In a prospective study of 35 cemented press fit condylar knees, we studied the prevalence of radiolucent lines 1 mm wide or wider using two methods of detection: conventional (plain) and fluoroscopically-guided radiographs. All films were evaluated in accordance with the Knee Society Total Knee Arthroplasty Roentgenographic Evaluation and Scoring System. A total of 12 radiolucencies were detected in nine knees (26% of all knees) using conventional radiographs versus 25 radiolucencies in 13 knees (37% of all knees) using fluoroscopically-guided radiographs. This was a statistically significant difference and suggests that the true prevalence of periprosthetic radiolucencies will be underestimated if conventional radiographs are used to evaluate the bone-prosthesis interfaces and questions the value of routine postoperative plain film radiographs to evaluate the results of knee arthroplasty.