Comparison of weight-bearing full-length radiographs and computed-tomography-scan-based three-dimensional models in the assessment of knee joint coronal alignment

The aim of the study was to determine any discrepancies among preoperative full-leg standing radiographs (LLR) and supine non-weight-bearing computed tomography (CT)-scan-based three-dimensional (3D) models in the assessment of the lower limb alignment prior to total knee arthroplasty (TKA) and answer the question of whether the LLR study can be obviated in preoperative planning when TKA is performed with patient-specific instrumentation (PSI).

LLR and CT-scan-based 3D models of 227 knees (183 patients) were measured. LLR data was then compared to 3D alignment data used to design the PSI for TKA surgery.

Alignment on LLR ranged from 153 to 194° versus 161.5 to 190.5° with CT-scan-based 3D models. The mean (standard deviation, SD) difference among techniques was 1.9° (1.15°) with a statistically significant difference ( P = 2e-16, namely P < .0001). Supine CT-scan-based 3D models underestimated the deformity in 167 cases (73.6%), exactly matched the value of LLR in 24 cases (10.6%) and overestimated the deformity in 36 cases (15.8%).

CT-scan-based models underestimate the degree of deformity at the knee joint. Despite the accurate information provided by the CT-scan and the 3D models (which is the basis for the planning of bone cuts), weight-bearing LLR should not be overlooked in the planning of TKA surgery to assess the extent of the coronal mediolateral instability.