2738. Ultrasound Sensitivity and Specificity in Brachial Plexus Trauma Compared to MRI
Authors * Denotes Presenting Author
  1. Jenifer Pitman *; Hospital for Special Surgery
  2. Lydia Ko; Hospital for Special Surgery
  3. Steve Lee; Hospital for Special Surgery
  4. Scott Wolfe; Hospital for Special Surgery
  5. Oganna Nwawka; Hospital for Special Surgery
We aim to evaluate the clinical utility of ultrasound (US) in the assessment of traumatic brachial plexus injuries; comparing sonographic findings to MRI, electrodiagnostic studies, surgical, and clinical outcomes.

Materials and Methods:
At a single institution, patients with history and symptoms suggesting traumatic brachial plexopathy underwent high-resolution US and magnetic resonance imaging (MRI). A retrospective chart review was performed with the addition of electrodiagnostic testing (EDX) and surgical outcomes, when available, and a final clinical diagnosis was determined. US was performed using a high-frequency (15-18 MHz) transducer. MRI consisted of high resolution fluid sensitive and anatomic FSE sequences. The brachial plexus was imaged from the paravertebral through the infraclavicular regions. The nerves were first localized at the paravertebral and then followed down to the supraclavicular and infraclavicular regions with transverse and longitudinal scans. On US, contralateral side imaging was obtained. Calculations of sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were performed at the root and branch levels. Surgical findings were used as the gold standard when present, and EDX findings were used as an adjunct in patients who underwent surgery, and as the gold standard in the subset of patients who did not undergo surgery.

Of 30 patients in the initial MRI and US imaging analysis, 13 underwent EDX and subsequent surgery, and 7 underwent EDX. Of these 20, 7 patients had high grade lesions (neuroma, root avulsion or nerve transection), and 12 patients had low grade lesions (scarring, segmental thickening). Sensitivity of US at the level of the roots ranged from 75-100% with the exception of the T1 nerve root, which was only visualized in 20% of patients. US specificity ranged from 93-100%. PPV ranges at US were from 83.33-100%, and NPV ranges were from 60.0-94.12%. Sensitivity on MRI at the root level ranged from 40-66.67% and specificity ranged from 83.33%-100%. PPV on MRI ranged from 40-100% and NPV ranged from 73.3-100%. At the level of the branches, US sensitivity ranged from 25-88.89%, and specificity from 81.82-100%. MRI sensitivity ranged from 58.33-100% and specificity from 75-92.31%. False negatives lowering US sensitivity were related to non-visualization of nerves secondary to obscuring anatomy (e.g., axillary nerve, suprascapular nerve at the clavicle), limited patient range of motion (e.g. axillary nerve), and limited range of the field (e.g. musculoskeletal nerve).

Overall, there was excellent specificity of both MRI and US at both the level of the roots and the branches of the brachial plexus. Sensitivity of US was superior to MRI for the root levels, while MRI sensitivity was superior at the branch levels. These findings suggest a complementary role for US and MRI in the workup of brachial plexus trauma, in conjunction with clinical diagnostic testing. Diagnostic imaging with US and MRI may be of use for guidance in surgical exploration and management.