General consent of patients was obtained. The institutional review board (Health and Welfare Directorate of the Canton of Bern, Switzerland; Cantonal Ethics committee for research, Project ID 2019-02142) waived the need for an informed consent. All methods were carried out in accordance with relevant guidelines and regulations. All the experimental protocols were approved by the institutional review Board (Health and Welfare Directorate of the Canton of Bern, Switzerland; Cantonal Ethics committee for research, Project ID 2019-02142).
Patients
The study group was a consecutive series of trauma patients with suspected spinal injuries admitted to our Level I trauma center. Between 02/2009 and 12/2012, 344 patients aged 16 years and older with an Injury Severity Score (ISS) equal to or greater than 16 underwent AP-LS and a full-body CT scan upon presentation in the emergency department. Data were retrieved from individual patient records and the picture archiving and communication system (PACS) image software (Sectra Workstation IDS7, Version 19.3, Sectra AB© Sweden). Independent variables included age, sex, mechanism of injury, and ISS.
Sample size calculation
The study will focus on a total of 335 participants. This is a diagnostic test accuracy study to estimate the sensitivity and specificity of LODOX in detecting fracture. Diagnostic measures will be estimated with 95% Wilson confidence intervals. We expect the prevalence of fracture to be between 20-40% [1, 15], sensitivity to be between 40%-70% [1] and specificity between 85%-100% [1]. A sample of 335 participants will result in a two-sided 95% Wilson confidence interval around the sensitivity and specificity as shown in the tables below (Supplemental material Table 4a, Sensitivity table; Table 4b Specificity table).
Radiographic imaging
The AP-LS was performed by a Lodox Statscanner (Statscan Critical Imaging System, Lodox Systems [Pty] Ltd, Johannesburg, South Africa). The LS C-arm rotates around the patient with an angle between 0 and 90 degrees and can provide an AP view within 13 seconds (138 mm/s) [1]. Detailed information about the Lodox Linear Slot Scanning Radiography System can be found at the companies online presence (http://lodox.com).
A full-body CT scan followed the performance of an LS. All CT examinations were performed using a 16-slice multidetector-row computed tomography system (Sensation 16, Siemens, Forchheim, Germany) with collimation of 16 by 1.5 mm and a reconstruction slice thickness of both 2 mm and 5 mm.
The full-body CT scans of all patients were analyzed in terms of spinal injuries by two independent (blinded) investigators. All injuries were classified according to the AO-Classification for spinal injuries [16–18]. If the classification was not concordant, the two investigators reached a consensus. The results of the CT scans were used as the diagnostic reference.
Image analysis
The full-body CT scans of all patients were analyzed in terms of spinal injuries by two independent (blinded) investigators. These were both senior physicians: An experienced radiologist and an experienced spine surgeon. All investigators went over various planes (ap, lateral and sagittal planes) while evaluating the CT scans for spinal injuries.
The AP-LS was assessed for signs of spinal injuries by three independent observers (physicians with the following speciality and experience: radiology attending with >5 years of experience, orthopedic attending with >5 years of experience, and orthopedic resident with <5 years of experience). If any spinal injuries were visible on AP-LS, the number (some patients presented with multiple spinal injuries) and level(s) of the spinal injury were noted. The level of the injury was classified as cervical, thoracic, lumbar, or sacropelvic.
Data sharing statement
All data generated or analyzed during the study are included in the published paper.
Statistical analysis
SPSS (IBM SPSS Statistics for Windows, Version 25.0, IBM Corp, Armonk, NY, 2017) was used to perform the statistical analyses.
Diagnostic accuracy (sensitivity, specificity, negative, and positive predictive values) was calculated for each of the following levels: cervical (occipitocervical and subaxial combined), thoracic, lumbar spine, and sacropelvic. The AUC (Area under the operator receiver characteristics curve) was computed for the three observers, namely the radiology attending (RA), orthopedic resident (OR), and the orthopedic attending (OA). The AUC ranges from 0.5-1. Values of 0.9-1.0 show that the test has an excellent discrimination ability, whereas values of 0.8-0.9 demonstrate a good, 0.7-0.8 a fair, 0.6-0.7 a poor, and 0.5-0.6 fail discrimination ability of the test [19].
The sensitivity and specificity were further calculated for stable versus potentially unstable injuries. Stable injuries were defined as the following:: A0-, A1- and A2- type subaxial and thoracolumbar injuries and: A- and B-type sacropelvic injuries. Potentially unstable injuries were defined as the following: A3 and A4-type as well as B- and C-type subaxial and thoracolumbar injuries, and C-type sacropelvic injuries according to AO-Spine classification [20]. If a patient presented with a stable and potentially unstable fracture, he or she was allocated as potentially unstable.
Interrater reliability between the three observers was calculated using Fleiss' Kappa and rated, according to Landis and Koch [21]. Kappa values range from −1 to +1 and are interpreted as follows: <0.00 poor, 0.00-0.20 slight, 0.21-0.40 fair, 0.41-0.60 moderate, 0.61-0.80 substantial, and 0.81-1.00 almost perfect interrater reliability.