Comparison of digital and plain radiography preoperative templating in total knee arthroplasty

Abstract

We present the results of prospective evaluation of digital compared to plain radiographic preoperative templating for primary total knee replacement. All consecutive patients undergoing primary knee replacement were eligible. Patients with previous knee replacement or without calibrated digital or plain radiographs were excluded. Plain radiographs were templated against acetate templates. Digital images were templated with the help of commercial software TraumaCad. A 25-mm spherical metal ball placed nearest to the affected knee joint acted as calibration object. The ICC value for intra-rater reliability was 0.846 for tibial templating and 0.840 for femoral templating. PFC sigma cruciate substituting components were implanted in all patients. Twenty-eight consecutive patients between April 2006 and June 2007 were included. Accurate digital templating score was 80% for tibial implant and 40% for femur. Accuracy of analog templating was 55% for tibial implant and 50% for femur. There was no mismatch of over one size. The differences between templated and implanted sizes were plotted against their mean in Bland–Altman plot. The 95% confidence interval of the differences between digital and actual sizes were: 0.78 to −0.75 sizes for tibial implant and 1.15 to −0.93 sizes for femoral implant. The 95% confidence interval of the differences between plain and actual sizes were: 0.38 to −0.99 size for tibial implant and 0.93 to −1.32 size for femoral implant. The two tailed P value for difference between digital and analog templating from a Wilcoxon matched pair signed rank test was 0.021 for tibia and 0.006 for femur. We found preoperative templating by the operating surgeon reliable and accurate but digital templating did not offer any additional advantage.