Report on a novel bone registration method: A rapid, accurate, and radiation-free technique for computer- and robotic-assisted orthopedic surgeries

Computer- and robotic-assisted technologies have recently been introduced into orthopedic surgery to improve accuracy. Each requires intraoperative “bone registration,” but existing methods are time consuming, often inaccurate, and/or require bulky and costly equipment that produces substantial radiation. We developed a novel method of bone registration using a compact 3D structured light surface scanner that can scan thousands of points simultaneously without any ionizing radiation. Visible light is projected in a specific pattern onto a 3 × 3 cm2 area of exposed bone, which deforms the pattern in a way determined by the local bone geometry. A quantitative analysis reconstructs this local geometry and compares it to the preoperative imaging, thereby effecting rapid bone registration. A registration accuracy study using our novel method was conducted on 24 CT-scanned femur Sawbones®. We simulated exposures typically seen during knee/hip arthroplasty and common bone tumor resections. The registration accuracy of our technique was quantified by measuring the discrepancy of known points (i.e., pre-drilled holes) on the bone. Our technique demonstrated a registration accuracy of 0.44 ± 0.22 mm. This compared favorably with literature-reported values of 0.68 ± 0.14 mm (p-value = 0.001) for the paired-point technique13 and 0.86 ± 0.38 mm for the intraoperative CT based techniques 14 (not enough reported data to calculate p-value). We have developed a novel method of bone registration for computer and robotic-assisted surgery using 3D surface scanning technology that is rapid, compact, and radiation-free. We have demonstrated increased accuracy compared to existing methods (using historical controls).