The aim of this study was to compare the effect of different image processi
ng modes (filters) on reproducibility and validity of the assessment of int
erproximal carious lesions on digitized radiographic images. Standardized r
adiographs of 34 extracted teeth exhibiting interproximal caries were obtai
ned. All radiographs were digitized and using the FRIACOM-software the cent
ral depth (CD) of each carious lesion was measured on the unchanged radiogr
aphic image and after use of five different filters with 7fold and 18fold m
agnification. All measurements were repeated after 1 week to estimate repro
ducibility. Histometric CD assessments provided a gold standard for compari
son with the radiographic measurements (validity). Neither reproducibility
nor validity of CD measurements were improved by one of the filters. Measur
ements of dentinal lesions showed a statistically higher variability than t
hose of enamel lesions (p<0.001). 18fold magnification provided more va lid
CD measurements than 7fold magnification (p<0.001). In this study digital
manipulations of radiographic images failed to result in statistically sign
ificantly improved reproducibility or validity of CD measurements.