When a scan or photocopy of the plaster model is properly taken, the produced photograph is expected to be perpendicular to the occlusal plane. This matter does not exist with the intra-oral photography of the dental arches using hand-held cameras and may introduce errors of projection. The angulation at which the intra-oral photograph is taken may vary so that a vertical component of the contact points displacement can be introduced to the measurement.25 Additionally, photographs may also have errors of magnification, so the calibration process (correction of magnification) must be taken into account before any measurement is carried out on a digital photo.
The study model measurements were considered as the gold standard for testing photograph-based measurements validity. Indeed, manual measuring on study models may suffer from some inherent errors (e.g., vibrations and sometimes calipers cannot reach the exact interproximal contact point of a tooth when that tooth is in contact with other teeth) .4, 26
Differences between the evaluated two methods of measuring that were greater than 0.25 mm were considered clinically important for small measurements (e.g., LII) and the threshold of clinically significant differences was raised to 1 mm for those greater measurements in the orthodontic study model analyses (e.g., TSALD, BOLTON ratios and sum of tooth size) .
In contrast to the current study, the previous papers 23−25 tested the validity of photographs when measuring LII only. Orthodontic diagnosis and treatment planning depends on several dental and alveolar analyses and the assessment of LII solely is not enough for decision making and outcomes’ analysis in the daily practice. Therefore, this study was accomplished to cover a wide spectrum of analyses that evaluate dental arches in the context of orthodontic evaluation.
Almasoud and Bearn used a millimeter ruler placed with wax onto the cusp tips of the premolars in the same level of the occlusal plane crossing the dental arch horizontally. This ruler helped in correcting the magnification of intra-oral photographs but was not practically easy to use since the applied wax may have not provided good and stable contact with the teeth as well as being cumbersome to the evaluated patients. Additionally, the fixation of the ruler at the middle of the dental arch prevented the ability of measurements in the posterior regions, so less chance to apply several dental arch analyses. Therefore, in the current trial, the ruler was fixed on the periphery of the lip retractor enabling the operator to take photos without any additional time-wasting procedures.
Validity and reliability of measurements made on intra-oral images
The differences in teeth sizes, basal bone length, LII, and partial Bolton measured on intra-oral images by Examiner A were statically significant when compared to those obtained from study models. However, the differences were equal to or less than 0.25 mm for the LII and 0.45 mm for the other variables which appeared to be of no clinical importance. On the other hand TSALD, overall Bolton value that measured by Examiner A and all variables measured by Examiner B did not show any statistically significant differences from those obtained from study models indicating that all measurements on intra-oral photographs were of a high, and sometimes a very good, level of agreement with those of plaster models. The current findings are similar to those of Almasoud and Bearn who evaluated only the LII measurement. Nevertheless, the current 95% confidence intervals of mean differences were narrower than those of Almasoud and Bearn. This may be due to differences in the calibration method, camera and lens settings, or mirror reflecting properties that may have affected picture resolution.
Validity and reliability of measurements made on plaster models' images
The comparison between measurements made on plaster models and the corresponding images of these study models showed high ICCs which were greater than 0.88 and the differences were less than 0.19 mm for LII and 0.48 mm for the other variables. Since the differences in the LII and TSALD measurements were less than the aforementioned clinical thresholds, good validity and reliability could be inferred. These findings are consistent with those of Almasoud and Bearn, Mushtaq et al and Tran et al who employed digital images of study models in their assessments.23−25 Tran et al reported that the assessment of LII in Class II crowded mandibular arches was very reliable and valid; unfortunately, they presented only correlation coefficients without performing significance testing. 24
Intra- and inter-examiner reliability of the different measuring techniques
The analysis of intra-examiner reliability of measurements made on plaster models revealed high ICCs. On the other hand, significant differences smaller than 0.27 mm (and 0.17 mm for LII) were found between the repeated measurements of one examiner (i.e. Examiner A) regarding LII and TSALD values but these differences could be neglected since they do not affect the clinical decision.
The intra-examiner reliability of measurements made on intra-oral images of dental arches showed high ICCs and all obtained values were greater than 0.85. In addition, mean differences were not significantly different between the two sets of data. This indicated a high agreement between the first and second measurements and high intra-examiner reliability.
The analysis of inter-examiner reliability of plaster models’ measurements and those of intra-oral photos’ showed high ICCs that ranged between 0.92 and 0.99. Significant differences with p-values less than 0.05 were found in teeth size and LII variables between the two examiners (i.e. Examiner A and Examiner B) on plaster models but these differences were less than 0.10 mm which did not have any clinical importance. No significant differences were identified between the two examiners regarding measurements made on intra-oral photographs. This indicated a high level of agreement between the first and second examiners when using the same intra-oral photographs and high inter-examiner reliability. This is similar to the previous reports by Almasoud and Tran.24, 25
Actually, intra-oral photographs method were better than direct measurement on study models in terms of intra-/inter-examiner reliability, particularly for the TSALD analysis due to obtained high ICC values. A similar conclusion was arrived at by Almasoud and Bearn regarding LII measurement. This can be attributed to the error introduced in manual measuring if any vibration occurs and secondly to the inability of calipers’ heads to rest at the exact interproximal contact points of neighboring teeth when they are in close contact or overlapped.
Reproducibility of measurements made on intra-oral images
The comparison between results obtained from the first and second intra-oral photographs showed a significant difference just for LII but the mean difference was lower than the aforementioned clinical threshold (i.e. 0.25 mm for LII).
In the light of the current results, measurements on intra-oral photographs of the dental arches were very accurate, reliable and repeatable when performing dental arch analysis. Additionally, they are suggested to replace study models and eliminate the need for repeated impressions during orthodontic treatment with fixed appliances in the context of prospective research projects.