Antigranulocyte scintigraphy in infected hip prosthesis: the diagnostic importance of delayed 20–24-h imaging and semiquantitative analysis

AIM To evaluate clinical efficacy of a dual-time acquisition protocol, which included 4 and 20/24-h imaging with antigranulocyte antibody scintigraphy (LeukoScan) combined with semiquantitative analysis in the diagnosis of infection in painful hip prosthesis. METHODS Sixty-seven consecutive patients with hip prosthesis were enrolled in this research project: 35 females, 32 males, mean age of 56.3 years. All patients had clinical and biochemical suspicious of infection. Each prosthesis had been implanted 3 months to 12 years before enrollment in this study. Twenty-four patients were on antibiotic therapy at the time of scintigraphy. Seven patients had bilateral hip prosthesis, one painful and the other painless: the seven painless prostheses were considered controls. LeukoScan examination was performed both at early (4 h) and delayed (20/24 h) times. The scintigraphic data were assessed both by visual and semiquantitative methods by three experienced nuclear medicine physicians blinded to clinical, laboratory and radiographic results. The uptake was graded visually by a 4-point scale: intense=3, moderate=2, mild=1 and absent=0. The semiquantitative analysis was obtained by a region of interest (ROI) analysis used in the anterior views to measure the ratio between the mean radioactivity in the prosthesis and the background radioactivity in the early and delayed images. An increase in the intensity of uptake of at least one scale-step at visual analysis and 20% at semiquantitative ROI analysis at the dual-time (early vs. delayed) LeukoScan was considered consistent with infection, whereas a stable or decreasing pattern was judged a negative result. Three-phase 99mTc-hydroxymethane diphosphonate bone scan was also performed routinely. Final diagnosis was determined at surgery and/or long-term clinical and imaging follow-up. RESULTS At visual analysis, sensitivity for both early and delayed imaging was 94%, whereas specificity was 71% for early imaging and 83% for early and delayed imaging approach. At semiquantitative ROI analysis, sensitivity remained 94%, whereas specificity rose slightly to 73% for early imaging and to 90% for early and delayed imaging combined. Of note, four false-positive early scans were diagnosed correctly as negative on delayed imaging showing a decreasing pattern in uptake intensity. Sensitivity and specificity were similar whether patients were on or off antibiotic therapy. CONCLUSION Our data show that early imaging LeukoScan is highly sensitive in evaluating septic prosthesis, but it is not optimally specific. Although the dual-time LeukoScan is capable of significantly increasing specificity for detecting infection. The semiquantitative ROI analysis further increased the specificity. Concomitant antibiotic treatment did not seem to influence the diagnostic efficacy of LeukoScan scintigraphy in detecting infected hip prosthesis.