Radiographic lunate type prediction: A comparison of fluoroscopy, cross-sectional radiography, and...

The lunate bone, as the central bone of the proximal carpal row, shows anatomical variability regarding its possible additional articulation with the hamate bone. This allows the morphological classification of the lunate bone into two subtypes according to the absence or presence of an accessory articular facet. Such additional articulation markedly influences the biomechanical behavior of the carpal rows, which makes radiological identification of such variants highly relevant. However, radiographic analysis has shown limited accuracy. This study evaluated the reproducibility and reliability of classic 2D- and 3D-radiography for accurately predicting lunate morphology.

Fifty fresh-frozen hand-specimens were obtained; of these, two had to be excluded, leaving 48 specimens for 2D- and 3D-radiography. Two independent observers predicted lunate type and intra- and inter-observer-reliability were assessed. Macroscopic lunate morphology was verified in a specimen-subset.

Mean intra-observer Cohen's kappa was 0.524-0.892, indicating moderate to almost perfect agreement. Mean inter-observer Cohen's kappa was 0.292-0.636, indicating fair to substantial agreement. All 12 dissected specimens showed an additional hamate articular facet. Direct imaging measurements led to a correct identification in 25.0-50.0% of cases, indicating that accessory medial facets were frequently missed radiographically.

Radiographic assessment did not reliably identify all lunate bones with two facets in this study. Capitate-triquetrum distance (CTD) analysis showed the highest observer agreement, supporting its value as a reproducible measurement-based method, but it did not eliminate misclassification. Radiographic lunate type prediction should therefore be interpreted with caution, particularly when no additional articular facet is reported.