21 Tracing analysis Four profile tracings were available for each patient: pre-operative, computerized prediction, manual prediction and actual post-operative. All tracings were digitized and entered into the computerized cephalometric software system PORDIOS (Purpose On Request Digitizer Input-Output System, Institute of Orthodontic Computer Sciences, Aarhus, Denmark), sellckchem which calculated all the cephalometric variables used in this study. In order to compare the computerized and manual prediction profiles and to test the prediction validity of the manual method (comparison between manually predicted and actual post-operative profiles) the author used the Profile Analysis cephalometric appraisal (included in the PORDIOS software), which incorporates variables from different well-known cephalometric analyses.
26 Profile Analysis includes 30 landmarks and 59 linear and angular variables.27 For each patient, 30 cephalometric landmarks where identified on the computerized prediction, manual prediction and actual post-treatment profile tracings (Figure 2). Identification of landmarks, tracings, superimpositions, digitizing of cephalograms and computer printouts were performed by the author. Figure 2 Dentoskeletal and soft tissue cephalometric landmarks used in the comparison of the prediction and post-treatment computer profile printouts. G=glabella; S=sella; N=nasion; N��=soft tissue nasion; P=porion; O=orbital; Ba=basion; Pn=pronasale; Pns=posterior … Statistical analysis Paired t-tests were used to determine any statistically significant differences (P < .
05) of cephalometric variables for both the computerized and manual soft tissue predictions; statistically significant differences between manually predicted and actual post-operative patient profile were also determined. Correction of type 1 error level was done by the Bonferroni method. Method error Eleven randomly selected manual prediction tracings were digitized twice. All 59 cephalometric variables of the Profile Analysis were compared by means of paired t-test. No statistically significant differences (P > .05) were found for any of the variables. The error of superimposition was estimated by performing double superimposition and double measurements for all patients. All measurements were analyzed by means of the method error test. No statistically significant differences were found.
The error of landmark displacement during computer simulation of jaw repositioning was estimated by using paired t-tests. No statistically significant differences (P >.05) were Brefeldin_A found. The error of landmark identification and, digitizing of Dentofacial Planner prediction printouts and post-treatment tracings was estimated by digitizing twice the Dentofacial Planner predictions and by calculating error magnitude for all cephalometric variables. No statistically significant differences were found for any of the variables.