Systematic Review on Reproducibility of Nuclear Imaging in the Assessment of Painful Hip and Knee Replacements

摘要

Abstract Nuclear imaging is used in the evaluation of painful arthroplasties for diagnosing loosening or periprosthetic joint infection (PJI). The purpose of this systematic review is to depict the reproducibility of the various nuclear imaging modalities used in the assessment of painful hip and knee arthroplasties. A systematic review of the literature was conducted with a comprehensive search of MED LINE to identify clinical studies investigating the intra-and inter-observer agreement of nuclear imaging modalities in diagnosing PJI and mechanical loosening. A total of 3.000 studies, published between 2004 and 2020, were reviewed and 16 studies met the inclusion criteria. Quality assessment identified concerns with regard to the internal validity in approximately 40% of the included studies. A meta-analysis could not be performed because of insufficient available clinical data. The intra-observer agreement was poorly investigated. The included studies reported an inter-observer agreement of "slight" to "excellent" for bone scintigraphy, "moderate" to "substantial" for leukocyte scintigraphy, "substantial" to "almost perfect" for combined leukocyte and bone marrow scintigraphy, and "fair" to "substantial" for anti-granulocyte scintigraphy. Hybrid imaging with SPECT/CT and FDG-PET/CT demonstrated "substantial" and "almost perfect" inter-observer agreement for symptomatic hip prostheses, respectively. The reproducibility of nuclear imaging is underreported in clinical studies investigating painful knee and hip arthroplasties. Moreover, the included studies demonstrated methodological concerns with a high risk of bias. The available evidence demonstrated a wide range of inter-observer agreement using scintigraphy. Hybrid imaging with SPECT/CT and FDG-PET/CT may improve the accuracy of interpretation and reproducibility. However, literature provides limited data to support this assumption.