Wednesday, September 26, 2018 7:35 AM–9:00 AM ePosters: P29. Does the true AP radiograph for scoliosis differ from ancient convention? Lessons from 3D imaging

Abstract

BACKGROUND CONTEXT Scoliosis is a three-dimensional deformity. As such, the rotational component must be considered during its evaluation and management. However, the severity of scoliosis has long been measured in the 2D coronal plane using the Cobb angle – a modality unable to account for axial vertebral rotation (VR). In the past, the Stagnara lateral radiograph has attempted to compensate for VR. To date, no similar radiographic view has been described for the AP plane. Recent advances in 3D CT imaging may allow clinicians to more accurately measure Cobb angles in scoliosis by compensating for axial VR. Here, we present a retrospective study comparing Cobb angle measurements taken from both traditional 2D radiographs, and rotationally compensated CT images. PURPOSE The purpose of this study is to compare Cobb angle measurements routinely made for AIS in both rotationally compensated and standard imaging. STUDY DESIGN/SETTING Retrospective matched cohort study. PATIENT SAMPLE A total of 66 patients diagnosed with AIS. OUTCOME MEASURES PT, MT, and TL/L curve Cobb angle values, coronal balance, thoracic and lumbar apical vertebral translation (T/L-AVT), thoracic and lumbar apical vertebral rotation (T/L-AVR), thoracic kyphosis, and sagittal balance. METHODS In this retrospective study, standard AP, lateral, and bending radiographs of 66 patients age 10-19 with AIS were obtained. Using multiplanar 3D CT, the coronal plane was reconstructed within the plane of the vertebral axis to form a rotationally compensated 3D CT image (RC-3DCT). The following measurements were made on both imaging modalities: PT, MT, TL/L curve, coronal balance, T/L-AVT, T/L-AVR, thoracic kyphosis, and sagittal balance. RESULTS Mean MT curve was found to be 40.56° on AP radiograph versus 35.39° on RC-3DCT (n=63, P=.00039). In 28 patients (42.4%), rotational compensation altered the end vertebra of the major curve by at least 1 level. The mean TL curve obtained on radiograph, was significantly larger than that of RC-3DCT (n=62; 35.34° vs. 30.98° ; P=.0046). AP films overstated coronal balance by over 50% when compared to RC-3DCT Mean T-AVT was measured to be 25.28 mm on conventional films and 19.46 mm on RC-3DCT (n=62; P=.00033). RC-3DCT L-AVT was significantly smaller than that of standard films (P=.0094). Mean T-AVT on standing film was 26.2±31.0°. The mean proximal thoracic kyphosis did not differ between the two modalities: 20.66° on radiograph versus 21.062° on RC-3DCT (P=.41). CONCLUSIONS Measurements traditionally made on AP radiograph to evaluate scoliosis may be inaccurate due to VR. Our data suggest that axial rotation of a patient 19° per Nash–Moe grade with respect to the coronal plane may adequately compensate for the rotational component of deformity without the need for CT. Additionally, rotational compensation may yield end vertebrae levels different from those obtained traditionally. In 42% of patients, VR compensation altered end vertebral designation by at least 1 level.