Reliability of detecting prosthesis/cement interface radiolucencies in total hip arthroplasty

A laboratory study assessed the reliability of detecting radiolucencies at the prosthesis/cement interface in femoral components for total hip replacement. Radiolucencies (thicknesses = 0.1, 0.3, 0.5 and 0.7 mm) were created by randomly masking non-tip Gruen zones (six per stem) in a group of 72 matte-finish femoral components. Only half of all Gruen zones were masked. The femoral components were reproducibly implanted in composite fiberglass replicate femurs. Anteroposterior and lateral radiographs taken using conventional techniques were then presented to 10 experienced readers, who scored each of 432 non-tip Gruen zones for radiolucency presence or absence. The series-average radiolucency detection rate was 47.1%, with a breakdown of 9.8%, 20.7%, 69.8% and 88.0% for radiolucency thicknesses of 0.1, 0.3, 0.5 and 0.7 mm, respectively. The false positive rate was 7.5%. The findings argue strongly that in many or most instances, initial cement-prosthesis debonding is not reliably detected from conventional plain film radiographs. Radiolucencies of at least 0.7 mm width are necessary to be accurately detectable by most viewers.