Purpose: The influence of angular deformity of the scaphoid on wrist function and arthrosis is debated and the reliability of the described quantitative measurements of deformity has been questioned. We hypothesized that the inherent imprecision with which computed tomography scanning planes are selected introduces another source of variability in measurements of scaphoid deformity, further diminishing their reliability.
Methods: Sagittal plane images of 15 computed tomograms of normal scaphoids were evaluated in 3 different reconstruction planes. Four observers measured the lateral intrascaphoid angle, the dorsal cortical angle, and the height-to-length ratio of the 45 images in random order and then measured them again in a distinct random order 2 weeks later. The variability of each observer's measurements (intraobserver reliability) was evaluated with Pearson correlation coefficients. The agreement of the measurements made by the 4 observers (interobserver reliability) and the agreement of the measurements of the same bone in different reconstruction planes (interplane reliability) were evaluated using interclass correlation coefficients.
Results: The intraobserver reliability was poor for 27 of 36 comparisons. The interobserver reliability of the dorsal cortical angle and the intrascaphoid angle was poor for all reconstruction planes. The interobserver reliability of the height-to-length ratio was good for 2 planes and poor for the third plane. The interplane reliability was poor for 7 of 12 comparisons, with no single measurement technique remaining consistent for all observers across reconstruction planes.
Conclusions: Quantitative measurements of scaphoid deformity have very limited reliability for individual observers, between different observers, and depending on the plane in which the image of the scaphoid is produced. Even the most reliable measure of deformity (height-to-length ratio) was not consistent between reconstruction planes. Unless more reliable scanning and measurement techniques are developed ideas about the effect of scaphoid deformity on wrist function will remain to a large degree speculative.