Pectus deformities: tomographic analysis and clinical correlation
To assess, with computed tomography (CT) studies, features of anterior chest wall development that can be related to different types of pectus deformities.
Materials and methods
From 71 patients with pectus deformities and chest coronal CT scans, 48 (40 male and 8 female), with a mean age of 15.8 years (ranging from 5 years to 38.4 years) were selected and divided into five groups, according to clinical type of deformity and image quality. A similar CT study was performed in a sixth group of 14 individuals with no underlying pectus deformity (control group), six male and eight female, mean age 19.3 years, (range 10.8 years to 30.5 years), totaling 62 subjects. Tomographic studies were performed on a 64-section CT scanner, with parameters varied according to the subject’s body mass index (BMI). Coronal reconstructions were used to assess six features of the sternum and costal cartilages in the groups. Two other factors, a sternal index, created to estimate the sternal body width, and the sternocostal angles, were also studied.
Feature I was noted in 13 patients and in no controls (P = 0.002), feature II in 39 patients and in one control (P = 0.000), feature III in 37 patients and in two controls (P = 0.002), and feature IV in two patients and in no controls (P = 0.002). The sternal index was significant to one group of patients.
The features studied and the index provide measurable and applicable data for the interpretation of anterior chest wall tomography, with possible implications for prognosis and treatment of different types of pectus deformities.
KeywordsChest Deformities Imaging studies of pectus deformities Pectus carinatum Pectus excavatum Studies on the development of the anterior chest wall by computed tomography
Analyses of imaging studies of sternal deformities from a development and growth standpoint are not usual, and radiologic reports of ‘normal’ chests are common, despite clear irregularities in sternal ossification and maturation . Sternal sutures are reported in the literature [2, 3, 4, 5, 6], but they do not exist in the sternum. Anterior chest wall formation and growth is endochondral [7, 8], and there are cartilaginous growth plates between the segments of the growing sternum [1, 9, 10, 11, 12] and at the costochondral junctions .
According to many authors [2, 3, 12, 14, 15, 16], the main development stages of the anterior chest wall can be summarized as follows. Sternum formation begins from two mesodermal longitudinal bands, visible at 6 weeks of embryonic life in the axillary regions. The costal arches elongate ventrally from the spine and, as the embryo grows, the tips of the ribs migrate forward with the sternal bands and fuse in the mid-ventral line to form a single structure. The union of the two sternal bands occurs in a cranial–caudal direction, simultaneously with a chondrification process. At 9 weeks, the sternum is uniformly cartilaginous and resembles, in shape, the future bone. When the chondrification process has been completed, rarely will any sign of its bilateral origin be observed. The ossification of the sternum is normally initiated at the manubrium and the three superior segments of the sternal body during the last 3–4 months of fetal life, whereas the ossification of the fourth and lower segment of the sternal body usually takes place during the first year after birth. Bony union among the four segments of the sternal body usually occurs from below upward, starting in early childhood; it is said to be completed between the 16th and the 25th years of life. The xyphoid process may ossify after birth, or it may remain cartilaginous throughout life. A fusion between the xyphoid process and the sternal body is found in approximately 30% of individuals after the second decade of life, and the manubrium–sternal synchondrosis usually remains open throughout life.
Axial computed tomography (CT) has been used for the three-dimensional (3-D) evaluation of thoracic dimensions , without any mention of development of sternum and costal cartilages. Some features of the anterior chest wall development have not yet been completely studied. The purpose of this study was to analyze with coronal CT the features that can be related to different types of pectus deformities.
Materials and methods
Nine hundred and twenty patients with pectus deformities had sought medical assistance over a 5-year period. Patients with important deformities had the indication for an imaging study of their condition. Seventy-one of them, or their legal guardians, agreed to proceed with a chest CT scan. A consent form, previously approved by the Institutional Review Board of our major institution, was signed by each individual or legal guardian.
Of the 71 patients, 23 were excluded from the study: 19 due to inadequate reconstructed images, three due to previous surgery with costal cartilage resection, and one with iatrogenic pectus  caused by previous sternal puncture during the neonatal period. All other patients had an idiopathic pectus deformity [12, 19]. The remaining 48 patients were divided into five groups, according to their different types of pectus deformities, as follows. Group 1, 11 patients with PCI; group 2, 11 patients with PCL; group 3, six patients with PCS; group 4, six patients with PEW and group 5, 14 patients with PEL. A sixth group, the control group, consisted of 14 patients with no underlying skeletal deformities who had undergone CT for unrelated reasons and who had also been informed and had consented to take part in this study. The 48 patients with pectus deformities and the 14 individuals from the control group resulted in a total of 62 subjects to be studied.
The 62 subjects consisted of 47 children and adolescents, 5–19 years old, and 15 adults, 20–38 years of age. In the groups of patients with pectus deformities the mean age was 15.8 years (range 5–38.4 years). In the control group the mean age was 19.3 years (range 10.8–30.5 years). Of those 48 patients with pectus deformities (groups 1to 5), 40 were male (83%) and eight were female (17%). In the control group (14 individuals), six were male (43%) and eight were female (57%).
All patients were assessed by multislice, non-enhanced, thorax CT, with axial sections and sagittal and coronal reconstruction of the anterior chest wall, performed on a Philips Brilliance® 64-section scanner, using the following parameters as a baseline (the examinations were tailored according to the body mass index (BMI) of the individual): section thickness 0.9 mm; matrix 512 pixels × 512 pixels; pitch 0.8; rotation time 0.5 s; tube voltage 80–120 kV. The maximum tube current–time product was limited to 200 mAs for the pectus group [mean CT dose index (CTDI) 10.7 mGy and mean dose–length product (DLP) 381.9 mGy.cm] and 300 mAs for the control group (mean CTDI 13.85 mGy and mean DLP 549.16 mGy.cm); with xy-plane dose online modulation (D-DOM [automatic current selection (ACS)]. All examinations were performed with the patient supine and the arms extended over the head. The CT examinations were scored by one of the authors, a radiologist, who reviewed the images with various window width and level settings. Only the coronal reconstruction was selected for study of the morphology and the development of the sternum and costal cartilages.
Features present in the anterior chest wall development and their schematic illustrations
We used the SPSS version 15.0 software package. To compare the tomographic features of the sternum and costal cartilages between all groups, we employed Pearson’s chi-square test (cross-tabulation). We also used an independent samples t-test to compare the mean ages of those who either had or had not had different tomographic features detected in their studies. We used one-way analysis of variance (ANOVA) for independent groups to analyze differences between means of the SC angles and WL index between the groups. A P value smaller than or equal to 0.05 was considered to be statistically significant. To calculate the sensitivity and the specificity of CT in detecting such deformities, we used only features that presented themselves as statistically significant in relation to pectus deformities.
Feature I was found in two out of 11 patients from group 1 (PCI) (ages 18 years and 20 years); in three out of 11 patients from group 2 (PCL) (ages 14 years, 28 years, and 38 years); in five out of six patients from group 3 (PCS) (from 14 years to 36 years of age); in one out of six patients from group 4 (PEW) (18 years old); in two out of 14 patients from group 5 (PEL) (14 years and 23 years of age); and in none of the 14 individuals from the control group.
Feature II was noticed in nine out of 11 patients from group 1 (PCI) (aged 12–20 years); in ten out of 11 patients from group 2 (PCL) (5–38 years of age); in five out of six patients from group 3 (PCS) (5–36 years of age); in four out of six patients from group 4 (PEW) (5–20 years of age); in 11 out of 14 patients from group 5 (PEL) (7–37 years of age); and in one out of 14 subjects from the control group (28 years old).
Feature III was found in ten out of 11 cases of PCI (patients aged 10–20 years); in ten out of 11 cases of PCL (patients between 5 and 38 years of age); in five out of six cases of PCS (5–36 years of age); in three out of six cases of PEW (between 5 and 20 years of age); and in nine out of 14 cases of PEL (between 7 and 37 years of age), and it was present in two out of 14 individuals from the control group (15 years and 30 years old).
Feature IV was found only in two patients, a 13.8-year-old adolescent and a 30-year-old adult, both with PCS, and was not observed in any other type of pectus or in the group control.
Features V and VI were noted in all groups, and there was no statistical significance between groups regarding the presence or absence of these features.
Prevalence of CT features of the sternum and costal cartilages in each group and their statistical significance (chi-square tests/cross-tabulations). See Table 1 for a description of the features
CT feature I
CT feature II
CT feature III
CT feature IV
CT feature V
CT feature VI
(n = 11)
(n = 11)
(n = 6)
(n = 6)
(n = 14)
(n = 14)
Of 14 patients from the control group, 11 had none of the statistically significant CT features. When at least one of the aforementioned statistically significant features was found, CT was able to detect pectus deformities with 96% sensitivity and 79% specificity. When two or more features were found, the specificity increased to 100%.
By comparing the mean age of individuals in which the features were detected and the mean age of individuals in which the features were not detected, in all groups, we found significant values only for features I (P = 0.002), V (P = 0.002), and VI (P < 0.001). Feature I was detected at the mean age of 22.0 years, and it was not detected at a mean age of 14.86 years. Feature V was detected at a mean age of 12.67 years, and it was not detected at a mean age of 18.69 years. Feature VI was detected at a mean age of 21.99 years, and it was not detected at a mean age of 12.29 years.
Mean values of the WL index of the sternal body. The greater the index, the larger the sternal body
Mean WL index
Group 1 (PCI)
Group 2 (PCL)
Group 3 (PCS)
Group 4 (PEW)
Group 5 (PEL)
Group 6 (control)
Comparison of the sternocostal (SC) angles between the groups through one-way ANOVA showed no statistically significant relationship (P > 0.21). For equivalent right and left costal cartilages, the mean values of the groups were: 2nd SC angle in groups 1–5 (pectus) 88.17°, in group 6 (control) 92.86°; 3rd SC angle in groups 1–5 (pectus) 73.03°, in group 6 (control) 74.96°; 4th SC angle in groups 1–5 (pectus) 57.31°, in group 6 (control) 57.64°; 5th SC angle in groups 1–5 (pectus) 45.49°, in group 6 (control) 43.96°; 6th SC angle in groups 1–5 (pectus) 35.84°, in group 6 (control) 36.14°; 7th SC angle in groups 1–5 (pectus) 27.68°, in group 6 (control) 24.5°. Only one patient with PCS, aged 14 years and shown in Fig. 4a, b, demonstrated particularly small values for the 5th (28°), 6th (17°) and 7th (10°) SC angles.
According to Haje and Haje, pectus deformities are more common in male subjects (73%) than in female subjects (27%) . That explains the higher male prevalence in the pectus groups in our study. The chest wall poses diagnostic difficulties for both the clinician and the radiologist . Psychological and psychosomatic aspects of pectus deformities in children, adolescents, and young adults are solved when correction is brought about by treatment . To achieve accurate and timely diagnoses that facilitate the appropriate treatment of pectus deformities, physicians must be familiar with chest wall growth and development disturbances .
The interpretation of our results from studied CT features is discussed on the basis of a possible correlation with pectus deformities. An influence on prognosis and eventually on the treatment is noted when necessary.
Features I, II and III had a statistically significant prevalence in all pectus groups when compared with the control group, and the presence of one of those features could suggest the existence of a pectus deformity, although it may not specify which type of deformity is present. Feature I was seen in all types of pectus deformities, but not in the control group. Feature II was noted in just one subject of the control group, but showed a high prevalence in groups with deformities. Feature III had a high prevalence in all pectus groups and was seen in just two subjects of the control group. It could be interpreted as resulting from a primary irregularity in the growth and development of the sternum, with consequent mismatch between the sternal growth and the costal growth. The two individuals noted in the ‘normal’ group had only a slight costal asymmetry, which was not severe enough to result in deformity.
The prevalence of feature IV in only two patients of one group (PCS) and in none of the subjects from the other groups was statistically significant and suggests that this is a feature unique to PCS. Features I and VI were also present in these two cases, indicating the closure of all cartilaginous growth plates of the sternum. Early fusion of sternal cartilages explains the usual rigidity in PCS, and this is an important observation that indicates orthotic treatment in childhood for this type of deformity [9, 12, 19]. Regarding other types of pectus deformities, such a treatment is ideally started before adolescence (PEW, PEL) or during adolescence (PCI, PCL), depending on the clinical judgment of the flexibility of the deformity and on the remaining cartilage in the sternum . Thus, an understanding of the studied features and how they may be related to the development of the anterior chest wall has important implications on the prognosis and, consequently, on the timing of the conservative treatment of a pectus deformity. The remaining cartilaginous structures—or growth cartilaginous plates—in the sternum also may have implications to avoid the iatrogenic deformity reported by Haje  and to avoid the complications and poor long-term results of operative techniques mentioned by Haje and Bowen .
The separate analysis of the first four features, those that showed a significant correlation between the groups, demonstrated that multislice coronal CT reconstruction of the anterior chest wall is capable of detecting pectus deformities with high sensitivity and specificity. We understand that the clinical examination of a pectus deformity is the most important aspect of the diagnosis, but complementary CT scans could convey a better understanding of developmental abnormalities that may lead to each specific type of pectus reported in this study. A clear rationale for prognosis and, consequently, treatment, could emerge.
Features V and VI apparently did not show any correlation with pectus deformities. However, such features demonstrated significant correlation with the age of the subject. In the pectus groups, feature V was more frequently seen in children and adolescents, and, in the control group, such feature was seen only in two subjects: clearly in an 11-year-old girl, and traces in a 25-year-old man. Regardless of the group, feature VI was more frequently seen in older adolescents and adults. This finding agrees with reports on normal sternal development, where the process of fusion in the sternum is said to be complete “between the 16th and the 25th year of life” . Our findings suggest that a normal process of complete sternal ossification tends to occur towards the end of adolescence, but it may eventually take place in early adulthood. The most commonly seen feature in the ‘normal’ group was VI, which was observed in five adults and three older adolescents: a 17-year-old boy, a 19-year-old woman, and a 14-year-old girl. At those ages, and considering the subjects’ gender, such a feature (the fusion of all sternal body segments) is usually expected to be found. That feature was found important to be evaluated in this study, due to previous reports of premature bony fusion of the sternal body in children and adolescents with a given type of pectus deformity [1, 3].
In a study of skeletally immature cadavers, Ogden et al. showed different patterns of ossification of the sternum, but they did not report whether those children and adolescents had pectus deformities or not . We understand that different growth patterns of the sternum and costal cartilages can exist in healthy subjects, and that the clinical correlation is always important. Slight growth irregularities may not lead to a deformity, just to a variation from normal.
The factors WL index and SC angles were interpreted according to our findings and to previous studies.
Haje et al., in a study with plain radiographs, described the BM index, resulting from the length of the sternal body (B) divided by the length of ossified manubrium (M); lower values depict short sternal bodies, and the shortest were found in PCS deformities . The WL index in our study, a ratio resultant from the larger width (W) of the ossified body of the sternum divided by its length (L), implies higher values for wide sternal bodies. The WL index showed higher values in patients with PCS. Our findings and those of the aforementioned study demonstrate that PCS really results from a disproportion between continued growth of the ribs and impaired growth of the sternum, due to premature closure of its cartilaginous growth plates, with consequent sternal enlargement and shortening. In the aforementioned study, the authors concluded that disturbances in the growth of the sternum occur in the three basic types of pectus carinatum (PCI, PCL, PCS), and in the localized type of pectus excavatum (PEL), but not in the wide type of excavatum (PEW) . Ours demonstrates that features correlated to abnormal sternal and costal development can be present in all types of pectus deformities, including PEW.
The analysis of means for the SC angles showed similar values between the groups and within the same group, not allowing any differentiation between clinically ‘normal’ individuals and pectus patients. No relationship to age was found with regard to this factor. However, the PCS patient shown in Fig. 4a, b presented the lowest 5th, 6th, and 7th angle values of all individuals. This finding implies that the inferior costal cartilages were at a more vertical position than usual, and contributes to the explanation that PCS deformity is a result of growth arrest from the sternum, as reported by Haje et al. . Future studies may find that such angles may be useful for imaging interpretation of the chest wall.
We chose to employ only coronal reconstructions of the anterior chest wall in this study, having in mind that it would provide an improved and more comprehensive understanding of the sternum and costal cartilage development. However, we imagine that future studies using the sagittal plane, besides the possibility of showing anatomical rotation of the sternal axis and the distance from the sternum to the spine, could eventually yield additional information on the development of pectus deformities and justify ongoing investigation.
Owing to radiation concerns, we do not recommend chest CT for pectus study as a routine, and we emphasize that we do not advocate the use of CT for diagnostic purposes in the assessment of these deformities. We believe that all the features studied, except feature III, will eventually be able to be seen in oblique and lateral standard radiographs of the sternum, although image quality is certainly superior in CT scanning. Therefore, CT could be indicated, according to clinical judgment on an individual basis, mainly for prognosis purposes. Future equipment and studies may result in better CT protocols, especially designed to deliver the smallest possible radiation dose.
The features studied and the WL index provide measurable and applicable data for the interpretation of anterior chest wall tomography, with possible implications for the prognosis and treatment of different types of pectus deformities.
- 3.Currarino G, Silverman N. Premature obliteration of the sternum sutures and pigeon breast deformity. Radiology. 1958;70:352.Google Scholar
- 4.Gabrielsen TO, Ladyman GH. Early closure of sternal sutures and congenital heart disease. AJR Am J Roentgenol. 1963;89:375–83.Google Scholar
- 5.Haje SA, Raymundo JLP. Considerações sobre deformidades da parede torácica anterior e apresentação de tratamento conservador para as formas com componentes de protrusão. Rev Bras Orthop. 1979;14:167–78.Google Scholar
- 6.Kuhn JP. The Thorax. In: Silverman FN, editor. Caffey's pediatric X-ray diagnosis. Chicago: Year Book Medical Publishers Inc; 1985. pp. 1097–9.Google Scholar
- 11.Haje SA, Bowen JR, Harcke HT, Guttenberg ME, Bacon CR. Disorders in the sternum growth and pectus deformities: an experimental model and clinical correlation. Acta Orthop Bras. 1998;6:67–75.Google Scholar
- 12.Haje S, Haje DP. Tórax e cintura escapular. In: Hebert SK, Barros Filho TEP, Xavier R, Pardini Jr AG, editors. Ortopedia e traumatologia: princípios e prática. 4th ed. Porto Alegre: Artmed Editora; 2009. pp. 147–65.Google Scholar
- 13.Shimomura Y, Wezeman FH, Ray RD. The growth cartilage plate of the rat rib: cellular differentiation. Clin Orthop. 1973;(90):246–54.Google Scholar
- 14.Paterson AM. The Human sternum. Liverpool: Williams & Norgate, 1904.Google Scholar
- 16.Gray H. Gray's anatomy. Carmine D. Clemente, editor. 30th American Ed. Philadelphia: Lea & Febiger; 1985.Google Scholar
- 19.Haje SA, Haje DP. Abordagem ortopédica das deformidades pectus: 32 anos de estudos [Orthopaedic approach to pectus deformities: 32 years of studies]. Rev Bras Orthop. 2009;44:193–200.Google Scholar
- 21.Einsiedel E, Clausner A. Funnel chest. Psychological and psychosomatic aspects in children, youngsters and young adults. J Cardiovasc Surg. 1999;40:733–6.Google Scholar