneity in scans due to the multiple MRI platforms used. Special thanks to my collaborators. Thank you.

Diagnostic Accuracy

Abbreviated STIR MRI Sequence

Pediatric Spine Trauma

QuickSTIR MRI

Sensitivity