Accuracy of computer navigation for acetabular component placement in THA

Abstract

The accuracy and precision of any computer-aided surgical device is critical to its utility. We asked the following question: how accurate and precise are the values measured by an imageless computer navigation system as compared with those measured using postoperative CT scans? Twenty-five patients (26 hips) underwent primary THA using an imageless computer navigation system for placement of the acetabular component. Inclination and anteversion were measured in the operative coordinate system as defined by Murray. Accuracy, precision, and bias were computed, and Bland-Altman analysis was used to assess levels of agreement. The accuracy (mean +/- standard deviation of the absolute difference between computer-assisted navigation and CT) was 1.8 degrees +/- 1.2 degrees for inclination and 2.0 degrees +/- 2.0 degrees for anteversion. Precision was 3.4 degrees for inclination and 5.5 degrees for anteversion. Bias was 0.52 degrees for inclination and 0.35 degrees for anteversion. Limits of agreement were 4.26 degrees for inclination and 5.58 degrees for anteversion. An imageless computer navigation system can precisely determine acetabular cup position.

Fig. 1

With the patient in the supine position, two Steinmann pins are placed using a sterile technique into the iliac crest approximately two fingerbreadths proximal to the ASIS on the operative side. This photograph shows placement of the first pin. Points corresponding to bilateral ASISs and bilateral pubic tubercles are referenced superficially to the pelvic pins using a blunt probe equipped with optical sensing spheres. Great care is taken to palpate the midpoints of the bony landmarks.

Fig. 2

This photograph shows placement of the acetabular cup with a curved cup inserter equipped with an optical array. The optical array is evident at the patient’s ASIS.

Fig. 3A–B

Scatterplots of Pearson’s r correlation coefficient for CAS versus CT values for (A) inclination and (B) anteversion are shown. There was a correlation for inclination (r = 0.850) and anteversion (r = 0.919).

Fig. 4A–B

Bland-Altman plots of agreement between CAS and CT scan methods for (A) inclination and (B) anteversion are shown. These plots show 96% of inclination values and 92% of anteversion values were within 2 SDs of the mean.

Fig. 5A–B

Scatterplots show accuracy versus BMI for (A) inclination and (B) anteversion. We observed no correlation between the absolute mean difference of the CAS and CT values and BMI for inclination (p = 0.30) or anteversion (p = 0.98).