Reliability of assessing interproximal bone loss by digital radiography: intrabony defects

Aim: The aim of the present study was to assess the reproducibility and validity of linear measurements of interproximal bone loss in intrabony defects on digitized radiographic images after application of different filters and magnifications. Methods: Immediately before surgery 50 radiographs of 50 periodontally diseased teeth exhibiting interproximal intrabony defects were obtained by a standardized technique in 50 patients. Intrasurgically the distances from the cemento-enamel-junction (CEJ) to the alveolar crest (AC) and from the CEJ to the deepest extension of the bony defect (BD) were assessed. All radiographs were digitized by a flatbed scanner (resolution: 600 X 1200 dpi). Using the FRIACOM-soft ware, the linear distances CEJ to AC and CEJ to BD were measured at 50 intrabony defects on digitized but unchanged radiographic images and also after use of 2 different basic image processing modes (filters: enhancement of grey level differences, spreading of grey values) with 7-fold and 14-fold magnification by 2 different examiners. Results: Repeated measures MANOVA revealed reproducibility of the measurement of the distance CEJ to AC to be significantly influenced by examiner (p = 0.027) and filter in combination with the height of 2 wall component of the intrabony defect (p = 0.066). For the distance CEJ to BD filters had significant influence on reproducibility in correlation with vertical angulation difference (p = 0.001). On the average in this study radiographic measurements tended to overestimate the amount of bone loss as assessed by intrasurgical measurements (CEJ-AC: 0.74-1.91 mm; CEJ-DB: -0.04-0.77 mm). Validity of measurement of the distance CEJ-AC was shown to be significantly influenced by the depth of the intrabony defect (p < 0.003). Validity of the distance CEJ-BD was significantly influenced by intrasurgically assessed bone loss (p = 0.029), horizontal angulation (p = 0.066). Filters influenced the validity only in combination with examiner (p < 0.001). Conclusions: In this study, the chosen digital manipulations (filters: spreading, structure) of radiographic images failed to result in statistically significantly more reproducible or valid measurements of interproximal bone loss within intrabony defects when compared to the digitized but unchanged images. All radiographic assessments on the digitized images except for use of enhancement of grey level differences (structure) came close to the intrasurgical gold standard.