Pectus excavatum: the effect of tricuspid valve compression on cardiac function

Background Pectus excavatum (PE) is a common congenital chest wall deformity with various associated health concerns, including psychosocial impacts, academic challenges, and potential cardiopulmonary effects. Objective This study aimed to investigate the cardiac consequences of right atrioventricular groove compression in PE using cardiac magnetic resonance imaging. Materials and methods A retrospective analysis was conducted on 661 patients with PE referred for evaluation. Patients were categorized into three groups based on the degree of right atrioventricular groove compression (no compression (NC), partial compression (PC), and complete compression(CC)). Chest wall indices were measured: pectus index (PI), depression index (DI), correction index (CI), and sternal torsion. Results The study revealed significant differences in chest wall indices between the groups: PE, NC=4.15 ± 0.94, PC=4.93 ± 1.24, and CC=7.2 ± 4.01 (P<0.0001). Left ventricle ejection fraction (LVEF) showed no significant differences: LVEF, NC=58.72% ± 3.94, PC=58.49% ± 4.02, and CC=57.95% ± 3.92 (P=0.0984). Right ventricular ejection fraction (RVEF) demonstrated significant differences: RVEF, NC=55.2% ± 5.3, PC=53.8% ± 4.4, and CC=53.1% ± 4.8 (P≥0.0001). Notably, the tricuspid valve (TV) measurement on the four-chamber view decreased in patients with greater compression: NC=29.52 ± 4.6; PC=28.26 ± 4.8; and CC=24.74 ± 5.73 (P<0.0001). Conclusion This study provides valuable insights into the cardiac consequences of right atrioventricular groove compression in PE and lends further evidence of mild cardiac changes due to PE. Graphical abstract Supplementary Information The online version contains supplementary material available at 10.1007/s00247-024-05971-z.


Introduction
Pectus excavatum (PE) is common with an incidence of 1 in 300 to 400, and a male to female ratio of 2:1 [1].The ensuing chest wall deformity can impact patients' psychosocial, academic [2], and cardiopulmonary function and be associated with chest pain, palpitations, and decreased exercise tolerance [3][4][5][6].The cardiopulmonary effects of PE can range from very mild to severe and appear to be related to the severity of the chest wall deformity; however, the cardiopulmonary effect does not have a tight relationship to the severity of deformity [3].One could state that the pathophysiologic effects of PE are controversial, but most surgeons treating this condition would state that the pathophysiologic effects of PE are real [6][7][8].Given that the changes affecting the heart function usually do not have a clinical manifestation, the controversy continues.Medical evidence is mixed on the impact PE has on cardiopulmonary function, but many of the cardiopulmonary exercise tests, pulmonary function tests, and cardiac functional evaluations are still within normal limits [3,4,9].As a population, patients with PE have cardiac systolic function on the low side of average with the RVEF and LVEF bell curves of the PE group shifted leftward from the normal bell curve [3,6].
Cardiopulmonary signs and symptoms were recognized in case reports and small series of case in the early, pioneering days of pectus repair and is nicely summarized in a paper by Wachtel et al. [6].In their report of 13 subjects, they reviewed the literature from 1939 to 1956.The degree of PE was severe in most cases and symptoms were extensive.Yet, in most cases, the subject had substantial or complete resolution of symptoms post-surgery.Of course, these early studies selected for only the severe subjects.In 1974, Guller and Hable looked at 54 subjects with moderate or severe PE and state that on physical examination had finding suggestive of cardiovascular disease: murmurs on auscultation, electrocardiogram changes, vector electrocardiogram changes, and chest X-ray changes [7].Nearly all were asymptomatic, 52 out of 54, which led them to conclude these findings were due to displacement of the heart and had minimal clinical significance [7].In the modern era of cardiac MRI, Saleh et al. compare pectus excavatum with match normal controls using CMR and found that the RVEF was reduce by 6% in the pectus group compared to the controls but did not find a change in pulmonary circulation time or perfusion indices [10].Another group evaluated the cardiac systolic function pre-and postoperatively showing that RVEF was reduced preoperatively as a group and postoperatively either remained stable or improved.RVEF less than 45% predicted improvement [11].Using CMR and stress echocardiography, PE subjects were found to have lower peak exercise capacity, changes in both the left and right ventricular E/A ratio, increased tricuspid gradient, and abnormal septal motion with flattening during inspiration [12].This brief review shows that PE affects cardiac systolic function, particularly affecting the right side of the heart.
The evolution of PE repair has evolved from an uncommonly performed, truly invasive surgery to a minimally invasive surgery which is performed in centers throughout the world.In the early days, resection of the anterior chest wall was practiced from 1910 to 1920.The Ravitch procedure was developed in 1949 which left many children with constrictive chest wall physiology leading to a "modified Ravitch procedure" using Adkins struts that dominated from the mid-1950s until the 1980s.Nuss introduced the first "Nuss bar" repair in 1987 [13,14].Since the first "Nuss bar" repair, the technique and equipment have been refined, creating the minimally invasive repair of PE that is used today [14].Whereas earlier methods of repair were relatively risky and more painful, after the advent of the minimally invasive repair of pectus excavatum, repair of PE has become increasingly common.
Around the same time that Nuss was developing his novel technique for repair, J. Alex Haller et al. published an article that employed CT to evaluate chest wall shape of patients with PE and introduced a CT-based pectus index (PI) [15].A PI of >3.25 defined a cutoff value for a group of patients that underwent repair in their institution [15].This was adopted as the threshold for performance of PE repair.Subsequently, there has been a proliferation of cross-sectional imaging indices from CT and MRI.
The chest wall deformity of PE causes variable compression of the right ventricle which is associated with decreased right ventricle function.Tandon et al. found that compression of the right ventricle free wall resulted in significantly lower right ventricle ejection fraction when compared to those without compression, and further decrease with compression at the right atrioventricular groove [16].The hypothesis of this study is that compression of the atrioventricular groove results in a progressive decrease in right ventricular systolic function.This study aims to fully assess the effect that right atrioventricular groove compression has on the cardiac systolic function as measured by cardiac MRI and relate this to standard measures of pectus excavatum.

Patients and methods
Approval was obtained from the institutional review board to retrospectively examine the electronic medical record of patients with PE referred to our chest wall center for evaluation.Demographic data included age, gender, race, weight, and height.Inclusion criteria were referral to the chest wall center and cardiac MRI.Exclusion criteria included congenital heart disease, prior chest wall surgery, underlying pulmonary disease, and connective tissue disorders.If a patient had repeat studies, only the most recent presurgical study was included.All patients seen by our chest wall center undergo cardiac MRI unless there is a contraindication.
Cardiac magnetic resonance imaging was performed on an Ingenia 1.5-Tesla scanner (Philips Healthcare, Best, Netherlands) to assess pectus indices, cardiac compression, and cardiac function.The protocol included axial singleshot steady-state free precession (SSFP) with cardiac end systolic gating and respiratory gating, standard cine SSFP with retrospective electrocardiogram gating obtained during suspended expiratory breath hold.These were obtained in the vertical long axis, horizontal long axis (four-chamber view), short-axis, and left ventricular outflow tract planes.Our standard protocol also included cine steady-state free precession in a trans axial plane, free breathing cine phase contrast of the aortic valve, main pulmonary artery, and the atrioventricular valves on one phase contrast cine slice.At the end, a 3D SSFP, cardiac, and respiratory-gated wholeheart sequence was performed.No contrast was administered for the exam.
The cardiac functional values for the right and left ventricles were obtained from the clinical database.Therefore, they were performed by multiple cardiac-trained radiologists and cardiologists.The radiologist and cardiologist both review the images on each exam and produce a single report.The cardiac functional values collected were indexed to the body surface area, and included right ventricular ejection fraction (RVEF), end diastolic volume (RVEDV i ), end systolic volume (RVESV i ), stroke volume (RVSV i ), and end diastolic mass (RVEDM i ).Our institution adopted definitions for all cardiac MRI cutoff values based on the literature which provides clinical consistency.Low normal is defined as RVEF of ≥48 to <50% and mild right ventricular functional impairment is defined as ≥40 to <48%.The analogous variables were collected for the left ventricle: ejection fraction (LVEF), end diastolic volume (LVEDV i ), end systolic volume (LVESV i ), stroke volume (LVSV i ), and end diastolic mass (LVEDM i ).All variables are indexed to the body surface area.Our institutional cutoff values for the LV are as follows: low normal is defined as LVEF of ≥52 to <55% and mild right ventricular functional impairment is defined as ≥40 to <52%.
The chest wall indices were collected from the clinical reports.The following indices are included: pectus (Haller) index (PI); correction index (CI) [1]; depression index [17] (DI); chest wall asymmetry ratio [18]; and sternal torsion [19].Normal average and standard deviation for PI reported by Haller et al. in the original paper was 2.56 ± 0.35 [15], and >3.25 is considered the threshold for PE repair.A CI threshold of 10.0% and above is proposed to define pectus patients with significant defects who should be considered for repair [18].The depression index (DI), a measure of the pectus depth divided by the width of a vertebral body, has a reported threshold for DI to indicate a PE deformity severe enough to consider surgical correction is >0.2 [17].Cardiac compression indices included modified cardiac compression index (mCCI) and modified cardiac asymmetry index (mCAI).These are labeled "modified" because they were modified for measurement on our MRI images which are obtained during end expiration and not during full inspiration on CT as described by Kim et al. [18] who used the xiphoid as a point of reference.We have modified the mCCI and mCAI so that it is measured at a uniform anatomic level on the heart since cardiac MRI is performed in expiratory breath hold.We measure at the level one slice above the edge of the visible eustachian valve on the axial single shot SSFP MR images, which approximates the level of the mid-tricuspid valve (TV) on axial images.Kim et al. report normal values of 1.1 to 2.0 for the CCI and 0.9 to 1.4 for the CAI [18].A guide to performing these measurements is provided in the Supplemental Material found online.These measurements were also the result of clinical studies and were performed by the radiologists and cardiologists who both read each exam at our institution.
Additionally, novel measurements of the cardiac structures include a semi-quantitative assessment of TV compression (Fig. 1).TV compression was evaluated as absent if the TV remained to the right of the sternum on the axial single-shot SSFP (Fig. 1a) which is obtained at end systole.TV compression was evaluated as partial compression if the valve plane was in contact with sternum, but to the right of midline (Fig. 1b, Supplemental Fig. S1) and complete if the valve was to the left of sternal midline (Fig. 1c, Supplemental Fig. S2).Categorization of TV compression was performed by cardiac imaging attending and second year medical student with, respectively, 20 years and 8 weeks of experience with cardiac imaging and performing these measurements.
The percentage of the cardiac transverse diameter displaced left of midline was also determined (Fig. 2).The TV and mitral valve (MV) were measured on the fourchamber cine images during diastole and the ratio of the two calculated as TV diameter divided by mitral valve diameter (Fig. 3).The TV and MV were also measured on cine phase contrast imaging (Fig. 4).The valves were measured with a linear measurement that would reflect the valve measurement on the four-chamber cine, but measured during peak flow across the valve which would represent early diastole.Additionally, the cross-sectional area of the flow across valve planes was measured (Fig. 5).Phase contrast cine MRI was performed according to accepted standards.A group of 100 measurements were repeated to obtain the interobserver agreement.The categorical values for classifying TV compression were scored by second year medical student (8 weeks of training and experience) and post medical school clinical research fellow with 2 years of experience.The research fellow utilized descriptions of the measures similar to that contained in this paper.The Kappa statistic for classifying TV compression  a, b, c).16-year-old male with pectus index of 3.9 and tricuspid valve (TV) compression absent (a).The TV (short arrow) is to the right of the sternal edge (long arrow) and not in contact with the sternum.Arrowhead is the center of the sternum.11-year-old female with pectus index of 3.6 and partial TV compression (b).The tricuspid valve (short arrow) is to the left of the sternal edge (long arrow) in contact with the sternum but not beyond the center of the sternum (arrowhead).During diastole, the tricuspid valve is still in contact with the lung pleura (Supplemental Fig. 1).14-year-old male with pectus index of 6.3 and complete TV compression (c).The tricuspid valve (short arrow) is to the left of the middle of the sternum (arrowhead) and in contact with the sternum.During diastole, the tricuspid valve is in contact with the chest wall (Supplemental Fig. 2)

Statistical analysis
Descriptive analysis was performed to describe the characteristics of the variables of interest.Continuous variables were presented as mean and standard deviation (SD) and were compared between three groups using one-way ANOVA.The interobserver agreement was evaluated using intraclass correlation coefficient with 95% confidence interval.All analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).A P-value less than 0.05 was considered statistically significant.

Results
A total of 661 different patients were identified with a complete set of data, duplicate exams on patients were removed, and all had pectus excavatum.The median age of these patients was 15 (13,16) and 532 (80%) were male.The overall median pectus index was 4.5 (3.9, 5.6).TV compression was present in 195 (29.5%), partial TV compression was present in 139 (21.0%), and no TV compression was present in 327 (49.5%) patients.The average age of the patients with TV compression was 16.3 ± 4.6 years, partial compression 15.3 ± 3.1 years, and those without TV compression were 14.9 ± 3 years (P=0.0006).There were no differences in height, weight, body mass index, or body surface area (Table 1).

Pectus-related indices
Of the chest wall indexes that are routinely reported at our institution, there were highly significant differences between the subjects with complete TV compression and those without or with partial TV compression (Table 2).
The pectus index (PI) measured as follows: without compression, 4.15 ± 0.94; with partial compression, 4.93 ± 1.24; and complete TV compression, 7.2 ± 4.01 (P<0.0001).The depression index (DI), a measure of the pectus depth divided by the width of a vertebral body [17], showed a similar pattern of mild worsening between no compression and partial compression, and a substantial increase in subjects with complete TV compression (Fig. 1).Subjects with no compression had a DI of 0.47 ± 0.2; subjects with partial compression had a DI of 0.56 ± 0.2; and subjects with complete TV compression demonstrated a DI of 0.82 ± 0.3 (P<0.0001).
Again, a similar pattern was observed in the correction index (CI): no compression, 23.8% ± 11.1; partial compression, 30.7% ± 12.1; and subject with complete TV 14-year-old male with pectus excavatum and pectus index of 6.3.TV compression was scored as complete in this patient.The contour labeled 1 represents the MV area of 6.66 cm 3 and the contour labeled 2 is the TV with an area of 6.79 cm 3  compression demonstrated a CI of 44.8% ± 15.5 (P<0.0001).Two other abnormal chest wall indexed values that showed the same consistent pattern were the chest wall asymmetry index (CWAI) and sternal torsion: chest wall asymmetry index (CWAI), no compression=0.98± 0.11, partial compression=1.01 ± 0.08, and complete TV compres-sion=1.08± 0.64, P=0.0005; and sternal torsion (degrees), no compression=10.7± 8.3, partial compression=11.7 ± 7.9, and complete TV compression=18.5 ± 10.7 (P<0.0001).

Cardiac indices
Cardiac functional measurements were all normalized by dividing the individual patient values by body surface area.It is a notable observation that there was no statistical difference in the height, weight, body mass index, or body surface area between groups (Table 1).The main finding in cardiac functional measurements is that there are differences in the ejection fraction and end systolic volume index of the right ventricle between patients with and without sternal compression of the TV (Table 3).

Valve measurements and heart displacement
The novel measurements of TV and MV diameter during diastole and their ratio were statistically significant to a high degree (Table 4).The measurements of the TV on the four-chamber view (mm) are as follows: no com-pression=29.52± 4.6; partial compression=28.26

Interobserver agreement
An interobserver agreement was determined between the two junior readers, a second year medical student with 8 weeks of experience and training with feedback and a physician research fellow with 2 years of experience with instructions as outlined in the manuscript.The intraclass correlation coefficient (Table 5) was excellent (>0.90) for heart transverse and heart left of midline measurements and good (>0.70) for measurements of the four-chamber TV, linear width and area of the TV, and area of the MV on the cine phase contrast images.However, the interclass correlation was poor between the linear measurements of the MV on the four-chamber cine and the cine phase contrast images.
The agreement between the evaluations of TV compression was measured using the Kappa statistic (Table 6).Second year medical student with 8 weeks of training and instruction had a Kappa=0.72,whereas research fellow using instructions as outlined in this paper demonstrated a Kappa=0.63.Both values are substantial in agreement with the standard reference, a cardiovascular imaging attending with 20 years of experience.

Discussion
This study demonstrates that increasing compression or restriction of the tricuspid valve results in lower RV systolic function compared to the group without compression of the tricuspid valve.For the tricuspid valve to be compressed by the sternum, a displacement of the heart toward the left hemithorax is required and is evident in the percentage that the heart is displaced into the left thorax across midline.Not unsurprisingly, all these are related to increased chest wall deformity indices.Overall, the relationship of chest wall indexes to decreased RV systolic function as a cause and effect is elusive and certainly is not a direct one-to-one relationship.Below we discuss the findings and possible suggestions as to the cause of decreased RV systolic function in the low normal range.
The demographics of the three groups of patients (no compression, partial compression, and complete compression of the TV) were notably consistent in height, weight, body mass index, and body surface area, and only different in age with patients progressively older with increasing severity of TV compression.The consistency of the demographics is likely due to the limited range of age in this pediatric referral center.The age difference is probably because the severity of the chest wall deformity worsens with age and the worsening of the chest wall deformity causes a greater shift in the cardiac structures to the left [20].
The finding of a consistent progressive worsening of the chest wall indices from no compression to complete TV compression is not surprising.First, the chest wall indices are all tied together by the shape of the chest wall with the use of either the distance from the sternum to spine, or the depth of the chest wall deformity in the numerator and another measure of body size in the denominator.The exceptions to this calculation are the CWAI and the sternal torsion which were both statistically significant.Both are tied to the chest wall deformity, though.Sternal torsion was measured in degrees, but the direction of the rotation was not noted as clockwise or counterclockwise which may come into play with shifting of the heart leftward.Ours was only an absolute value.Capunay et al. evaluated the effect of sternal torsion on PE and found that HI, CI, CCI, and TV/MV width were all affected by sternal torsion and that if the right side was down (counterclockwise), the effect was greatest [19].If counterclockwise was recorded as negative and clockwise as positive, the results may be different because of the competing positive and negative values in the analysis.Here, as an absolute value in degrees, it may only be linked to the severity of the chest wall deformity.
The mCCI and mCAI also reflect the severity of chest wall deformity and the lack of space for the heart in its usual location.As such, it is not surprising that the percentage of the heart displacement to the left of midline is also increased.Compression of the TV depends on the displacement of the heart to the left by the chest wall deformity.The severity of the chest wall deformity is related to cardiac function in many studies of PE [3][4][5].
However, the main hypothesis of this study was that increased compression of the right atrioventricular valve leads to decreased right ventricular systolic function.The ejection fraction of the right ventricle demonstrated a statistically significant difference with the highest difference occurring when progressing from no TV compression to partial TV compression, a difference of 1.41% and 0.69% to complete TV compression.It would seem the critical change is when the chest wall deformity is even partially limiting the motion of the TV and base of the RV.In patients without compression of the TV, the excursion of the TV annulus is sliding against the curvature of the pericardium that is up against lung pleura, whereas when partial or complete TV compression is present, the excursion is sliding horizontally up against the sternum and chest wall, limiting motion to two dimensions.This compression also limits the expansion of the base of the RV.The RVEDV i trends toward significance and RVESV i demonstrated a statistically significant increase.The overall RVSV is unchanged.This suggests that there is mild dilation of the RV to maintain stroke volume.The lack of statistical significance for the RVEDV i may be due to several factors.First, the RV is known to be more difficult to contour consistently due to the trabeculations, partial volume effects due to RV shape, and difficulty contouring at the plane of the atrioventricular valve resulting in higher interobserver and intraobserver variability than the left ventricle [21][22][23][24].Second, our research uses the clinical data derived by many readers rather than having a small group repeat the contours.The percent displaced, that portion of the heart displaced to the left of midline, is also increased from 65% ± 5 when no compression of the TV is present to 79% ± 7 when there is complete TV valve compression.When the heart is more displaced into the left chest, the RV is more able to dilate because of a larger interface with the lung, the rotation of the heart posteriorly and the anterior chest wall outward.This is supported by the abnormal RV systolic and diastolic function found during stress echocardiography and decreased tricuspid valve area compared to children without PE [9].Perhaps, using newly available FDA approved, clinically utilized artificial intelligence contour detection, and review of contours by one or two expert scorers would provide cleaner and more consistent results that are statistically significant [22,25,26].We looked at the TV size in relation to the degree of categorical compression: no TV compression, partial TV compression, and complete TV compression.In each group, the measurement of the TV decreased.When evaluating this on the four-chamber cine during diastole, the TV size (mm) decreased with each group from no compression to complete compression.The largest decrease is from partial compression to complete compression.This was also true when the TV was indexed to the size of the MV.Initially, the thought was that the MV would reflect the patient size, but the MV became larger with progressive chest wall deformity and compression of the TV.This is an interesting observation that is not due to the size of the subjects.Age may be a possibility in that each TV compression group is older as one progressed from no TV compression to complete TV compression.However, since there is no size difference in height, weight, body mass index, or body surface area, it suggests another possibility.Leftward displacement of the heart pushes the MV further away from the spine and "freeing it up" may allow the MV to expand slightly when it is not compressed.Another issue with measuring the TV on the four-chamber cine is that the TV flow is teardrop shaped and the width measured depends on where the placement of the imaging plane is and where it bisects the TV.It is for this reason that we also measure the TV linearly and in cross-sectional area on the phase contrast images.These were measured at peak flow because it is easiest to visualize the valve flow edges than the valve edges, whereas the four-chamber view was measured during diastole.On the PC of the TV, the relationship between TV compression was the same for both the linear and cross-sectional area measurements.Therefore, it is conclusive that TV compression is present when the sternum/chest wall deformity and TV are in contact throughout the cardiac cycle.Interestingly, Capunay et al. performed a similar evaluation when looking at sternal torsion and reported similar findings [19].The change in the cross-sectional area of the TV is unlikely to cause a substantial change in resistance to flow across the valve but adds support to the change in the shape of the right atrium and right atrioventricular junction.
Hypotheses for the cause of changes in right heart function are many, but in this study, the small change in the TV size raises the question as to whether that is contributing to the cause.We do not think the change in the TV size has any significant contribution to the decreased RV ejection fraction.From principles in fluid mechanics, the small change in size would cause no significant resistance or flow difference.However, the decrease in the TV valve size is not only associated with more compression of the heart in general but also with more deformity of the atrium and the right ventricle base.We hypothesize that one of the factors related to decreased RVEF in PE is the effect of the shape change of the atrium, TV, and RV on the normal flow vortices that maintain the momentum of blood in a normal heart as it enters the right atrium from the inferior and superior vena cava and crosses the TV [27,28].Prior authors have expressed several other hypothetical causes including twisting of the venous inflow, cardiac compression of the atria, limitation of cardiac expansion, and respiratory limitation that contributes to the mild cardiac function changes in PE [18].It would be very interesting to evaluate these vortices in normal subjects versus patients with PE using 4D flow imaging.Others have found RV diastolic function and systolic function in PE worsened with exercise stress [9].The mid-RV circumferential strain was found to be decreased by Mazur et al. with a concomitant increase in basal and apical strain which was hypothesized to be compensatory [29].Since a large fraction of patients with PE have compression of the mid-RV wall, this compression could restrict the end diastolic expansion of the RV circumference resulting in decreased strain and contractility.As the compression of the heart moves toward the AV groove, the affected volume of RV would have larger decreased strain and thus decrease function.Evaluation of strain based on where the sternum is compressing the RV is feasible.
The left ventricular cardiac indices did not show statistical significance, but it is interesting that LVEF trended downward with LVESV i and approached statistical significance.The differences were very small, and the LVEDV i showed no such trend and stroke volume index were very consistent between the subjects.Overall, the cardiac function in PE patient is shifted to the low part of the normal cardiac functional distribution [3,9,12,30].
There are limitations to this study.First, the values used for cardiac function and chest wall indices were derived from clinical reports which likely increased variability over having a few experts reproduce these analyses.Second, it is a retrospective study without normal controls.Third, some would state that using body surface area to normalize the cardiac functional values is overly influenced by obesity.However, patients with PE tend to be a very asthenic group and only four had a BMI over 30.Fourth, tricuspid valve compression related to Haller index, depression index, shift of the heart to the left, and other measures of chest wall deformity suggest that the usefulness of this finding is limited.Additionally, the small decrease in the right ventricular systolic is of doubtful clinical significance as this study does not include clinical symptomology or outcomes.Overall, it does describe changes in the region of the tricuspid valve and adds to the cardiac function in pectus excavatum literature.

Conclusion
Our study demonstrates that compression of the tricuspid valve is associated with decreased function of the right ventricle and increased pectus excavatum indices.In most patients, RV systolic function remains in the normal range but on the left side of the bell curve.It also provides further evidence of decreased cardiac function due to PE highlighting the need for further research to better understand cause and effect between chest wall deformity and cardiac function in affected individuals.

Fig. 1
Fig. 1 Axial single-shot steady-state free precession with cardiac end systolic gating and respiratory gating (a, b, c).16-year-old male with pectus index of 3.9 and tricuspid valve (TV) compression absent (a).The TV (short arrow) is to the right of the sternal edge (long arrow) and not in contact with the sternum.Arrowhead is the center of the sternum.11-year-old female with pectus index of 3.6 and partial TV compression (b).The tricuspid valve (short arrow) is to the left of the sternal edge (long arrow) in contact with the sternum but not

Fig. 2 Fig. 3
Fig. 2 Percentage of the cardiac transverse diameter displaced left of midline (cardiac displacement).Axial single-shot steady-state free precession of a 13-year-old male with pectus index of 7.1.The thin yellow line is the center of the chest, and the gray dashed line (a) represents the portion of the heart left of midline.The yellow double arrow line (b) is the transverse diameter of the heart.Percent of cardiac displacement=(a/b)×100

Fig. 4 Fig. 5
Fig. 4 Tricuspid (TV) and mitral valve (MV) measurement.Cine phase contrast across the atrioventricular valve during peak flow.15-year-old male with pectus index of 4.5.The yellow double arrow line (TV) is the tricuspid valve measurement.The yellow dash double arrow line (MV) is mitral valve measurement

Table 1
Demographics of each group.BMI body mass index, BSA body surface area

Table 3
Cardiac function indices results.RV right ventricle, LV left ventricle, EF ejection fraction, EDV end diastolic volume, BSA body surface area, ESV end systolic volume, ED end diastolic, ml milliliter, m 2 meter squared,

Table 6
Agreement between reader and the reference.CI confidence interval