Radiographic evaluation of orthodontic treatment by means of four different cephalometric superimposition methods

INTRODUCTION: Despite discussion on the merit of various cephalometric superimposition methods, there remains a need to assess which one can be used in daily practice with reasonably accuracy and less working time. OBJECTIVE: The aim of this study was to investigate four methods of cephalometric superimposition by means of assessing the longitudinal changes in craniofacial morphology caused by growth and response of adolescents with Class I malocclusion to orthodontic treatment involving first premolar extraction. METHODS: Pretreatment (T1) and post-treatment (T2) standardized lateral cephalometric radiographs of 31 adolescents (20 females and 11 males), with Angle Class I malocclusion and indication of premolar extraction, participated in this study. Radiographs were digitized, traced and had structures identified by means of a cephalometric software. Four superimposition methods were used: Björk structural method, Steiner/Tweed SN line, Ricketts N-Ba line at N-point and Ricketts N-Ba line at CC-point. Positional changes were quantified by horizontal and vertical linear changes in the following cephalometric landmarks: anterior/posterior nasal spine (ANS and PNS), gnathion (Gn), Gonion (Go), Pogonion (Pog), A-point and B-point. Differences between T1 and T2 in horizontal and vertical positional changes for all superimposition methods were assessed by one-way analysis of variance (ANOVA) and Bonferroni correction (p < 0.05). RESULTS: There were no statistically significant differences among the cephalometric superimposition methods or when patients' sex was considered. CONCLUSION: Björk structural method, Steiner/Tweed SN line, Ricketts N-Ba line at N-point and Ricketts N-Ba line at CC-point methods were reliable and presented similar precision when the overall facial changes due to active growth and/or orthodontic treatment were examined.


INTRODUCTION
The need to visualize and understand the behavior of craniofacial structures in response to orthodontic treatment and continuous growth and development has motivated the emergence of several cephalometric superimposition techniques. 1-6 Different anatomical structures, cephalometric landmarks, lines and planes of reference have been used for this purpose, allowing quantitative analysis of growth and treatment based on changes of the facial skeleton of a particular individual over a period of time. [7][8][9][10][11] Cephalometric superimposition is of great importance when assessing orthodontic-orthopedic treatment response and orthognathic surgery outcomes. [11][12][13][14] Pretreatment and post-treatment cephalometric tracings should be carefully superimposed in order to provide reliable assessment of orthodontic/ growth structural changes. 13 Longitudinal changes in craniofacial morphology caused by growth and treatment response can be measured by superimposing a series of lateral cephalograms, using relatively stable landmarks, such as cranial base, cranial points, lines or regional contours, as reference. 15 Several superimposition methods have been described in the literature. Björk and Skieller 4,5 state-ofthe-art structural superimposition method based on Björk implant studies on craniofacial growth has been used widely. Superimposition is made on specific anatomical bone structures. This method, however, relies on the quality of the radiograph, particularly with regard to optimal contrast and density. Steiner/Tweed SN line method, 2 Ricketts 6 N-Ba line at N-point and Ricketts N-Ba line at CC-point methods have also been described in the literature. Superimposition on the cranial base provides an overall assessment of growth and treatment changes of facial structures. It will not identify specific sites of growth, but will aid assessment of the amount and direction of maxillary and mandibular growth, as well as the overall displacement of teeth and associated soft tissue changes.
Superimposition methods have revealed deficiencies or difficulties in comparison to others. 16,17 Some studies 9,18 demonstrate inaccuracy of cephalometric superimposition methods; while others 13,19,20 suggest the use of more than one method in order to increase the procedure reliability, provide additional information and make it possible to assess sagittal skeletal and dental changes more accurately. Although some practitioners may opt to do so, it is time consuming and may not be ideal in private settings. Nevertheless, this is not to suggest that cephalometric superimposition is not a useful measurement tool used to assess the extent of dentofacial changes. Rather, studies 13,16,17 indicate that it may be used with sufficient degree of accuracy for clinical diagnosis and treatment.
The most significant limitation of cephalometric superimposition is that three-dimensional changes are measured in two dimensions. The advent of cone-beam computed tomography has provided new insights of three-dimensional changes induced by normal growth and orthodontic treatment. Will it be the substitute for the traditional superimposition methods used today? Nevertheless, the present emphasis on minimizing radiation exposure prevents the use of such diagnosis resource on routine conventional orthodontic practice. 21,22 Which superimposition method is best suited to assess changes caused by growth and/or orthodontic treatment response? Are superimposition methods equally accurate and reliable? The aim of this study was to assess four different methods of cephalometric superimposition by means of examining the results of Angle Class I treatment of growing individuals treated with upper and lower premolar extractions.

MATERIAL AND METHODS
The present study was approved by Universidade Federal de Goiás Institutional Review Board under protocol number 055/2007.

Sample selection
This retrospective observational study was conducted on pre (T 1 ) and post (T 2 ) treatment standardized lateral cephalometric radiographs of 31 adolescents (20 females and 11 males) with mean age of 13 years and 4 months at T 1 and 17 years and 6 months at T 2 . Radiographs were obtained from the archives of Universidade Federal de Goiás, School of Dentistry, postgraduate program in Orthodontics.
Patients had been referred to orthodontic treatment due to Angle Class I malocclusion, and indication of upper and lower premolar extraction due to severe crowding or dental protrusion in permanent dentition. Only individuals presenting high-quality original article Lenza MA, Carvalho AA, Lenza EB, Lenza MG, Torres HM, Souza JB lateral cephalometric radiographs at the beginning and end of treatment, which allowed clear visualization of dentoskeletal structures, soft tissue and facial contour, were included. These individuals did not present any systemic conditions that could hinder the results of the study.

Cephalometric superimposition analysis
All pre and post-treatment lateral cephalometric radiographs were digitized with a resolution of 150 dpi by means of a flatbed scanner (Hewlett-Packard Company, Palo Alto, Ca, USA) attached to a transparency reader. Images were saved in TIFF format. Each cephalogram was traced and all anatomical structures necessary for the superimposition methods were identified (Fig 1). Both tracing and superimposition procedures were performed by the same operator. Data were collected by means of Radiocef Studio 2.0 cephalometric software (Radiomemory, Belo Horizonte, Brazil).
Four superimposition methods were used to assess the positional changes caused by orthodontic treatment and associated growth, taking into account the stability of reference points, as well as their precision and visualization, and ease of the method: • Björk structural method (M1): radiographs were superimposed on the reference bone structures in the anterior cranial base, as described by Björk and Skieller. 4,5 Anterior contour of sella turcica wall, anterior contour of the median cranial fossa, the mean intersection point of the lower contours of the anterior clinoid processes, the inner surface of the frontal bone, contour of the cribriform plate, contours of the bilateral frontoethmoidal crests and contour of the median border of cerebral surfaces of the orbital roofs.
• Steiner/Tweed SN line method 2,3 (M2): radiographs were superimposed on the SN line with registration at the S-point.
• Ricketts N-Ba line at N-point method 6 (M3): radiographs were superimposed on the N-Ba line with registration at the N-point.
• Ricketts N-Ba line at CC-point method 6 (M4): radiographs were superimposed on the N-Ba line with registration at the CC-point (center of the cranium) where the Ba-N plane intersects the Ptm-gnathion line.
The positional changes assessed by the superimposition methods were quantified on the basis of horizontal and vertical linear changes in the following cephalometric landmarks: anterior nasal spine (ANS), posterior nasal spine (PNS), gnathion (Gn), gonion (Go), pogonion (Pog), A-point and B-point; following the criteria described by Baumrind and Frantz. 7,8,9 Post-treatment tracings were then superimposed on pre-treatment ones so as to quantify the horizontal and vertical positional changes according to each superimposition method. Vertical alterations were measured in millimeters with a line perpendicular to Frankfort horizontal plane (FHP), whereas horizontal alterations were also measured in millimeters with a horizontal line perpendicular to Nasion (Nperp) (Fig 2). Seven horizontal measurements A reference coordinate system was established for pre and post-treatment cephalometric radiographs. For the vertical measurements obtained by a line perpendicular to the Frankfort horizontal plane, the X-axis was used; all measurements below this axis were negative while those above it were positive. For the horizontal measurements obtained by a line perpendicular to the nasion-perpendicular (Nperp) line, the Y-axis was used; all measurements to the left of this axis were negative while those to the right were positive.
Raw numbers were registered in an Excel spreadsheet for later statistical calculations. The positional change of each landmark was then compared within this coordinate system for each superimposition method. 9 Pre and post-treatment differences determined the horizontal and vertical linear changes of the cephalometric landmarks.

Error of the method
Intraexaminer reliability was determined by reassessing ten randomly selected cephalometric radiographs (five pretreatment and five post-treatment) which were digitized, traced and measured by the same examiner twenty-one days after the first measurement. The difference between  first and second cephalometric measurements was determined for each radiograph, and casual error calculated by Dahlberg 23 formula: E 2 = Σ d 2 /2n, in which "d" represents the difference between the values obtained in the first and second measurements and "n" represents the number of cases in which measurements were repeated. Systematic error was calculated by paired t-test, according to Houston. 24 Significance level was set at P < 0.05.

Statistical analysis
Descriptive statistics (mean and standard deviation) were calculated for all measurements obtained by the superimposition methods in both observational periods. Paired t-test was used to assess the amount of displacement according to each superimposition method on pre and post-treatment cephalometric radiographs. Differences in the amount of horizontal and vertical positional changes between T 1 and T 2 were assessed by one-way analysis of variance (ANOVA) and Bonferroni test for all superimposition methods. Data were analyzed by SPSS for Windows (SPSS 17.0, SPSS Inc, Chicago, III). Differences were considered statistically significant at P < 0.05.

RESULTS
Changes caused by growth and/or orthodontic treatment, assessed by four cephalometric superimposition methods, were quantified by horizontal and vertical linear alterations in the following cephalometric landmarks: anterior nasal spine (ANS), posterior nasal spine (PNS), gnathion (Gn), gonion (Go), pogonion (Pog), A-point (subspinale) e B-point (supramentale). A total of seven horizontal and seven vertical measurements were carried out. The amount of positional changes was calculated by the difference between pre and posttreatment cephalometric measurements for each superimposition method.
The calculated reliability coefficient for examiner reliability was 0.97 for all measurements, which yields sufficient reliability. Tables 1, 2     Values = mean ± standard deviation. Student's t-test. p < 0.05 ANOVA (one-way). Bonferroni. NS = nonsignificant. Radiographic evaluation of orthodontic treatment by means of four different cephalometric superimposition methods original article in comparison to T 1 . Negative values mean that the cephalometric measurement at T 2 was lower than T 1 -that is, the cephalometric landmark is closer to the Nperp line (horizontal) or to Frankfort horizontal plane (vertical) at T 2 in comparison to T 1 .
One-way analysis of variance (ANOVA) and Bonferroni test were used to assess the mean values of each variable and, thus, indicate the positional change of all cephalometric landmarks according to each superimposition method. Paired t-test was used to assess the amount of displacement for every two superimposition methods over pre and post-treatment differences. Significance level was set at 5%.

DISCUSSION
Serial radiographic cephalometry has been used since it was simultaneously but independently discovered by Hofrath 25 and Broadbent 26 in 1931, as a means of measuring craniofacial changes caused by growth and/or treatment. However, its clinical application resulted in confusion and misunderstanding. For this reason, accuracy of cephalometric measurements has always been questioned. Validity and reliability of superimposition methods have been a source of research of different studies 9,18 investigating the magnitude of superimposition errors and the extent of potential positional changes of facial structures. 27 The principle behind cephalometric superimposition is to compare radiographs taken at different time intervals, most often pre and post-treatment or even longer, in which radiographs taken years post-retention are included. This comparison provides the orthodontist with a general overview of growth and/or treatment outcomes through changes of the facial skeleton by comparing linear and angular measurements on serial cephalograms of the same patient. However, lateral cephalometric radiographs taken at different time intervals and by different operators are difficult to reproduce with a satisfactory degree of accuracy.
Positional changes of the anatomical structures caused by growth and/or treatment are studied by means of cephalometric radiographs superimposed to one another on structures that are considered stable over a period of time. Broadbent 1 was the first to publish a technique for superimposition of successive cephalometric films which demonstrated a child's facial growth. Traditionally, the method of best fit has been used, meaning that bone structures that apparently do not change over time are used. It was thought that superimposition of such structures allowed growth and/or treatment changes in other skeletal structures to be demonstrated. Anterior cranial base superimposition proposed by De Coster 28 was based on that principle.
Other superimposition methods have been proposed. Simpler methods based on two or three easily identified cephalometric landmarks have been used. Thus, the sella-nasion line, with sella as the registering landmark, was proposed as an ideal superimposition method by Steiner 2 and Tweed 3 . However, this method implies that sella is stable and that an increase in the SN line would be due to positional changes of the nasion. Ricketts 6 developed a superimposition method based on the posterior region of the cranial base and the nasion-basion line, reasoning that growth of the spheno-occipital synchondrosis would play an important role until the end of puberty. Melsen 29 found that sella is not stable during growth due to remodeling of the fossa. In addition, the author found that nasion and basion changed considerably in position, direction and the amount of growth, which made them unreliable structures on which cephalometric superimposition could be based.
Based on their implant studies, Bjork and Skieller 4,5 identified the location of natural reference markers on the anterior cranial base, mandible and maxilla. This method is known as the structural method. In this study, superimposition on the anterior cranial base was used.