Accuracy of computer-assisted radiographic measurement of interproximal bone loss in vertical bone defects

Objective. The objective of the present study was to compare 2 different computer-assisted analysis systems with respect to the measurement of interproximal bone loss on radiographs. Study Design. In 14 patients with untreated advanced periodontal disease, 90 standardized radiographs were taken presurgically and during postoperative follow-up. During periodontal surgery for 30 vertical bone defects and 28 adjacent sites (22 vertical, 3 horizontal, and 3 without bone loss), the distances from the cementoenamel junction (CEJ) to the alveolar crest (AC) and from the CEJ to the bottom of bony defect (BD) were measured. In all radiographs, the linear distances from the CEJ to the AC and from the CEJ to BD were assessed by using 2 computer-assisted analysis devices: linear measurement (LMSRT) and FRIACOM (FRIADENT, Mannheim, Germany). A comparison between the radiographic and intrasurgical assessments was performed by using the paired t test. Results. With respect to the linear distance from the CEJ to the BD, the study failed to detect statistically significant differences between the computer-assisted techniques and intrasurgical measurement as the gold standard (LMSRT,: 0.26 +/- 2.16 mm [P >.05]; FRIACOM, 0.35 +/- 2.5 mm [P >.05]). In terms of the distance from the CEJ to the BD and from the CEJ to the AC, the correlation between both computer-assisted methods was excellent (r(s) > or = 0.75) and there were no statistically significant differences detectable between LMSRT and FRIACOM for both distances. With regard to the distance from the CEJ to the AC, both techniques overestimated interproximal bone loss as compared with intrasurgical measurements (CEJ-AC, LMSRT: 1.63 +/- 2.52 mm [P <.01]; FRIACOM: 2.00 +/- 2.48 mm [P <.01]). The amount of overestimation of interproximal bone loss did not differ significantly in either computer-assisted technique (P >.05). For double measurements, both LMSRT and FRIACOM showed excellent reproducibility with coefficients of variation that ranged between 3.8% and 4.5% (CEJ-BD) and between 5.7% and 6.6% (CEJ-AC). There were no statistically significant differences between the reproducibilities of both computer-assisted techniques (P <.05). Conclusions. Both LMSRT and FRIACOM are reproducible methods to measure the distances from the CEJ to the BD and from the CEJ to the AC on radiographs and can be recommended for clinical use. With respect to the distance from the CEJ to the AC, it must be kept in mind that both computer-assisted methods tend to overestimate the real distance on an average of 1.5 to 2.0 mm.