Abstract
BACKGROUNDS
Total knee arthroplasty (TKA) using an imageless navigation is widely used in these days. Despite the usefulness of navigation-assisted TKA, there are still limitations of accuracy. From previous studies, many factors have been suggested as causes of the discordance between pre-op planning and post-op results. In Addition, Registration of reliable landmark is very important factor in navigation-assisted TKA, fundamentally. Nevertheless, current method of registration process is substantially affected by subjective preference of operators. Until now, However, there is no consensus about the optimal range of reference point. Moreover, the tolerance of imageless navigation system is still questionable. We investigated the effect of variation during the manual registration in this study. We compared the measured alignment and calculated plan of navigation system which were collected from repeated independent registration processes.
METHODS
From 7 March 2016 to 13 May 2016, 44 patients (49 knees) underwent navigation assisted TKA with Orthopilot® Aesculap system. The subject group were severe osteoarthritis patients, they have evaluated radiographically and clinically before the operation. we excluded candidates who have shown very severe mal-alignment (>20 °) and metaphyseal bowing in Pre-op radiographic evaluation. All patients were followed for postoperative long axis film that could measure the correction angle, and followed clinically for functional score. Authors executed multiple registration trials in a single case, each trial was implemented by different surgeons (Senior surgeon JHJ and trainee LJH1, LJH2). At first, Senior surgeon (JHJ) start the operation from initial approach. Standard sub-vastus approach was applied to all-patients. After the procedure of joint exposure, each participating surgeon did the examination of knee anatomy and registered optimal point of his own. It was repeated three times (J,L1,L2) via imageless navigation system. Then, we collected the information of measured limb alignments and calculated plans of tibia cutting from navigation system.
RESULTS
33 knees were evaluated as Gr. 4 in Kellgren-Lawrence classification. The other 16 knees were Gr. 3. In repeated registration processes, patients who were scored Gr. 3 have shown no significant differences in mechanical limb alignments, both coronal and sagittal. There were also no significant differences in Gr. 4 patients, too. Initial tibia planning has shown the largest variance between medial and lateral cutting level (0.4 ± 1.3 mm, in neutral alignment). But, no statistical significance was observed. There is a tendency that the deviation of tibia planning has diminished gradually with the progression of this study. In radiographic evaluation, all cases have satisfactory limb alignments postoperatively.
CONCLUSION
Our experiment suggest that variation of landmark registration alone couldn't have a significant effect on the calculated alignment of navigation system. In this study, we concluded that tolerable range of registration process for alignment calculation is relatively wide. Additionally, we think that the cutting depth is more vulnerable than alignment calculation, and it may need further study with more cases. Measured limb alignment is almost reliable in imageless navigation. Even though operators were not so experienced for the registration process.
