Our data confirms that SPECT/CT arthrography has a high diagnostic accuracy for detecting aseptic loosening in patients with persistent pain following primary hip and knee arthroplasties (sensitivity of 86%, specificity of 98%, PPV of 86%, NPV of 98%, and overall accuracy of 97%). Comparable results have been reported in previously published studies. In a previous retrospective study of a smaller series of 38 SPECT/CT hip and knee arthrograms from our institution (different patient cohort than included in this study), this technique was found to have a sensitivity of 100.0%, specificity of 96.0%, PPV of 92.9%, NPV of 100.0%, and overall accuracy of 97.4% for loosening12. In another study by Murer et al., examining SPECT/CT arthrography for the evaluation of loosening of total knee arthroplasties, a sensitivity of 96%, specificity of 100%, PPV of 72.7% and NPV of 100% was reported for detecting tibial component loosening. As well, a sensitivity of 95%, specificity of 100%, PPV of 42.9%, and NPV of 100% was reported for detecting femoral component loosening 13. The study concluded that SPECT/CT helped in diagnosing aseptic loosening in patients with persistent pain after primary total knee arthroplasty.
While reported studies have consistently shown high sensitivity and specificity for this technique, our evaluation of 63 patients did demonstrate 1 FP and 1 FN case. In the FN case, while the SPECT/CT was reported as negative for loosening, the revision surgery performed four months after the SPECT/CT scan demonstrated a subsiding tibial tray. The report in this case did describe substantial streak artifact on the associated CT study precluding assessment of prosthetic/osseous structures (metal artifact suppression not applied). This artifact may have contributed to the false negative result and decreased sensitivity. With metal artifact suppression techniques now available for SPECT/CT scanners image quality may be improved, potentially reducing the impact of this artifact. In the FP case, a review of the images demonstrated mis-registration between the SPECT and CT acquisition, likely related to patient motion. With the images re-aligned, activity is seen extending into the bone-prosthetic interface of the medial condyle region, however this does not follow the expected contour of this interface suggesting it is likely artifact.
To diagnose loosening, several imaging modalities have been described including plain radiography, bone scintigraphy, planar nuclear arthrography, SPECT/CT, 18F-FDG-PET and/or MRI 14. A meta-analysis by Temmerman et al. reported a pooled sensitivity and specificity for plain radiography of 82% and 81%, respectively15. The most recent study on bone scintigraphy by Claassen et al. reported a sensitivity of 76% and a specificity of 83% 16. Interestingly, planar nuclear arthrography on its own has a low reported sensitivity and specificity of 87% and 64% [15]. Comparatively, we found a much higher sensitivity and specificity with SPECT/CT arthrography demonstrating the importance of the hybrid nuclear medicine imaging technique in the evaluation of aseptic loosening.
Our study has multiple limitations. First of all, as a retrospective study design, it is inherently prone to selection bias. Moreover, as the surgeons were not blinded to the SPECT/CT reports, the decision to perform the surgical revision reference standard was influenced by the outcome of the SPECT/CT diagnostic test therefore confounded by verification bias. As well, since the study involves clinical reports from multiple different readers, there is also the potential for variability in interpretation of the scans. Finally, while our results demonstrate a high diagnostic accuracy for our SPECT/CT arthrogram technique, it is uncertain if this technique is optimal. Specifically, a wait time of 30 minutes with ambulation or 60 minutes without ambulation may be too short. It is uncertain if a longer wait time may have impacted our FN result. There are no reported studies examining the optimal time from injection to imaging. Further study is warranted in this regard.
In summary, SPECT/CT arthrography appears to be highly accurate for the diagnosis of aseptic loosening in patients with hip or knee arthroplasties. Given this, SPECT/CT arthrography should be considered as an important diagnostic adjunct, particularly for patients where there is clinical uncertainty.