Low-Dose Radiation 3D Intraoperative Imaging

MINI: The objective of this study was to evaluate the accuracy and reliability of pedicle screw placement using O-Arm at dosages below the manufactured recommended dose. O-Arm at reduced dose showed a 90% accuracy when compared with computed tomography; however, about 30% medial breaches were misclassified. STUDY DESIGN: Cadaveric study. OBJECTIVE: The objective was to evaluate O-Arm's ability at low-dose (LD) settings to assess intraoperative screw placement. SUMMARY OF BACKGROUND DATA: Accurate placement of pedicle screws is crucial because of proximity to vital structures. Malposition of screws may result in significant morbidity and potential mortality. O-arm provides real-time, intraoperative imaging of patient's anatomy and provides higher accuracy in scoliosis surgeries, avoiding risk to vital structures. We hypothesize using LD or ultra-low doses (ULDs) to obtain intraoperative images allow for accurate assessment of screw placement, both minimizing radiation exposure and preventing screw misplacement. METHODS: Eight cadavers were instrumented with pedicle screws bilaterally from T1 to S1. Screws were randomly placed using O-arm navigation into three positions: contained within the bone, OUT-anterior/lateral, and OUT-medial. O-arm images were obtained at three dosage settings: LD (kVp120/mAs125-lowest manufacturer recommended), very-low dose (VLD) (kVp120/mAs63), and ULD (kVp120/mAs39). Computed tomography (CT) scan was performed using institution's LD protocol (kVp100/mAs50) and gross dissection to identify screw positions. RESULTS: LD, VLD, ULD, and CT for identifying "IN" screws relative to gross dissection had, a mean (standard deviation) sensitivity of 84.2% (±5.7), specificity of 76.1% (±9.3), and accuracy of 79.9% (±3.1) from all three observers. Across the three observers, the interobserver agreement was 0.67 (0.61-0.72) for LD, 0.74 (0.69-0.79) for VLD, 0.61 (0.56-0.66) for ULD, and 0.79 (0.74-0.84) for CT. Effective doses of radiation (mSV) for LD O-arm scan was 2.16, VLD 1.08, ULD 0.68, and our LD CT protocol was 1.05. CONCLUSION: Accuracy of pedicle screw placement is similar for O-arm at all doses and CT compared to gross dissection. Interobserver reliability was substantial for VLD and CT. Approximately 30% of medial screw breaches are, however, misclassified. ULD and VLDs can be used for intraoperative navigation and evaluation purposes within these limitations. LEVEL OF EVIDENCE: N/A.