Soft tissue gaps created in total knee replacement rely on the creation of symmetrical spaces that accommodate prosthetic implants. We studied a new custom surface registration protocol in a computer navigation system to accurately and precisely measure these gaps. In eight cadaver lower extremities, gaps were measured from the proximal tibial cut surface to the registered most distal surfaces of the medial and lateral femoral condyles, measured from 0 to 120 degrees. The computer measurement was compared against metrology spacers precise to 200 microns. Tensor reproducibility was assessed using a typical teeter-toter tensor in four specimens with cruciate retained and four with sacrificing technique.

Generalised MANOVA tests were used for assessment of means of repeated measures involving the three separate experiments. There was no difference between the measurements obtained using computer navigation compared to the metrology spacers in one specimen including the re-registration group (P = NS, Beta = 0.9). The sagittal position of the knee (Flexion/Extension) did affect the magnitude of the measurements obtained. (P=.001) For comparison, descriptive statistics of spacer block versus navigation measure revealed for the medial compartment measurement, a mean (n=200) of 0.006 mm (SD: 0.32 mm) and lateral compartment measure (n=200) of 0.12 mm (SD: 0.41 mm). The projected maximum error was 1.0 mm capturing 100% of values to 90 degrees. The re-registration repeated measures experiment varied as a function of knee flexion and the repetition number. Descriptive statistics for comparison revealed a mean medial compartment measure (n=200) of 0.24 mm (SD: 0.54 mm) and lateral compartment measure(n=200) of 0.01 mm(SD: 0.42 mm).

The tensor study compared the ability of the surgeon to produce a consistent gap measure over eight separate trials. Hypothesis testing revealed significant differences as a function of degree of flexion, order of testing (with later tests having greater gaps), and the specimen being measured (P<.001, P<.001, and P<.001).

The overall conclusion of the block studies was that the computer system was accurate to at least one millimeter for measuring the gaps of the knee. The tensor study demonstrated stretching or permanent strain of the ligaments, significant differences between the angles of flexion and between the individual specimens. This is to say that each specimen was unique with variability of measurements through the range of motion.