From image to index: diagnostic accuracy of a novel semi-quantitative approach for assessing suspected periprosthetic joint infection with triple-phase bone scintigraphy

Abstract

Background

Total knee and hip arthroplasty are common procedures for patients with osteoarthritis when conservative therapy fails. Aseptic loosening and periprosthetic joint infection are major complications, with periprosthetic joint infection being a leading cause of pain post-surgery. This study aims to assess the accuracy of a novel semi-quantitative parameter, the Blood Pool-To-Delayed Ratio (BPrDr) variation index, for assessing suspected periprosthetic joint infection in the hip and knee.

Methods

A retrospective analysis was conducted using data from the Nuclear Medicine department at Hospital Clínico Universitario de Valladolid, between October 2018 and December 2020, including patients who underwent surgery for suspected prosthetic joint infection. The Blood Pool ratio (BPr) was calculated by comparing the affected and contralateral joint areas during the blood-pool phase. The Delayed Ratio (Dr) was determined similarly in the delayed phase, and the BPrDr variation index was calculated as [(Dr–BPr)/BPr] × 100. Findings were compared with microbiological culture and intraoperative joint fluid analysis.

Results

Sixty-four patients were included; microbiological examination was positive for infection in 17. The BPrDr variation index showed an area under the ROC curve (AUC) of 0.71 (95% CI 0.57–0.85; P = 0.009). The optimal cut-off for differentiating infection from aseptic loosening was 14.73%, with sensitivity of 88.2%, specificity of 55.3%, positive predictive value of 41.6%, and negative predictive value of 92.8%.

Conclusion

The BPrDr variation index may be a useful screening tool for ruling out periprosthetic joint infection in the hip and knee after joint replacement.