Validity and reproducibility of cephalometric measurements obtained from digital photographs of analogue headfilms 11 4 Stom atolog ija, Bal ti c Dental an d Max ill of aci al Journ al, 2007, Vol . 9, N o. 4 Validity and reproducibility of cephalometric measurements obtained from digital photographs of analogue headfilms Simonas Grybauskas, Irena Balciuniene, Janis Vetra SCIENTIFIC ARTICLES Stomato log ija, Baltic Dental and Maxillo facial J o urnal, 9 :114 -12 0 , 2 00 7 SUMMARY The emerging market of digital cephalographs and computerized cephalometry is over- whelming the need to examine the advantages and drawbacks of manual cephalometry, meanwhile, small offices continue to benefit from the economic efficacy and ease of use of analogue cephalograms. The use of modern cephalometric software requires import of digital cephalograms or digital capture of analogue data: scanning and digital photography. The validity of digital photographs of analogue headfilms rather than original headfilms in clinical practice has not been well established. Digital photography could be a fast and inexpensive method of digital capture of analogue cephalograms for use in digital cephalometry. AIM. The objective of this study was to determine the validity and reproducibility of measurements obtained from digital photographs of analogue headfilms in lateral cephalometry. MATERIAL AND METHODS. Analogue cephalometric radiographs were performed on 15 human dry skulls. Each of them was traced on acetate paper and photographed three times independently. Acetate tracings and digital photographs were digitized and analyzed in cephalometric software. Linear regression model, paired t-test intergroup analysis and coefficient of repeatability were used to assess validity and reproducibility for 63 angular, linear and derivative measurements. RESULTS AND CONCLUSIONS. 54 out of 63 measurements were determined to have clinically acceptable reproducibility in the acetate tracing group as well as 46 out of 63 in the digital photography group. The worst reproducibility was determined for measurements dependent on landmarks of incisors and poorly defined outlines, majority of them being angular measurements. Validity was acceptable for all measurements, and although statistically significant differences between methods existed for as many as 15 parameters, they appeared to be clinically insignificant being smaller than 1 unit of measurement. Validity was acceptable for 59 of 63 measurements obtained from digital photographs, substantiating the use of digital photography for headfilm capture and computer-aided cephalometric analysis. Key words: cephalometry, reproducibility, dry skull, acetate tracing, digital photography. 1Institute of Odontology, Faculty of Medicine, Vilnius University, Vilnius, Lithuania 2Institute of Anatomy and Anthropology, Riga Stradins University, Riga, Latvia Simonas Grybauskas1 – D.D.S., MOS RCSEd Irena Balciuniene1 – D.D.S., PhD, Dr. habil. med., prof. Janis Vetra2 – M.D., Dr.habil.med., prof. Address correspondence to Dr. Simonas Grybauskas, Institute of Odontology, Faculty of Medicine, Vilnius University, Zalgirio 115, Vilnius LT-08217, Lithuania. E-mail: simonas.grybauskas@gmail.com nience in generation of treatment predictions have con- tributed to a shift from manual tracing on acetate pa- per towards digital computer-aided cephalometry [4]. Digital cephalometry has offered even more advan- tages, i.e., option to manipulate the image for size and contrast, image enhancement, ability to archive and improve access to images, superimposition of images [5]. Moreover, patients benefit from reduced dose of radiation if a digital cephalograph is chosen for image capture, whereas the lack of user-sensitive chemical development process and instantaneous image forma- tion save both space and time in the clinician’s prac- tice [6]. By now, many offices have not yet switched to the use of digital cephalographs, therefore the digiti- INTRODUCTION Variety of emerging computer software for lat- eral cephalometry in clinical orthodontics simplified the analysis and reduced time needed to perform certain measurements [1,2,3]. The ease of use and ability to perform several analyses at a time as well as conve- Stom atolog ija, Bal ti c Dental an d Max ill of aci al Journ al, 2007, Vol . 9, N o. 4 11 5 S. Grybauskas et al. SCIENTIFIC ARTICLES zation process of analogue head films is the only op- tion if the benefits of digital cephalometric analysis are anticipated. The two known methods of headfilm capture are scanning and digital photography. Stud- ies have shown that images captured from flatbed scanner can be reliable as compared to their corre- sponding analogue headfilms for use in clinical prac- tice, not so good for research [7-11]. Little data ex- ists on the reliability of images captured by means of digital photography – a poorly documented operator- sensitive technique with some speculations on distor- tion of images [12]. Computer-aided cephalometry and digitizing process of analogue headfilms were reported by numerous authors [2,8,12-21]. However, results of comparison of digitizing methods with ana- logue measurement methods were contradictory [2,9,14,18,22,23]. The aim of this study was to evaluate validity and reproducibility of measurements obtained from digital photographs of headfilms as compared to those obtained from traditional acetate paper tracings. Vali- dation of digital photography can enable its use in digital capture of analogue data for computer-aided cephalometric analysis without need for specific hard- war e. MATERIAL AND METHODS A set of 15 human dry skulls was obtained from the Department of Anatomy, Histology and Anthro- pology at Vilnius University. The skulls were chosen according to following criteria: occlusion was stable and reproducible with at least three pairs of antagonist teeth; posterior occlusal height was present; at least one of the condyles was intact and fit into glenoid fossa. The mandible part was related to the maxilla of its couterpart skull on the basis of occlusal interdigitation or maximal contact, and condylar seating in the gle- noid fossa. Since soft tissue components of the tem- poromandibular joint (TMJ) were missing on dry skulls, interpositional items were used to support the condyles in the center of the glenoid fossa preventing them from contact with the bone surface, thus, mimicking natural intra-articular space. Subsequently, the mandible was secured in this position with scotch tape around the skull. Fifteen lateral cephalograms were performed on the series of skulls by securing skulls in the cephalostat (Moviplan 8000 CE, Villa Sistemi Medicali, Italy) with the ear rods in the external auditory meati, and the distance between film and midsagittal plane at 13 cm. Preliminary work led to using the following radiographic setting: 77 kVp, 12 mA, 0.10 s. Headfilms were traced on acetate paper for three times with one week interval between independent tracings by the same operator, hence, the acetate trac- ing group was composed of 45 cephalometric acetate tracings. Two ruler points were marked on every trac- ing 108 mm apart. Following this, headfilms stayed on the view box and 15 digital pictures were taken (Canon 350D, Macro lens 100mm f2.8; 5Mp resolu- tion, image resolution 3200 x 2400 pixels) at a right angle from a distance of 2 meters. A transparent ruler of 108 mm was present on the radiograph whereas the camera was secured on the tripod when taking pictures of every radiograph. Three pictures were taken for every headfilm and the camera was dis- mounted a nd remounted a fter ever y picture to immitate independent attempts. The digital photogra- phy group was composed of 45 digital photographs of lateral headfilms. Digital pictures were imported into Dolphin 9.0 cephalometric software (Doplhin Imaging, USA) and digitizing procedure was performed on the series of 15 triplets of digital pictures. Images were sharpened, saturated, contrasted and brightened if needed to achieve best visibility of landmarks. Acetate tracings were stuck to the computer screen (hardware: IBM T60p, 1.8GHz, 2GB RAM, ATI Mobility FireGL V5200, screen resolution 1600x1200 dpi, 32bit color quality) with scotch tape and identical digitizing pro- cedure was performed on every tracing. The magni- fication factor was known to be 1.08 for the given cephalograph, therefore, 108 mm distance between ruler points was attributed to 100 mm distance on the software. The error was inherent in landmark identification process and was known to be variable depending upon the clarity of nature and definition of landmarks [9]. Hand measuring was abandoned in this study. Instead, once the digitizing procedures were finished for the 3 sets of acetate tracings and 3 sets of digital photo- graphs, software generated 6 sets of linear and angu- lar measurements that were exported and used to as- sess reproducibility and validity of digital photographs of headfilms (Table 1). Since measurements were gen- erated in automatic fashion by the software, no mea- suring errors were introduced in this part of the study. Data was imported and statistical analysis processed with SPSS 15.0. Assessment of reproducibility. Bla nd a nd Altman’s formula (1999) was used for the statistical analysis of reproducibility to determine coefficient of repeatability of every measurement for two different methods ( ∑ = ××−= n i ii s n R 1 21 1 ν ) [24]. Measurements were ranked as reproducible if both R coefficient and standard deviation (SD) of dif- 11 6 Stom atolog ija, Bal ti c Dental an d Max ill of aci al Journ al, 2007, Vol . 9, N o. 4 SCIENTIFIC ARTICLES S. Grybauskas et al. ferences from the average values were less than 1 unit of measurement. One “unit of measurement” in this study was an equivalent of one millimeter, one degree or one per- cent. It was used as a substitute in order to avoid rep- etition of bulky explanations of reproducibility for lin- ear angular and derivative measurements. The esti- mated reproducibility in this study was classified into four groups: ultra high reproducibility of measurements (R value and SD of differences smaller than 0.5 units), high reproducibility (R and SD of differences greater than 0.5 unit but smaller than 1 unit), moderate repro- ducibility (R value and SD of differences are between 1 and 2 units), and poor reproducibility (R value and SD of differences greater than 2 units). Mean of dif- ference is twice the mean of differences from the av- erage, therefore the limit of 2 units for R value was considered to be the range of clinical acceptance. All differences were taken for absolute numbers in this study. Assessment of validity. Validity was rated as ac- ceptable or non-acceptable in this study. Validity of measurements obtained from digital photographs were considered to be acceptable provided both of the two following conditions were met: first, paired t-test analy- sis revealed no statistically significant intergroup dif- ference (P>0.05) that would also be clinically signifi- cant (both mean and SD of differences greater than 2 units) between measurements obtained from digital photographs and those obtained from acetate tracings. Second, linear regression analysis showed strong cor- relation between methods: the intraclass correlation co- efficient r>0.75 ) )var()var()var( )var( ( errormethodskull skull r ++ = ; standardized beta coefficient >0.7 and confidence in- tervals for beta contained value 1; there was no sys- tematic offset in values and confidence intervals for alpha contained 0 value. The use of linear regression was essential in testing the agreement between two series of paired measurements that were shown to have few statistically significant differences between means, nevertheless, could have poor agreement [25]. Acetate tracing was the independent method, whereas digital photography was a dependent cephalometry method in linear regression model. RESULTS Reproducibility of measurements obtained from acetate tracings Fifteen (23.81%) out of 63 measurements used in lateral cephalometry were highly reproducible, with the standard deviation (SD) of differences of measure- ments being less than 0.5 unit (one unit equals one millimeter, one degree or one percent). Eight of them (12.70%) were characterized by ultra small R coeffi- cient (<0.5 unit) whereas 7 measurements by a small R value (0.5-1 of a unit). Thirty two (50,79%) mea- surements fell into moderate level of R value and SD of differences of 1 unit, 8 measurements demonstrated SD lower than one unit with R exceeding one unit. Nine (14.29%) parameters demonstrated both R and SD of differences being beyond 2 units of measure- ment. Reproducibility of measurements obtained from digital photographs of headfilms Eleven (17.46%) out of 63 measurements used in digital photography group were characterized by ultra high reproducibility with both R value and SD of differences being smaller than 0.5 of a unit of mea- surement. Twenty seven (42.86%) of 63 measure- ments showed high reproducibility with both R coef- ficient and SD of differences being smaller than 1 unit, 8 more measurements demonstrated SD lower than 1 unit, however R values were higher than 2 units. Seventeen (26.98%) of 63 measurements showed R values greater than 2 units, and four (6.35%) of them were characterized by SD of dif- ferences greater than 2 units. Overall characteristics of least reproducible measurements is presented in Table 2. Validity of measurements obtained from digi- tal photographs of lateral headfilms Validity was acceptable for all measurements except LI/Occ, S-Go, UFH/TFH and N-ANS (Table 3). There was a high correlation between methods for 59 out of 63 measurements: linear regression model showed interclass correlation coefficient r>0.8; standardized coefficient Beta>0.9; confidence inter- vals for Alpha and Beta values contained values 1 and 0 respectively. Non-acceptable validity was de- termined for 4 measurements: LI/OC, S-Go, UFH/ TFH and N-ANS. In 60 out of 63 lateral cephalom- etric measurements differences between the two methods were less than 0.5 units and less than 1 unit in the rest three measurements. There were no sta tistica lly significa nt differ ences between measurements obtained from digital photographs of lateral headfilms and corresponding acetate cephalo- metric tracings in 49 measures. A list of measure- ments for which paired t-test analysis and linear regression analysis showed statistically significant differences or poor correlation between the two methods is presented in Table 3. Stom atolog ija, Bal ti c Dental an d Max ill of aci al Journ al, 2007, Vol . 9, N o. 4 11 7 S. Grybauskas et al. SCIENTIFIC ARTICLES Table 1. Linear, angular and derivative measurements used for cephalometric analysis Measurement Definition Cranial base dimensions Linear measurements S-N Anterior cranial base length S-Ar Distance between sella and articulare S-Ba Posterior cranial base length Ba-N Total cranial base length Angular measurements N/S/Ar Angle between S-N and S-Ar lines N/S/Ba Cranial base saddle angle between S-N and S- Ba SN/FH Angle determined by S-N and Frankfort horizontal (FH) plane Facial height Linear measurements ANS-N UAFH, upper anterior facial height ANS-Me LAFH, lower anterior facial height N-Me TAFH, total anterior facial height LAFH/TAFH Ratio of lower anterior face height to total anterior face height UAFH/LAFH Ratio of upper facial height to lower facial height S-Go TPFH, total posterior face height S-PNS Posterior midfacial height Ar-Go Lower posterior facial height Jarabak ratio Ratio of total posterior to total anterior face height Vertical relationship Angular measurements SN/PP Angle determined by SN and Palatal plane SN/MP Angle determined by SN and Mandibular plane SN/OP Angle determined by SN and Occlusal plane FH/PP Angle between FH and Palatal plane FH/MP Angle determined by FH and Mandibular plane FH/OP Angle determined by FH and Occlusal plane MP/PP Angle determined by Mandibular and Palatal planes PP/OP Angle determined by Palatal and Occlusal planes Relationship of the maxilla to the cranial base Linear measurements A-Nv Distance from point A to Nv line Co-A Distance from Condylion to A point Ar-A Distance from Articulare to A point Ba-A Basialveolar length A-NPog Distance from point A to facial plane line Angular measurements S/N/A Angle determined by S-N and N-A lines NA/FH Angle determined by N-A line and FH plane Relationship of the mandible to the cranial base Linear measurements B-Nv Distance from point B to Nv line Angular measurements S/N/B Angle determined by S-N and N-B lines S/N/Pog Facial angle determined between S-N and facial plane lines FH/NPog Facial angle determined between FH plane and facial plane linbe N/S/Gn Y-axis, the angle determined by S-N and S-Gn lines S/Ar/Go Articulare angle, determined by S-Ar and Ar- Go lines Measurement Definition Relationship of the maxilla to the mandible Linear measurements Wit’s upraisal Distance between the projections of point A and B onto occlusal plane Angular measurements A/N/B Angle determined by N-A and N-B lines N/A/Pog Convexity angle, determined by N-A and A-Pog lines A/Ar/Gn Angle 1 from the A-Ar-Gn triangle A/Gn/Ar Angle 2 from the A-Ar-Gn triangle Ar/A/Gn Angle 3 from the A-Ar-Gn triangle Relationship of the maxillary dentition to the maxilla and the cranial base Linear measurements UIE-NA Distance from Upper incisor tip to N-A line UIE-APog Distance from Upper incisor tip to A-Pog line Angular measurements U1/NA Angle determined by maxillary incisor axis and N-A lines U1/FH Angle determined by maxillary incisor axis and FH plane U1/SN Angle determined by maxillary incisor axis and SN line U1/PP Angle determined by maxillary incisor axis and Palatal plane U1/OP Angle determined by maxillary incisor axis and Occlusal plane Relationship of the mandibular dentition to the mandible and the cranial base Linear measurements LIE-APog Distance from Lower incisor tip to A-Pog line LIE-NB Distance from Lower incisor tip to N-B line Angular measurements LI/MP Angle determined by Mandibular incisor axis and Mandibular plane LI/NB Angle determined by Mandibular incisor axis and N-B line LI/OP Angle determined by Mandibular incisor axis and Occlucal plane Relationship of the maxillary dentition to the mandibular dentition Angular measurements UI/LI Angle determined by Maxillary incisor axis and Mandibular incisor axis Maxillary or palatal dimensions Linear measurements ANS-PNS Palatal length, distance from Anterior nasal spine to Posterior nasal spine A-PNS Distance from posterior nasal spine to A point Mandibular length Linear measurements Go-Gn Length of mandibular corpus, distance between Gonion and Gnathion points Go-Co Ramus height, distance between Gonion and Condilion points Co-Gn Length of mandibular base, distance between Condilion and Gnathion points (Co-Gn)-(Co- A) Maxillo-mandibular length difference – difference between Co-Gn and Co-A values Angular measurements Co/Go/Gn Gonial angle, determined by Go-Co and Go-Gn lines 11 8 Stom atolog ija, Bal ti c Dental an d Max ill of aci al Journ al, 2007, Vol . 9, N o. 4 DISCUSSION In digital photography group the worst reproduc- ibility was seen for U1/FH, U1/L1, B-Nv and articu- lar angle (S-Ar-Go), followed by FH/OP, FH/OP, FH/ NPog, UI/SN, UI/FH, UI/OP. In the acetate tracing group poor reproducibility was determined for the measurements U1/SN, U1/PP, U1/OP, U1/NA, U1/ FH; L1/OP, L1/NB, L1/GoGn, U1/L1. Obviously, ma- jority of theses measurements depend on landmarks and references of incisor teeth and poorly defined outlines or low contrast area such as Articulare, Go- nion, PNS and Porion. Our data agrees with results reported by Chen et al (2000), who stated that least reliable landmarks are those that are located on curved anatomical boundaries or on axis on teeth, thus re- sulting in greatest inaccuracies of following measure- ments: U1/SN, U1/L1, L1/OP, L1/MP [9]. Our data is also in line with Baumrind and Frantz (1971a) who described “errors in identification” being specific for different landmarks and arising from inability to lo- cate anatomical landmarks [26]. Definition was later expanded by Vincent et al (1997) who classified er- rors of identification caused by: poor outline of the curvature of the line upon which the landmark is po- SCIENTIFIC ARTICLES S. Grybauskas et al. sitioned; contrast of the area; noise and superimposi- tion of other structures; poor definition of the land- mark [6]. In the acetate tracing group all nine measure- ments with poor reproducibility were angular, as well as 3 out of 4 in the analogue cephalometry group. Angular measurements showed worse reproducibil- ity than linear measurements and it is line with stud- ies conducted by Baumrind and Frantz as well as Savinsu et al [27,10]. The comparative analysis showed that there were few statistically significant differences between meth- ods, however all of them were clinically insignificant with mean and SD of differences smaller than 0.5 unit, thus substantiating the use of digital photogra- phy and tracing of digital photographs in orthodontic practice. According to linear regression model, the validity of measurements obtained from digital pho- tographs was acceptable: r>0.8, standardized beta coefficient >0.9 and confidence intervals for alpha and beta values were containing values 0 and 1 re- spectively (p<0.05) for majority of measurements (poor correlation between groups for 4 measurements needs further investigation). It is in agreement with Chen et al (2004) and Schulze et al (2002) results Table 2. Characteristics of least reproducible measurements obtained from digitized acetate tracings and digital photographs of headfilms Acetate tracing group Digital photography group Confidence interval of R (95%) Confidence interval of R (95%) Measurement R value SD of differences lower bound upper bound R value SD of differences lower bound upper bound Articular Angle S-Ar-Go(º) 1.91 1.52 -1.07 4.88 3.39 2.28 -1.07 7.85 B-Nv (mm) 0.92 0.75 -0.56 2.40 3.37 2.51 -1.54 8.28 Facial Angle (FH-NPog) (º) 0.59 0.49 -0.37 1.56 2.09 1.56 -0.96 5.14 FH / MP (º) 1.13 0.86 -0.57 2.82 2.25 1.59 -0.87 5.38 FH / OP (º) 1.44 1.04 -0.61 3.49 2.57 1.87 -1.09 6.23 FH / PP (º) 0.78 0.61 -0.42 1.98 2.15 1.59 -0.97 5.26 Interincisal Angle (UI/LI) (º) 4.40 4.22 -3.87 12.67 3.11 2.19 -1.17 7.40 LI / GoGn (º) 2.96 2.50 -1.94 7.86 2.12 1.45 -0.73 4.97 LI / NB (º) 2.65 2.43 -2.11 7.40 1.99 1.36 -0.68 4.67 LI / Occ Plane (º) 2.49 2.25 -1.92 6.89 2.08 1.54 -0.94 5.10 Lower Posterior Facial Height Ratio (Ar-Go/S-Go x 100) (%) 1.63 1.35 -1.03 4.28 2.52 1.87 -1.15 6.19 Mandibular Body Length (Go-Gn)(mm) 2.04 1.65 -1.18 5.26 2.67 1.85 -0.97 6.30 Maxillary Depth FH / NA (º) 0.83 0.65 -0.45 2.12 2.05 1.54 -0.98 5.07 U1 / FH (º) 3.32 2.69 -1.94 8.58 3.07 2.56 -1.94 8.09 U1 / NA (º) 3.03 2.65 -2.16 8.22 2.27 1.70 -1.05 5.60 U1 / Occ Plane (º) 2.71 2.43 -2.05 7.48 2.28 1.57 -0.80 5.36 U1 / Palatal Plane (º) 3.41 2.68 -1.85 8.66 2.47 1.71 -0.87 5.82 U1 / SN (º) 3.20 2.70 -2.08 8.48 2.15 1.59 -0.98 5.27 Stom atolog ija, Bal ti c Dental an d Max ill of aci al Journ al, 2007, Vol . 9, N o. 4 11 9 S. Grybauskas et al. SCIENTIFIC ARTICLES stating that although statistically significant differ- ences between digitized and analogue measurements existed, they were clinically insignificant [28,29]. It also agrees with the study conducted by Macri and Wenzel (1993) who stated that it was possible to achieve reliability of digital images comparable to that obtained with conventional equipment for radiographs of good quality [2]. Collins et al (2007) compared mea s ur ements fr om p hotogr a phed la t er a l cephalograms and scanned cephalograms and found statistically significant differences in linear measure- ments by using Dolphin software [30]. Although digi- talization of acetate tracings rather than scanning was used in our study, 11 out of 15 measurements that were shown to have statistically significant differ- ences were linear measurements suggesting a need for more thorough investigation of magnification fac- tors in computer-aided cephalometry. Table 3. Intergroup comparison of measurements obtained from digital photographs of analogue cephalograms and corre- sponding acetate tracings Measurement Compa- red methods Mean of diffe- rence SD of diffe- rence t Sig. (2- tailed) Unstandardized Coefficient Alpha Standardized Coefficient Beta Sig. 95% Confidence Interval for Alpha and Beta A-Gn-Ar (Angle 3) (°) Dig photo -0.09629 0.153659 2.516 0.036 -1.159 0.171 -2.955 0.637 Acetate 1.023 0.999 0 0.99 1.055 Anterior Cranial Base (S-N) (mm) Dig photo 0.31524 0.193747 -7.246 0 0.631 0.298 -0.697 1.958 Acetate 0.986 1 0 0.966 1.006 Ba - A (mm) Dig photo 0.4199 0.187261 -7.536 0 -0.491 0.664 -3.053 2.07 Acetate 1.001 1 0 0.972 1.029 L1 - Occ Plane (°) Dig photo -0.18518 0.699007 0.828 0.432 -3.478 0.01 -5.846 -1.11 Acetate 1.052 0.999 0 1.019 1.086 Lower Facial Height (ANS-Me) (mm) Dig photo 0.20422 0.292465 -3.118 0.014 0.222 0.867 -2.803 3.247 Acetate 0.993 0.999 0 0.947 1.04 Mandibular length (Co-Gn) (mm) Dig photo 0.50009 0.595312 -5.551 0.001 -0.304 0.897 -5.638 5.03 Acetate 0.998 0.999 0 0.953 1.044 Midfacial length Co- A (mm) Dig photo 0.40611 0.463005 -7.176 0 -0.408 0.711 -2.903 2.087 Acetate 1 0.999 0 0.971 1.029 Mx/Md diff (Co-Gn – Co-A) (mm) Dig photo 0.22033 0.259186 -5.897 0 -0.338 0.39 -1.21 0.535 Acetate 1.005 0.999 0 0.969 1.041 N - Ba (mm) Dig photo 0.46236 0.509173 - 10.128 0 0.255 0.757 -1.614 2.123 Acetate 0.993 1 0 0.974 1.012 PNS-A (mm) Dig photo 0.19733 0.274228 -3.341 0.01 0.63 0.539 -1.679 2.94 Acetate 0.983 0.999 0 0.935 1.031 Posterior Cranial Base (S-Ar) (mm) Dig photo 0.10776 0.157548 -2.981 0.018 -0.02877 0.943 -0.944 0.886 Acetate 0.998 1 0 0.972 1.023 Posterior Cranial Base (S-Ba) (mm) Dig photo 0.22644 0.255059 -7.913 0 -0.859 0.065 -1.786 0.068 Acetate 1.015 1 0 0.993 1.038 Posterior Face Height (SGo) (mm) Dig photo 0.27545 0.695934 -1.308 0.227 4.889 0.02 1.044 8.734 Acetate 0.934 0.998 0 0.885 0.983 Saddle/Sella Angle (SN-Ar) (°) Dig photo -0.18889 0.271058 3.104 0.015 -1.045 0.544 -4.92 2.83 Acetate 1.01 0.999 0 0.979 1.04 SN - MP (°) Dig photo 0.24074 0.436686 -2.494 0.037 0.333 0.468 -0.694 1.36 Acetate 0.979 0.999 0 0.943 1.016 Total Face Height (N-Me) (mm) Dig photo 0.46666 0.548392 -5.93 0 1.395 0.385 -2.168 4.957 Acetate 0.983 0.999 0 0.952 1.015 UFH/TFH (N- ANS:N-Me) (°) Dig photo 0.05556 0.15411 -1.17 0.276 2.127 0.018 0.488 3.767 Acetate 0.949 0.999 0 0.911 0.987 Upper Face Height (N-ANS) (mm) Dig photo 0.26003 0.350888 -3.596 0.007 2.012 0.009 0.671 3.353 Acetate 0.953 0.999 0 0.925 0.981 12 0 Stom atolog ija, Bal ti c Dental an d Max ill of aci al Journ al, 2007, Vol . 9, N o. 4 Received: 16 09 2007 Accepted for publishing: 21 12 2007 REFERENCES 1. Chen SK, Chen YJ, Yao CC, Chang HF. Enhanced speed and precision of measurement in a computer-assisted digital cephalometric analysis system. Angle Orthod 2004;74:501-7. 2. Macri V, Wenzel A. Reliability of landmark recording on film and digital lateral cephalograms. Eur J Orthod 1993;15:137- 48. 3. Power G, Breckon J, Sherriff M, McDonald F. Dolphin Imaging Software: an analysis of the accuracy of cephalometric digitization and orthognathic prediction. Int J Oral Maxillofac Surg 2005; 34:619-26. 4. Ongkosuwito EM, Katsaros C, Bodegom JC, Kuijpers-Jagtman AM. Digital cephalometrics. Ned Tijdschr Tandheelkd 2004; 111:266-70. 5. Forsyth DB , Davis DN. Assessment of an automated cephalometric analysis system. Eur J Orthod 1996; 18:471-8. 6. Vincent AM, West VC. Cephalometric landmark identification error. Austral Orthod J 1987;10:98-104. 7. Bassignani MJ, Bubash-Faust L, Ciambotti J, Moran R, McIlhenny J. Conversion of teaching file cases from film to digital format: a comparison between use of a diagnostic-quality digitizer and use of a flatbed scanner with transparency adapter. Acad Radiol 2003;10:536-42. 8. Bruntz LQ, Palomo JM, Baden S, Hans MG. A comparison of scanned lateral cephalograms with corresponding original radiographs. Am J Orthod Dentofacial Orthop 2006; 130:340- 8. 9. Chen YJ, Chen SK, Chang HF, Chen KC. Comparison of landmark identification in traditional versus computer-aided digital cephalometry. Angle Orthodontist 2000; 70: 387-92. 10. Sayinsu K, Isik F, Trakyali G, Arun T. An evaluation of the errors in ceph alom etric m easu rements on scann ed cephalometric images and conventional tracings. Eur J Orthod 2007;29:105-8. 11. Turner PJ, Weerakone S. An evaluation of the reproducibility of landmark identification using scanned cephalometric images. J Orthod 2001;28:221-9. 12. Nimkarn Y. Miles PG. Reliability of computer generated cephalometrics. Int J Adult Orthod Orthogn Surg 1995;10:43- 52. 13. Davis DN, MacKay F. Reliability of cephalometric analysis using manual and interactive computer methods. Br J Orthod 1991;18:105-9. 14. Geelen W, Wenzel A, Gotfredsen E, Kruger M, Hansson L-G. Reproducibility of cephalometric landmarks in conventional film, hardcopy and monitor-displayed images obtained by the storage phosphor technique. Eur J Orthod 1998; 20:331–40. 15. Houston WJ. The application of computer aided digital analysis to orthodontic records. Eur J Orthod 1979;1:71-9. 16. Lim KF, Foong KW. P hosphor stim ulated compu ted cephalometry: reliability of landmark identification. Br J Orthod 1997;24:301-8. 17. Liu J-K, Chen YC, Cheng KS. Accuracy of computerized automatic identification of cephalometric landmarks. Am J Orthod Dentofac Orthop 2000; 118:535-40. 18. Oliver RG. Cephalometric analysis comparing five different methods. Br J Orthod 1990;27:277-83. 19. Richardson A. A comparison of traditional and computerized methods of cephalometric analysis. Eur J Orthod 1981;3:15- 20. 20. Rudolph DJ, Sin clair PM , Coggin s JM . Au tomatic compouterized radiographic identification of cephalometric landmarks. Am J Orthod Dentofac Orthop 1989; 113:173-9. 21. Stirrups DR. A comparison of the accuracy of cephalometric landmark location between two screen/ film combinations. Angle Orthodontist 1989;59:211-5. 22. Cohen JM. Comparing digital and conventional cephalometric radiographs. Am J Orthod Dentofacial Orthop 2005;128:157- 60. 23. Santoro M, Jarjoura K, Cangialosi TJ. Accuracy of digital and analogue cephalometric measurements assessed with the sandwich technique. Am J Orthod Dentofacial Orthop 2006;129:345-51. 24. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8:135-60. 25. Stöckl D, Dewitte K, Thienpont LM. Validity of linear regression in method comparison studies: is it limited by the statistical model or the quality of the analytical input data? Clin Chem 1998;44:2340–6. 26. Baumrind S, Frantz RC. The reliability of head film measurements. Landmark identification. Am J Orthod 1971a; 60:111-27. 27. Baumrind S, Frantz RC. The reliability of head film measurements. Conventional angular and linear measures. Am J Orthod 1971b;60:505–17. 28. Chen YJ, Chen SK, Yao JC, Chang HF. The effects of differences in landmark identification on the cephalometric measurements in traditional versus digitized cephalometry. Angle Orthod 2004;74:155-61. 29. Schulze RK, Gloede MB, Doll GM. Landmark identification on direct digital versus film-based cephalometric radiographs: a human skull study. Am J Orthod Dentofacial Orthop 2002;122:635–42. 30. Collins J, Shah A, McCarthy C, Sandler J. Comparison of measurements from photographed lateral cephalograms and scanned cephalograms. Am J Orthod Dentofacial Orthop 2007;132:830-3. SCIENTIFIC ARTICLES S. Grybauskas et al. CONCLUSIONS 1. Both measurements obtained from acetate trac- ings and digital photographs of analogue cephalograms were shown to have adequate reproducibility with both R coefficients and SD of differences smaller than 2 units of measurement. Nine measurements in the acetate cephalometry group and seventeen in the analogue cephalometry groups failed to go within this limit and were shown to be less reproducible. 2. Majority of poorly reproducible measurements were angular or associated with least reproducible landmarks and references. 3. Validity of 59 out of 63 lateral measurements ob- tained from digital photographs was acceptable, thus, substantiated the use of digital photography for headfilm capture, digital tracing and computer-aided cephalom- etric analysis.