Breathe In... Breathe Out... Stop Breathing: Does Phase of Respiration Affect the Haller Index in Patients With Pectus Excavatum?

Transcription

1 Pediatric Imaging Original Research irkemeier et al. Haller Index in Patients With Pectus Excavatum Pediatric Imaging Original Research Krista L. irkemeier 1 Daniel J. Podberesky 1 Shelia Salisbury 2 Suraj Serai 1 irkemeier KL, Podberesky DJ, Salisbury S, Serai S Keywords: expiration, inspiration, Haller index, MRI, pectus excavatum, pediatric, phase of respiration DOI: /JR Received January 4, 2011; accepted after revision pril 17, Department of Radiology, incinnati hildren s Hospital Medical enter, 3333 urnet ve, ML 5031, incinnati, OH ddress correspondence to K. L. irkemeier (Krista.irkemeier@cchmc.org). 2 Division of iostatistics and Epidemiology, incinnati hildren s Hospital Medical enter, incinnati, OH. ME This article is available for ME credit. See for more information. WE This is a Web exclusive article. JR 2011; 197:W934 W X/11/1975 W934 merican Roentgen Ray Society reathe In... reathe Out... Stop reathing: Does Phase of Respiration ffect the Haller Index in Patients With Pectus Excavatum? OJETIVE. The purpose of this article is to determine whether the phase of respiration at the time of imaging affects chest wall measurements and compression of internal structures in patients with pectus excavatum. MTERILS ND METHODS. Forty-seven patients (median age, 14 years) imaged for preoperative pectus excavatum underwent limited axial balanced steady-state free precession MRI of the chest at inspiration, expiration, and stop quiet breathing. Two radiologists, who were blinded to prior measurements, independently calculated the Haller index, asymmetry index, and sternal tilt in each phase of respiration. ompression of internal structures was recorded. Statistical comparison was performed. RESULTS. The Haller index was significantly lower at inspiration, compared with stop quiet breathing and expiration, with medians (interquartile ranges) of 3.96 ( ), 5.16 ( ), and 5.09 ( ), respectively (p < for both). No significant difference in Haller indexes was observed between expiration and stop quiet breathing (p = ). Of 11 patients with a Haller index less than 3.25 at inspiration, eight (72.7%) had an index greater than 3.25 on expiration and stop quiet breathing, which accounted for 17% (8/47) of all patients imaged. ompression of the liver or vascular structures was present in 24 (51%) patients. There was no significant difference in the asymmetry index, sternal tilt, or right heart compression between phases of respiration. ONLUSION. Obtaining the Haller Index at inspiration may result in a value significantly lower than that at expiration, potentially affecting surgical and financial decision making. ompression of the liver and vascular structures was observed in 51% of patients, but additional research is needed to determine the clinical significance of this finding. T he most widespread currently accepted radiographic method of assessing pectus excavatum is to obtain the Haller index, also known as the pectus index, by either T or radiography. This index is commonly requested by surgeons for preoperative planning and assessment of severity and is often required by third-party payers as a prerequisite for payment [1 4]. s described by Haller and colleagues [5], the Haller index is the ratio of the inner transverse diameter of the thorax to the anteroposterior diameter from the anterior margin of the vertebral body to the posterior aspect of the sternum, obtained by a single T slice at the level of greatest chest depression. However, Haller et al. did not describe in which phase of respiration the scan was performed. The purpose of this work is to determine whether different phases of respiration yield significant differences in the Haller index, the asymmetry index, the sternal tilt, and compression of internal anatomic structures. The use of MRI allows rapid evaluation of the chest in multiple phases of respiration without the use of ionizing radiation. Materials and Methods Institutional review board approval was obtained for this retrospective HIP-compliant study. Imaging The MRI scans were acquired on a 3-T scanner (chieva, Philips Healthcare), equipped with 32 receiver channels and with a gradient performance of 40 mt/m using a 32-channel cardiac array coil, and on a 1.5-T scanner (HDxt, GE Healthcare), equipped with eight receiver channels with a gradient performance of 50 mt/m using an eight-channel cardiac array coil. None of the patients required sedation. rapid survey scan was obtained to identify W934 JR:197, November 2011

2 Haller Index in Patients With Pectus Excavatum Fig year-old boy with pectus excavatum imaged at full expiration. Horizontal dotted line denotes level of deepest depression of anterior chest wall because it is asymmetric and off midline, to obtain accurate anteroposterior dimension. Solid vertical line () represents anteroposterior measurement, and solid horizontal line (T) represents transverse internal thoracic measurement used to calculate Haller index of 6.37 in this patient. Vertical dashed lines represent right (R) and left (L) anteroposterior hemithorax dimensions used to calculate asymmetry index. Dashdot-dot lines (S) represent sternal angle. axial slice locations, using the proximal and distal ends of the lungs as landmarks. Next, three axial scans were acquired sequentially using a 2D balanced steady-state free precession pulse sequence. In a subset of patients who were also undergoing subsequent cardiac MRI, respiratory bellows were placed around the chest at the approximate level of the diaphragm, allowing the technician to monitor appropriate depth of inspiration or expiration. The first scan was acquired with a full inspiration breath-hold. reathing instructions were to take a deep breath in, blow it out, take a deep breath in, and hold it. The second scan was acquired with full expiration breath-hold. reathing instructions were to take a deep breath in, blow it all the way out, and hold it. The third scan was acquired with the patient holding his or her breath in the middle of a normal respiratory cycle. reathing instructions were to breathe normally and then stop breathing. ll patients adequately followed breathing instructions by subjective evaluation, and in those with respiratory bellows, adequate depth of respiration could be confirmed objectively as well. On the basis of the patient s torso size on the survey, respiratory pattern, and number of slices needed to cover the chest, the acquisition time for these three sequential scans was planned in advance to be approximately seconds each, with no need of physical patient movement in the longitudinal direction. The entire chest was imaged axially with 5-mm-thick slices. Patients Subjects were identified through the hospital s radiology report search program and the radiology PS (MIS PS 6.0 Workstation, Merge Healthcare). ll MRI examinations performed for pectus evaluation from November 2009 through September 2010 were assigned a randomized study identification number. Studies were placed in the PS under that identification number so that images could be viewed in a blinded fashion with no clinical information, radiology report, or prior measurements available to the reader. Of a total of 50 patients referred by pediatric surgeons at a large academic children s hospital for preoperative pectus evaluation, 46 patients 8 20 years old (median age, 14.0 years) and one 36-year-old patient with pectus excavatum were included. Two patients were excluded for pectus carinatum and another patient was excluded from phase of respiration comparison because of an incomplete MRI examination, thus yielding a total of 47 MRI examinations for review. Data cquisition The radiologists reviewed the MRI examinations and performed measurements to calculate the Haller index, asymmetry index, and sternal tilt in each phase of respiration. For each phase of respiration, the axial slice with the greatest degree of anteroposterior narrowing was identified. On that slice, the Haller index was calculated by obtaining maximum transverse diameter of the chest between the inner rib margins and dividing it by the minimum anteroposterior diameter from the anterior aspect of the spine to the most posteriorly displaced portion of the anterior chest wall (Haller index = transverse diameter / anteroposterior diameter). When the deepest anterior chest wall deformity was not in the midline, a transverse line parallel to the transverse diameter measurement was placed at the deepest extent of the anterior wall and the most anterior aspect of the vertebral body. The anteroposterior measurement between the two parallel lines was the anteroposterior measurement (Fig. 1). t the same slice, the asymmetry index was calculated by obtaining the greatest anteroposterior diameter of the right and left chest cavities (Fig. 1), dividing the right anteroposterior diameter by the left, and multiplying by 100 (asymmetry index = [right anteroposterior diameter / left anteroposterior diameter] 100). ll measurements were obtained using the PS electronic calipers. The sternal tilt was calculated using the PS angle tool at the level of greatest sternal tilt (Fig. 1). Right heart compression was subjectively graded as none, mild, moderate, or severe. The presence of compression of other internal structures was noted as free text. Statistical nalysis SS software (version 9.2, SS Institute) was used for all analyses. The primary outcomes of interest were the Haller index, asymmetry index, and sternal tilt. Statistical analysis to compare each outcome of interest among the three breathing stages was performed using the runner nonparametric method for repeated measures. The outcomes were transformed using the Rank procedure, and then the analysis of variance statistics and resulting p values were computed using the Proc Mixed procedure. The unstructured covariance matrix was specified, and estimations were based on the minimum variance quadratic unbiased estimation method. SS uses the second degree of freedom rather than the first degree of freedom to estimate the p value for the analysis of variance statistic, resulting in conservative p values. n additional step was performed to improve the estimated p values using the first degree of freedom and the analysis of variance statistic from the SS procedure. For each reader, the nonparametric Wilcoxon signed rank test was used to assess differences in TLE 1: Haller Index Data at Three Different Stages of Respiration Minimum Maximum reathing Stage Median Lower Upper Quartile a Inspiration 3.96 b Quiet breathing Expiration a The lower and upper quartiles refer to the 25th and 75th percentiles, respectively. b p < versus Haller index at quiet breathing and end expiration. JR:197, November 2011 W935

3 irkemeier et al. TLE 2: symmetry Index Data at Three Different Stages of Respiration Minimum Maximum reathing Stage Median Lower Upper Quartile a Inspiration Quiet breathing Expiration a The lower and upper quartiles refer to the 25th and 75th percentiles, respectively. TLE 3: Sternal Tilt Data at Three Different Stages of Respiration Quartile a reathing stage Median Lower Upper Minimum Maximum Inspiration Quiet breathing Expiration a The lower and upper quartiles refer to the 25th and 75th percentiles, respectively. heart compression between inspiration and expiration. Only descriptive statistics were used for compression of other internal structures. The Freq procedure was used to determine the distribution of subjects with Haller index above and below the value of 3.25 for the three breathing stages. Data are presented as median (interquartile range) for nonnormally distributed data and as frequencies and proportions for categoric data. ll tests were two sided, and p values of 0.05 or less were considered significant. Reported p values are not adjusted for multiple comparisons. Results There was a significant difference in the Haller index among the stop breathing stages. Fig year-old woman with pectus excavatum., Image was obtained at full inspiration. Horizontal and vertical lines depict anteroposterior narrowing used to obtain Haller index of 7.56., Image was obtained at full expiration. Horizontal and vertical lines depict anteroposterior narrowing used to obtain Haller index of 13.93, significantly higher than Haller index at inspiration., Image was obtained at stop quiet breathing. Horizontal and vertical lines depict anteroposterior narrowing used to obtain Haller index of 13.29, similar to Haller index at expiration. The Haller index was lower for inspiration compared with both stop quiet breathing and expiration (p < for both). The medians were 3.96 ( ) for inspiration, 5.16 ( ) for stop quiet breathing, and 5.09 ( ) for expiration. No statistically significant difference was observed between quiet breathing versus expiration Haller index (p = ). The Haller index data for each breathing stage are provided in Table 1, and Figure 2 depicts an MRI example. There was no statistically significant difference in the asymmetry index (p = ) or the sternal tilt (p = ) among the breathing stages. symmetry index data for each breathing stage are provided in Table 2. Sternal tilt data for each breathing stage are provided in Table 3. There were 11 subjects with a Haller index less than 3.25 on inspiration. Of these, eight subjects (72.7%) had a Haller index of 3.25 or higher with expiration (Fig. 3). The same eight subjects had a Haller index of 3.25 or higher at stop quiet breathing. There were three subjects with a Haller index less than 3.25 on expiration, and these same subjects had a Haller index of less than 3.25 during stop quiet breathing and full inspiration as well. These were the only three subjects with a Haller index less than 3.25 on stop quiet breathing. The subjective degree of right heart compression on expiration was not significantly W936 JR:197, November 2011

4 Haller Index in Patients With Pectus Excavatum different than in inspiration for either radiologist (reader 1, p = ; reader 2, p = ). We observed differences in mass effect on internal structures at different phases of respiration (Figs. 4 6). We observed compression of internal structures in 24 patients (51%), including the liver, inferior vena cava, brachiocephalic vein, and displacement of the aorta, celiac axis, and superior mesenteric artery. In one case, there was a greater degree of brachiocephalic vein compression at inspiration; in all other cases, when there was a difference in the degree of compression, it was subjectively less on inspiration. Fig. 3 8-year-old boy with pectus excavatum., Image was obtained at full inspiration. Horizontal and vertical lines depict anteroposterior narrowing used to obtain Haller index of 3.05, which would be considered normal., Image was obtained at full expiration. Horizontal and vertical solid lines depict anteroposterior narrowing used to obtain Haller index of 5.00, which would be considered abnormal. Horizontal dotted line denotes level of deepest depression of anterior wall because it is asymmetric and off midline, to obtain accurate anteroposterior dimension., Image was obtained at stop quiet breathing. Horizontal and vertical solid lines depict transverse and anteroposterior measurements, respectively, at level of greatest anteroposterior narrowing used to obtain Haller index of Horizontal dotted line denotes level of deepest depression of anterior wall because it is asymmetric and off midline, to obtain accurate anteroposterior dimension. Fig year-old boy with pectus excavatum., Image was obtained at full expiration. Inferior vena cava (long arrow) is nearly obliterated because of compression of liver and inferior vena cava from narrowed anteroposterior diameter of lower chest. orta (short arrow) is displaced to left., Image was obtained at full inspiration. Inferior vena cava (long arrow) returns to normal diameter at full inspiration. orta (short arrow) is closer to midline. Discussion The present study shows that the Haller index measured in full inspiration is significantly lower than that at either full expiration or stop quiet breathing. ecause Haller et al. [5] did not describe a specific protocol for their study in terms of depth of respiration at the time of T, we cannot be certain that we are performing comparable measurements today. t the time of the study by Haller and colleagues, the technology was such that a T scan of the entire chest could not be performed JR:197, November 2011 W937

5 irkemeier et al. Fig year-old man with pectus excavatum., Image was obtained at full expiration. Origin of celiac artery (arrow) is displaced and acutely angulated., Image was obtained at full expiration. Liver (long arrow) is compressed, inferior vena cava (thin short arrow) is mildly flattened, and aorta (thick short arrow) is displaced posteriorly to left., Image was obtained at full inspiration. There is resolution of liver (long arrow) and inferior vena cava (thin short arrow) compression, decreased aortic displacement, and loss of acute angulation of celiac artery (thick short arrow). in one breath-hold, and so the standard was to obtain studies during quiet breathing. s T scanners have evolved and become faster, it is now standard to obtain images at full inspiration unless specifically evaluating for air trapping. t our institution, performing unenhanced chest T scans at full inspiration for examinations ordered to evaluate pectus excavatum was the standard for many years. dditionally, Haller refers to the Welch index as one of the previous measures of chest wall deformity. Welch [6] describes specifically that the radiographs to obtain his measurements are taken neither in deep inspiration nor in full expiration, which we would interpret to mean that the images were obtained at stop quiet breathing. With our analysis, we have determined that Haller indexes taken at stop quiet breathing are significantly higher than those obtained at full inspiration. For this discussion, a Haller index of less than 3.25 will be referred to as normal, and an index greater than 3.25 will be referred to as abnormal, because this is the current convention. In patients with a normal Haller index of less than 3.25 at expiration or stop quiet breathing, there was also a normal Haller index on inspiration. However, in eight of 11 patients with a normal Haller index on inspiration, the Haller index on expiration and stop quiet breathing was abnormal. Depending on how stringently the surgeon uses the Haller index, this could affect surgical decision making. lso, depending on how strict the insurance company is regarding the index, it may affect payment. Three major insurance companies require a Haller index of greater than 3.25, and one requires a Haller index of greater than 3.2 to qualify for financial payment of surgical correction [1 4]. Of our patients imaged during inspiration, which is the way T is typically performed, 72.7% of patients with a normal Haller index may be abnormal if imaged in a different phase of respiration. This accounted for 17% of all the patients we imaged during the study period. We are not implying that the most accurate measure of surgical candidacy is decided by a scan in a certain phase of respiration. However, if the Haller index is to be held as the standard measure, it is important to consider the potential underestimation of the index if it is obtained during inspiration. There are very few published data regarding the differences in chest wall measurements, particularly the Haller index, at different phases of respiration. Raichura et al. [7] examined six patients with pectus excavatum and six matched control subjects and noted that the anteroposterior and transverse dimensions both decreased at deep inspiration for both groups. They did not specifically calculate and compare Haller indexes at different stages of respiration. Herrmann and colleagues [8] evaluated chest wall motion during respiration and measured the Haller index, but they did not look at Haller indexes at different phases of respiration to compare them. We did not find a significant difference in the asymmetry index or the sternal tilt between different phases of respiration. We also did not find a significant difference in right heart compression between inspiration and expiration. However, we did notice subjective differences in mass effect on internal structures, such as the inferior vena cava, left brachiocephalic vein, and liver, between different phases of respiration in some patients. Further research in this area should include outcomes studies of surgical repair and correlation with the Haller index and compression of internal structures in different phases of respiration to determine which phase of respiration is most useful in predicting surgical outcomes. W938 JR:197, November 2011

6 Haller Index in Patients With Pectus Excavatum In conclusion, the Haller index measured during MRI is significantly lower when measured at inspiration compared with stop quiet breathing and full expiration. Measurement at inspiration, as T is typically performed currently, resulted in a Haller index of less than 3.25 in 17% of patients referred for imaging in this investigation, who would have had a Haller index of greater than 3.25 if measured during another phase of respiration. This difference in measurement could affect surgical and financial decision making regarding pectus excavatum surgery. In addition, further research regarding the potential clinical significance of compression of internal structures is needed, because there may be hemodynamic consequences of compression of vascular structures. cknowledgments We thank lan rody, MD, for editorial assistance, Judy Hardin for collecting the necessary studies, Rose Martin for assistance with anonymizing the study patients and creating a PS study database, Debby Paradiso and Penny New for administrative assistance, and Ulrike ttenberger, MD, for assistance with German translation. References 1. etna. linical policy bulletin: pectus excavatum and Poland s syndrome surgical correction. No etna Website. data/200_299/0272.html. Published July 28, Updated July 23, ccessed June 21, nthem. Medical policy: cosmetic and reconstructive services of the trunk and groin. nthem Website. policies/mp_pw_a htm. Published Updated ccessed June 21, igna. igna medical coverage policy: surgical treatment of chest wall deformities. igna Website. www. cigna.com/customer_care/healthcare_professional/ Fig year-old boy with pectus excavatum., Image was obtained at full expiration. Thoracic inlet (horizontal lines) is narrowed, and left brachiocephalic vein (thick arrow) is narrowed. Trachea (thin arrow) is displaced to right but not narrowed., Image was obtained at full inspiration. Thoracic inlet (horizontal lines) is less narrowed, and left brachiocephalic vein (thick arrow) is not compressed. Trachea (thin arrow) is midline., Image was obtained at stop quiet breathing. Thoracic inlet (horizontal lines) is narrowed but there is no luminal narrowing of left brachiocephalic vein (thick arrow). Trachea (thin arrow) is slightly deviated to right but is not narrowed. coverage_positions/medical/mm_0309_coveragepositioncriteria_surgical_treatment_chest_wall_ deformities.pdf. Published March 15, Updated March 15, ccessed June 21, Humana. Humana medical coverage policy: chest wall deformities (pectus excavatum/carinatum and Poland syndrome) surgical treatments. Humana Website. apps.humana.com/tad/tad_new/ home.aspx. Published June 28, Updated June 24, ccessed June 21, Haller J Jr, Kramer SS, Lietman S. Use of T scans in selection of patients for pectus excavatum surgery: a preliminary report. J Pediatr Surg 1987; 22: Welch KJ. Satisfactory surgical correction of pectus excavatum deformity in childhood: a limited opportunity. J Thorac Surg 1958; 36: Raichura N, Entwisle J, Leverment J, eardsmore S. reath-hold MRI in evaluating patients with pectus excavatum. r J Radiol 2001; 74: Herrmann K, Zech, Strauss T, Hatz R, Schoenberg S, Reiser M. ine MRI of the thorax in patients with pectus excavatum (in German). Radiologe 2006; 46: FOR YOUR INFORMTION This article is available for ME credit. See for more information. JR:197, November 2011 W939