Abstract
INTRODUCTION: Reported prevalence rates of the fabella sesamoid bone vary widely, particularly between studies based on either X-ray imaging or anatomical dissection approaches. The purpose of this study was to directly compare these two methodologies in their detection of fabellae and investigate whether variability in the density of fabellae could explain any discrepancies. METHODS: Fifty cadaveric knee segments were examined for the presence of a fabella by both X-ray imaging and anatomical dissection. The relative density of each excised fabella specimen was then quantified using a separate set of radiographs. Results: Fabellae were detected in 40% of the sample knees via a manual dissection approach but in just 12% of those same specimens using X-ray imaging. Relative density measurements confirmed that fabellae identifiable only via dissection were significantly less dense than fabellae visible in whole knee radiographs but denser than the surrounding tissue. CONCLUSION: Radiology cannot reliably detect cartilaginous or incompletely ossified fabellae, which were found in 28% of the study population. Clinicians should consider the potential occurrence of a fabella when diagnosing posterolateral knee pain, even if it may not be visible via X-ray.