Accuracy and reproducibility of preoperative three-dimensional planning for total hip arthroplasty using biplanar low-dose radiographs : A pilot study

Abstract Background In total hip arthroplasty (THA), the acetabular cup and femoral stem must be correctly sized and positioned to avoid intraoperative and postoperative complications, achieve good functional outcomes and ensure long-term survival. Current two-dimensional (2D) techniques do not provide sufficient accuracy, while low-dose biplanar X-rays (EOS) had not been assessed in this indication. Therefore, we performed a case-control study to : (1) evaluate the prediction of stem and cup size for a new 3D planning technique (stereoradiographic imaging plus 3D modeling) in comparison to 2D templating on film radiographs and (2) evaluate the accuracy and reproducibility of this 3D technique for preoperative THA planning. Hypothesis Accuracy and reproducibility are better with the 3D vs. 2D method. Patients and methods Stem and cup sizes were retrospectively determined by two senior surgeons, twice, for a total of 31 unilateral primary THA patients in this pilot study, using 3D preplanning software on low-dose biplanar X-rays and with 2D templating on conventional anteroposterior (AP) film radiographs. Patients with a modular neck or dual-mobility prosthesis were excluded. All patients but one had primary osteoarthritis; one following trauma did not have a cup implanted. The retrospectively planned sizes were compared to the sizes selected during surgery, and intraclass coefficients (ICC) calculated. Results 3D planning predicted stem size more accurately than 2D templating: stem sizes were planned within one size in 26/31 (84%) of cases in 3D versus 21/31 (68%) in 2D ( P = 0.04). 3D and 2D planning accuracies were not significantly different for cup size: cup sizes were planned within one size in 28/30 (92%) of cases in 3D versus 26/30 (87%) in 2D ( P = 0.30). ICC for stem size were 0.88 vs. 0.91 for 3D and 2D, respectively. Inter-operator ICCs for cup size were 0.84 vs. 0.71, respectively. Repetitions of the 3D planning were within one size (except one stem), with the majority predicting the same size. Discussion Increased accuracy in 3D may be due to the use of actual size (non-magnified) images, and judging fit on AP and lateral images simultaneously. Results for other implant components may differ from those presented. Size selection may improve further with planning experience, based on a feedback loop between planning and surgical execution. Level of evidence Level III. Retrospective case-control study.