The reproducibility and accuracy of cephalometric analysis using different digital imaging modalities and image compression

grantor: University of TorontoThis study evaluated and compared the reproducibility and accuracy of cephalometric landmarks and measurements between conventional radiographic film and the corresponding indirect, JPEG compressed or direct digital images (DI). Twenty-one conventional films were indirectly digitized using a scanner, and TIFF images were converted to JPEG files using 2 levels of compression (12:1 and 25:1). Simultaneous acquisition of conventional and direct DI (stimulable-storage phosphor (SPP)) was obtained for thirty patients. Three observers for the indirect and one observer for the direct DI identified 17 and 22 landmarks respectively on the computer monitor. X and y-coordinates were recorded and cephalometric measurements calculated. Statistical analysis involved use of the paired samples t-tests and ANOVA models. The results showed that indirect DI have equivalent landmark reproducibility with conventional film. JPEG compression of indirect DI at 25:1 level results in loss in accuracy of cephalometric measurement when compared to non-compressed DI, whereas, JPEG at 12:1 did not. Direct DI were less accurate for 8 out of 21 landmarks but clinical accuracy of cephalometric measurement was equivalent to conventional film.M.Sc