MECHANICAL AXIS ALIGNMENT IN COMPUTER GUIDED TOTAL KNEE REPLACEMENT – COMPARISON OF CHANGES DURING SURGERY AND ASSESSMENT & VALIDATION OF SCANOGRAM

Abstract

Significant concerns remain in computer navigated surgery regarding potential errors due to inadequate tracker or array fixation, cutting guide block movements, saw blade deviation, variable component seating and standardisation and validity of radiographic measurements of alignment for outcome assessment. There are no studies in the literature comparing computer generated axes at different steps of operation as well as radiographic axes using scanograms to our knowledge. Long leg films involve significant radiation, which can be minimised by the use of scanograms.

A prospective study was performed to compare the per-operative and post-operative alignment of the lower limbs after navigated total knee replacements. All consecutive patients who underwent navigated total knee replacement between May 2006 and November 2006 were included in the study. Patients with inadequate data, patients who refused to participate in the study or lost contact, obvious measurement errors and patients having had recent operations were excluded. The intra-operative initial, trial and the final axes were recorded from the navigation system. Post-operatively a CT (Computer Tomogram) scanogram of the lower limbs was performed as per the scanogram protocol. Measurement of the mechanical hip-knee-ankle axis of the lower limb was performed on the computer. Results were analysed using appropriate statistical methods and comparison made between initial, trial, final and scanogram axes with assessment of their correlation coefficients.

Twenty-five patients were initially recruited in the study, of which, 15 were available with completed data. There were four males and 11 females with the age ranging from 57–80 (average 70) years. The right knee was replaced in 12 and the left knee in three patients. The average initial alignment was 0.09° valgus (0.5° varus to 1° valgus), trial alignment 0.59° varus (2° varus to 1° valgus), final alignment 0.56° varus (4° varus to 1.5° valgus) and average radiographic alignment was 0.52° varus (3.1° varus to 1.8° valgus) in maximum possible extension. Average deviation from initial to trial alignment was 0.69° varus, trial to final was 0.03° varus and final to radiographic alignment was 0.12° valgus.

Correlation co-efficient of 0.62 between the initial and final axes with average difference of 0.72° varus (p= 0.11, unequal variance 2 tailed) demonstrates reasonable reproducibility of the alignment with computer-guided surgery, also confirming the fact that there is some variation between the initial cut angles and final mechanical axes. Correlation co-efficient of 0.92 between final axes and radiographic axes suggests that scanogram is an imaging modality with reasonable accuracy for measuring mechanical limb alignment despite significantly low radiation and relatively low resolution. Potential errors in radiographic measurements due to rotational malposition combined with flexion deformity is highlighted.