Increased Accuracy in Templating for Total Knee Arthroplasty Using 3D Models Generated from Radiographs

Abstract

AbstractTemplating prior to total hip arthroplasty is a widely adopted practice that aims to improve operative efficiency and reduce clinical outliers. Predicting implant size before total knee arthroplasty (TKA), although less common, could increase operating room efficiency by reducing necessary equipment needed for the procedure. This study compared templating accuracy in TKA using two-dimensional (2D) digital radiographs to a novel imaging technology that generates a three-dimensional (3D) model from these 2D radiographs. Two hundred and two robotic-assisted primary TKA surgical cases using Persona Knee System (Zimmer Biomet, Warsaw, IN) were retrospectively analyzed. For all cases, 3D templating was completed preoperatively using a novel radiographic image acquisition protocol. Using the same radiographs, the knee was templated using a 2D digital templating program. All surgeons were blinded to the final implant sizes, and all templating was done independently. The accuracy of predictions within ± 1 from the final implant size was determined for the femoral and tibial components. The accuracy (within 1 size) of tibial size predictions was comparable between attending surgeons and residents (87 vs. 82%, p = 0.08), but attending surgeons more accurately predicted the femoral size (77 vs. 60%, p < 0.05). The 2D to 3D imaging technology more accurately predicted both tibial and femoral sizes compared with the attending surgeons (99.5 vs. 87%, p < 0.05; 84% vs. 77%, p < 0.05). However, the imaging technology, attending surgeons, and residents were all more likely to overestimate femur size (p < 0.05). Moreover, the 3D imaging technology predicted the exact tibial component size in 93.1% of cases, which was significantly greater compared with residents (40%, p < 0.01) and attending surgeons (53%, p < 0.01). The 2D to 3D imaging technology more accurately predicted tibial and femoral component sizes compared with 2D digital templating done by surgeons. All templating predictions were more accurate for the tibial implant size than for the femoral size. The increased accuracy of implant size predictions from this 3D templating technology has the potential to improve intraoperative efficiency and minimize costs and surgical time.