Radionuclide arthrogram with SPECT/CT for the evaluation of mechanical loosening of hip and knee prostheses

To evaluate the value of SPECT/CT in radionuclide arthrogram (RNA) for the assessment of mechanical loosening of hip and knee prostheses.

A retrospective audit of 117 RNA SPECT/CTs evaluated by a single reader--40 hips, (1 hemiarthroplasty) and 77 knees (12 unicompartmental). The detection of any radiotracer within the bone/prosthetic interface was deemed positive for loosening. The operative assessment of 29 hip and 44 knee prosthetic joints was known and used as the gold standard. A subsequent blinded reassessment of the planar images was performed and compared with the SPECT/CT results for 26 of the 29 hip and 42 of the 44 knee prostheses.

The respective SPECT/CT versus planar results were as follows: hips--acetabular cup: sensitivity: 73 versus 0%; specificity: 71 versus 100%; positive predictive value: 62% versus indeterminate; negative predictive value: 80% versus 72% (p = 0.0044). Hips--femoral component: sensitivity: 78 versus 63%; specificity: 90 versus 94%; positive predictive value: 78 versus 83%; negative predictive value: 90 versus 85% (p = 0.2482). Knees--femoral component: sensitivity: 75 versus 17%; specificity 63 versus 97%; positive predictive value: 43 versus 67%; negative predictive value: 87 versus 74% (p = 0.0001). Knees--tibial component: sensitivity: 86 versus 63%; specificity: 86 versus 76%; positive predictive value: 55 versus 38%; negative predictive value: 97 versus 90% (p = 0.6831).

For evaluation of mechanical loosening of the hip prosthesis SPECT/CT was significantly better than planar scanning for the acetabular cup, but not for the femoral stem. For evaluation of the knee prosthesis, a significant improvement was noted using SPECT/CT for the femoral component, and although superior results were also noted for the tibial component, statistical significance was not reached. Taking into account the limitations of this retrospective audit, the value of using RNA SPECT/CT appeared to lie in the exclusion of mechanical loosening.