Prediction of ^99mTc-HMPAO-labeled leukocyte scintigraphy results in patients with knee arthroplasty by means of a new semiquantitave parameter in three-phase bone scan

Background: Interpretation of three-phase bone scan in patients with suspected loosening of knee prosthesis relies on the evaluation of asymmetry of counts in blood flow and blood pool phases between the prothesized and controlateral knee. In case of three-phase bone scan pattern of loosening, autologous leucocyte scintigraphy is the gold standard non-invasive imaging method for distinguishing between septic and aseptic loosening. However, autologous leucocyte scintigraphy is a more complex, expensive and longer exam compared to three-phase bone scan. The aim of the present study was to predict the results of autologous leukocyte scintigraphy by identifying in three-phase bone scan cut-off values of count ratios between the prosthetic knee and the "healthy" knee in perfusion and blood pool phases and the use of a new semiquantitative parameter, named perfusion-to-blood pool ratio (P/BP ratio).

Methods: Thirty patients with history of unilateral knee arthroplasty underwent three-phase bone scan (after 1 year from arthroplasty) and subsequently ^99mTc-HMPAO-labeled leucocytes scintigraphy (within an interval of 2 months), used as standard of reference to establish presence of infection. Autologous leukocyte scintigraphies were evaluated visually and semiquantitatively. In three-phase bone scans, regions of interest (ROIs) were delineated in the perfusion and blood-pool phase images, incorporating the entire prosthetic region and applying an isocontour method (including only pixels within a threshold of 40% of the maximum pixel activity within the ROI); corresponding mirror ROIs were placed on the healthy knee, making minimal manual positioning corrections and applying the isountour method. Perfusion (Pr) and Blood Pool (BPr) ratios were calculated by dividing the total ROI counts in the prosthetic knee for the ROI counts in the healthy knee; P/BP ratio was calculated as follows: P/BP ratio = {[(Pr/BPr) x 100] - 100}. Corresponding ROC curves for each semiquantitative parameter were generated to identify the optimal cut-offs for predicting the results of the autologous leukocyte scintigraphy.

Results: Autologous leukocyte scintigraphy identified 13 cases of septic loosening and 17 negative cases. The optimal cut-offs of Pr (1.91), BPr (1.5) and P/BP ratio (25%) discriminated between septic and aseptic loosening (respectively above and below the cut-off values) with high accuracy (Pr: sensitivity = 100%, specificity = 58.8%; BPr: sensitivity = 92.3% , specificity = 41.18%; P/BP ratio: sensitivity = 100%, specificity = 76.47%).

Conclusions: Semiquantitative evaluation in three-phase bone scan may serve as a relevant tool in the assessment of suspected septic loosening of knee prostheses. In our experience, P/BP ratio is the semiquantitative parameter with the highest predictive power of septic loosening. These findings warrant confirmation in larger patient samples; furthermore we will assess whether the use of P/BP ratio may anticipate the timing of differential diagnosis between septic and aseptic loosening and influence clinical decision making.