The aim of this study was to analyze the diagnostic performance of the cervical vertebral maturation (CVM) method in estimating accurately the stages of maturation of the midpalatal suture.
Cone-beam computed tomography (CBCT) images from 142 subjects (84 female, 58 male; mean age, 14.8 ± 9.7 years) were analyzed by 2 calibrated examiners to define, by visual analysis, the maturational stages of the cervical vertebrae and the midpalatal suture. These CBCT images were required by orthodontists and surgeons for diagnosis and treatment purposes. Positive likelihood ratios (LHRs) were calculated to evaluate the diagnostic performance of the CVM stages in identifying the maturational stages of the midpalatal suture.
Positive LHRs greater than 10 were found for several cervical vertebral stages (CSs), including CS1 and CS2 for the identification of midpalatal suture stages A and B, CS3 for the diagnosis of midpalatal suture stage C, and CS5 for the assessment of midpalatal suture stages D and E. These positive LHRs indicated large and often conclusive increases in the likelihood that the CVM stages were associated with specific stages of midpalatal suture maturation. At CS4, there were a moderate positive LHR for stage C and low positive LHRs for stages D and E.
Most CVM stages can be used for the diagnosis of the stages of maturation of the midpalatal suture, so that CBCT imaging may not be necessary in these patients. In the postpubertal period, however, an assessment of the midpalatal suture maturation using CBCT images may be indicated in deciding between conventional rapid maxillary expansion and surgically assisted rapid maxillary expansion. On the other hand, if the CVM stage cannot be assessed, chronologic age may be a viable alternative to predict some midpalatal suture stages (particularly the early stages).
The diagnostic performance of the cervical vertebral maturation (CVM) method for midpalatal suture maturation was evaluated.
Positive correlations among CVM stages and maturation stages of the midpalatal suture were verified.
CS1 and CS2 identify the midpalatal suture stages A and B.
CS3 indicates midpalatal suture stage C.
For CS4 and CS5, individual assessment of midpalatal suture maturation is recommended.
Rapid maxillary expansion (RME) is an orthopedic procedure that produces separation of the midpalatal suture, thus widening the maxilla. RME has been used routinely in orthodontic practice for many reasons, including the correction of crossbites and dental crowding.
Histologic studies and investigations with microcomputed tomography on autopsy material have demonstrated large variability in the chronologic age for fusion of the midpalatal suture. Persson and Thilander observed fusion of the midpalatal suture in subjects from 15 to 19 years of age. On the other hand, patients at the ages of 27, 32, 54, and even 71 years have been reported to have no signs of fusion of this suture. Such findings indicate that the variability in the developmental stages of fusion of the midpalatal suture is not related directly to chronologic age, particularly in young adults.
The maturation of the facial sutures has been demonstrated by Björk to be related to growth in height, and the start of fusion of the midpalatal suture has been associated with the rate of skeletal growth. It is well known that skeletal growth has periods of acceleration and maturation that are not associated directly with chronologic age. Therefore, several biologic indicators have been proposed for individual assessment of skeletal maturity, including the hand-wrist method and the cervical vertebral maturation (CVM) method.
The CVM method is performed on lateral cephalograms that are used routinely for orthodontic diagnosis and treatment planning, avoiding the need for an additional radiograph. This method has demonstrated reliability and reproducibility for evaluating the pubertal peak and further maturation in skeletal growth.
Cone-beam computed tomography (CBCT) provides 3-dimensional images of the oral and maxillofacial structures with no image overlap, allowing a reliable diagnosis of the maturation of the midpalatal suture before RME. Angelieri et al proposed 5 maturational stages of the midpalatal suture: stage A, straight high-density sutural line, with no or little interdigitation; stage B, scalloped appearance of the high-density sutural line; stage C, 2 parallel, scalloped, high-density lines that are close to each other and are separated in some areas by small low-density spaces; stage D, fusion completed in the palatine bone with no evidence of a suture; and stage E, complete anterior fusion in the maxilla.
However, CBCT is characterized by additional radiation and increased costs for patients compared with the lateral cephalograms and panoramic radiographs obtained routinely in orthodontic practices. Therefore, the aim of our study was to analyze the diagnostic performance of the CVM method in identifying correctly the stages of maturation of the midpalatal suture in growing and adult patients.
Material and methods
Baseline diagnostic CBCT images acquired for clinical purposes in 142 subjects (84 female, 58 male) with a mean age of 14.8 ± 9.7 years (range, 5.3-58.4 years) were examined. The CBCT images were required by clinicians for diagnosis and treatment of these patients. These CBCT images were obtained from the archives of private practices of orthodontists and surgeons as well as from researches conducted at Methodist University of São Paulo, São Bernardo do Campo, Brazil, and the University of Michigan, Ann Arbor. This study was approved by the institutional review board of the University of Michigan.
The images of the midpalatal suture were analyzed using Invivo5 software (Anatomage, San Jose, Calif). The adjustment of the patient’s head in the 3 planes of space and the selection of the slice for evaluation the midpalatal suture maturation were performed according to the protocol described previously by Angelieri et al. The central cross-sectional axial slice in the superior-inferior dimension (from the nasal to the oral surfaces) of the palate was used for the staging of the midpalatal suture.
All axial central cross-sectional slices used for assessment of the midpalatal suture were arranged by the principal investigator (F.A.) in a PowerPoint (Microsoft, Redmond, Wash) presentation with a black background and codes that were displayed sequentially on a high-definition computer monitor. For subjects with a thick palate, 2 axial cross-sectional slices were used. No adjustments in contrast or brightness of these images were made. In a darkened room, each image of the midpalatal suture was classified blindly by an expert examiner (F.A.) according to the visual analysis method described previously by Angelieri et al.
Using the Invivo5 software, lateral cephalograms were derived from the same CBCT images. The cephalograms then were cropped so that only the images of the cervical vertebrae were visible; images of the dentition and adjacent skeletal structures were not visible. The cropped images were arranged in a similar PowerPoint presentation with a black background and identification codes. The vertebral images were analyzed according to the CVM method by 1 expert examiner (J.A.M.).
Thirty images of the cervical vertebrae and 30 images of the midpalatal sutures from the same subjects were selected randomly from the total sample and reclassified by the same examiners a month later. A weighted kappa coefficient was calculated for evaluation of the intraexaminer agreement for the CVM method and classification of the midpalatal suture maturation.
The correlations between the midpalatal sutures and the CVM stages were evaluated by the Spearman correlation test.
The relationship between skeletal maturity assessed with the CVM method and the maturational stages of the midpalatal suture was evaluated with a measure of diagnostic performance—the positive likelihood ratio (LHR). The positive predictive value of a test is the probability that the patient has the condition (in this case, a specific maturational stage of the midpalatal suture) when restricted to patients who test positive (specific stage of CVM). The LHR incorporates both the sensitivity and the specificity of the test and provides a direct estimate of how much a test result will change the odds of having a condition or “disease.”
The LHR for a positive result indicates how much the probability of the condition to be diagnosed (specific maturational stage of the midpalatal suture) increases when a test is positive (specific CVM stage). A result of 1 indicates no diagnostic performance (ie, no relationship), whereas a result smaller than 1 must be interpreted as a decrease in the likelihood of disease (negative relationship). An LHR greater than 1 indicates that the test result is associated with the disease.
An LHR between 1 and 2 can be interpreted as a minimal increase (15%) in the likelihood of disease. An LHR between 2 and 5 indicates a small increase (15%-30%) in the likelihood of disease. An LHR between 5 and 10 can be interpreted as a moderate increase (30%-45%) in the likelihood of disease. LHRs above 10 indicate large and often conclusive increases in the likelihood of the disease (ie, strong association). A positive LHR of 10 or more for any CVM stage was considered a reliable indicator for the diagnosis of any of the maturational stages of the midpalatal suture.
The statistical analyses were performed with a statistical software package (version 12; SPSS, Chicago, Ill) and an interactive statistical calculator ( statpages.org/ctab2x2.html ). Statistical significance for all statistical tests was set at P <0.05.
The weighted kappa coefficients for the evaluation of the intraexaminer agreement for the CVM method and the classification of the midpalatal suture maturation were 0.978 and 0.935, respectively. This result indicates very good intraexaminer reproducibility.
The prevalence rates of the different CVM stages according to the maturational stages of the midpalatal suture are reported in Table I , as are the demographics of the subjects at the different CVM stages. There were statistically significant positive correlations between the CVM stages and the maturational stages of the midpalatal suture (Spearman r = 0.908 at P <0.01).
|Skeletal maturation stages||Midpalatal suture stages||Correlation coefficient|
|Stage A||Stage B||Stage C||Stage D||Stage E||0.908*|
|CS1 (n = 41; 22 F, 19 M; mean age, 8.5 ± 2.7 y)||15 (36.6%)||26 (63.4%)||–||–||–|
|CS2 (n = 16; 7 F, 9 M; mean age, 10.4 ± 2.1 y)||1 (6.2%)||15 (93.8%)||–||–||–|
|CS3 (n = 23; 10 F, 13 M; mean age, 12.4 ± 1.3 y)||–||4 (17.4%)||19 (82.6%)||–||–|
|CS4 (n = 25; 18 F, 7 M; mean age,14.2 ± 1.7 y)||–||–||18 (72.0%)||3 (12.0%)||4 (16.0%)|
|CS5 (n = 37; 27 F, 10 M; mean age, 25.6 ± 13.4 y)||–||–||5 (13.5 %)||13 (35.1%)||19 (51.4%)|
The positive LHRs of the cervical vertebral stages (CSs) for the identification of the maturational stages of the midpalatal suture are given in Table II . The presence of CS2, CS3, CS4, and CS5 produces a minimal to small decrease in the likelihood of stages A, B, E, and C, respectively. The values of positive LHRs of CS1 for the diagnosis of stages A and B in the midpalatal suture and of CS5 for the identification of stage D were between 2 and 5, indicating a small increase in the likelihood (15%-30%) of detecting the maturational stages of the midpalatal suture. Both CS4 and CS5 showed a moderate increase (30%-35%) in the likelihood of detecting the maturational stages C and E, respectively. Only CS2 and CS3 showed positive LHRs greater than 10 for the diagnosis of stages B and C, respectively ( Table II ).
|Skeletal maturation stages||Midpalatal suture stages|
|Stage A||Stage B||Stage C||Stage D||Stage E|
In a previous study, the midpalatal maturational stages A and B were indicated as the stages in which a conventional RME approach would have encountered less resistant forces and probably more skeletal effects than at stage C. With either stage D or E, surgically assisted RME should be considered because fusion of the midpalatal suture already has occurred partially or totally.
To give more clinical relevance to the results of this investigation, data from the CVM stages and the midpalatal suture stages were combined ( Table III ). Five diagnostic performance parameters were calculated for prepubertal stages CS1 and CS2 for the identification of midpalatal suture stages A and B, pubertal stage CS3 for the diagnosis of midpalatal suture stage C, and postpubertal stage CS5 for the assessment of midpalatal suture stages D and E. In all cases, the positive LHRs were greater than 10, thus indicating a large and often conclusive increase in the likelihood that the CVM stage of diagnosis corresponds to a specific stage in midpalatal suture maturation.
|Diagnostic tests||Skeletal maturation stages|
|CS1 and CS2 Variable diagnosed: stages A and B||CS3 Variable diagnosed: stage C||CS5 Variable diagnosed: stages D and E|
|Value (95% CI)||Value (95% CI)||Value (95% CI)|
|Sensitivity (%)||93.4 (87.9-93.4)||45.2 (34.4-51.5)||82.1 (70.5-89.2)|
|Specificity (%)||100.0 (95.8-100.0)||96.0 (91.4-98.6)||95.1 (90.8-97.9)|
|Positive predictive value (%)||100.0 (94.1-100.0)||82.6 (62.9-93.0)||86.5 (74.3-94.0)|
|Positive LHR||infinity (21.1-infinity)||11.3 (4.0-37.8)||16.9 (7.6-41.7)|
According to the 4 parameters of sensitivity, specificity, positive predictive value, and positive LHR, the prepubertal stages CS1 and CS2 showed the highest diagnostic performance for identification of stages A and B in midpalatal suture maturation, with specificity and positive predictive values of 100%. However, the 95% confidence intervals for the positive LHRs of these CVM stages for the identification of maturational stages in midpalatal suture were variable, with the widest range seen for CVM stages 1 and 2 (21.1-infinity).