Pediatr Cardiol DOI 10.1007/s00246-014-0900-z
ORIGINAL ARTICLE
Yield of Cardiac Magnetic Resonance Imaging as an Adjunct to Echocardiography in Young Infants With Congenital Heart Disease Joyce T. Johnson • Kimberly M. Molina Molly McFadden • L. LuAnn Minich • Shaji C. Menon
•
Received: 18 November 2013 / Accepted: 25 March 2014 Ó Springer Science+Business Media New York 2014
Abstract Echocardiography provides adequate preoperative imaging for most young infants with congenital heart disease (CHD). When anatomic details require further clarification, cardiac magnetic resonance imaging (CMRI) may be useful but adds the risks of sedation or general anesthesia for a vulnerable population. This study aimed to determine the safety of CMRI and its yield of additional significant information for this population. The study identified all infants age 90 days or younger with preoperative echocardiography and a CMRI from the period 2002–2012. Indications, complications, and imaging results were collected. The additional CMRI information was defined as ‘‘significant’’ if it altered surgical management or ‘‘not significant’’ if it did not. Associations between indications for CMRI and the likelihood of new significant findings were sought. For 137 infants (58 % male), CMRI was performed at a median age of 5 days (range, 0–89 days). The CMRI yielded additional information for 76 % (104/137) of the patients. The additional findings were significant for 69 % (72/104) of these patients. The incidence of significant new findings was similar among indication categories. All the infants were intubated. Complications occurred for 5 % of the patients, including one subject with a bradycardic event that prevented completion of the exam and six patients with transient vital sign changes that allowed exam completion. More than 50 % of young infants with CHD who
J. T. Johnson (&) K. M. Molina L. L. Minich S. C. Menon Division of Cardiology, Department of Pediatrics, University of Utah and Primary Children’s Hospital, 100 N. Mario Capecchi Drive, Salt Lake City, UT 84113, USA e-mail: [email protected]; [email protected] M. McFadden Division of Epidemiology, Department of Medicine, University of Utah and Primary Children’s Hospital, Salt Lake City, UT, USA
underwent preoperative CMRI had new findings affecting surgical management. Among these patients, CMRI-associated complications were few and predominantly minor for intubated infants. Further studies to determine standard preoperative criteria for the use of CMRI for infants with CHD may help to define appropriate cost-effective use of this diagnostic method. Keywords Surgical planning Cardiac magnetic resonance imaging Congenital heart disease Infant
Introduction Echocardiography is the imaging method of choice for defining the preoperative anatomy and physiology of infants with congenital heart disease (CHD) [9]. For cases in which echocardiography fails to provide sufficient information, cardiac catheterization has been the traditional alternative approach despite the associated risks and exposure to ionizing radiation [8, 12]. Cardiac magnetic resonance imaging (CMRI) is an alternative noninvasive imaging method that avoids these risks [11]. Although acoustic windows in infants usually allow an excellent echocardiographic definition of cardiac anatomy and function, this technology has limitations. Noncardiac chest anomalies, lung disease, and extracardiac lesions may interfere with echocardiographic imaging. The yield of additional preoperative information that affects surgical management, the potential to identify anatomy missed by echocardiography, and the complications associated with CMRI have not been studied systematically. Previous studies have evaluated the yield of new imaging methods and their effect on surgical management and have provided guidance for their incorporation into practice [1, 6]. We used this methodology with the aim of determining the
123
Pediatr Cardiol
added clinical benefit and risks of CMRI for preoperative planning for young infants with CHD. We hypothesized that CMRI for preoperative evaluation of infants with CHD will safely provide additional information that affects surgical planning. In addition, we aimed to evaluate a group of patients without preoperative CMRI but with findings by intraoperative assessment that differed from echocardiographic findings to assess further the ideal indications for CMRI.
Methods Study Population In this retrospective cohort study performed to evaluate the utility of CMRI in young infants with CHD, we assessed two groups. The CMRI group was identified by searching the Pediatric CMRI Database for the period 2002–2012 to identify infants age 90 days or younger with CHD who had preoperative echocardiography and CMRI. Infants who had undergone cardiac surgery before CMRI were excluded from the study. The CMRI group allowed us to determine the utility of CMRI for patients who had the exam. We used a quality assurance (QA) group to determine whether surgically relevant information not obtained by preoperative echocardiography could have been identified by CMRI had it occurred. We searched the Pediatric Echocardiography Quality Assurance Database for all existing years (2005–2012) to identify surgically relevant discrepancies between preoperative echocardiographic diagnoses and intraoperative assessment at the first cardiac surgery. Those who had a CMRI before surgery and children older than 4 years at the time of surgery were excluded from the QA group because all surgically relevant information would likely have manifested by the age of 4 years. The QA group had no patients who also were in the CMRI group. The study was approved by the Primary Children’s Hospital Privacy Board and the University of Utah Institutional Review Board under a waiver of consent. Medical Record Data Collection The medical records of both groups were reviewed for demographic information, diagnoses, date and type of surgical procedures and imaging studies, and surgical descriptions of the anatomy. Complications and indications for intubation in the CMRI group were recorded. Imaging Data Collection For the CMRI group, the echocardiographic and CMRI images and reports from both imaging studies were reviewed for echocardiographic diagnoses, indications for
123
CMRI, CMRI diagnoses, details of anatomy, and function/ flow analysis. All studies were performed using a GE Medical Systems (Milwaukee, WI, USA) 1.5-T Signa LX Echospeed MRI scanner. Torso or cardiac-phased array coils were chosen according to the patient’s body size. For patients under general anesthesia, breathholding was performed to limit motion artifact. No changes were made in the scanner or sequences used at our center during the study period. The common imaging sequences used for structural and functional evaluation of structural heart disease in this cohort included a three-plane localizer; double- and tripleinversion recovery fast spin echo (FSE); electrocardiogram-triggered, segmented k-space, fast-spoiled, gradient-recalled cine sequences (SSFP); free-breathing cine phase contrast flow measurements (40 phases per cardiac cycle) with retrospective cardiac gating; multi-slab 3D whole heart SSFP T2 prep navigator sequence; and inversion recovery-prepared gradient-echo sequence 10 to 20 min after an intravenous bolus of 0.2-mmol/kg gadolinium-based contrast. The images were transferred to a work station for postprocessing (MASS; Medis, Leiden, The Netherlands). The subjects were grouped by indication into the following categories based on the anatomic area involved: pulmonary arteries, pulmonary and systemic veins, aortic arch, septation defects, coronary artery anatomy, valves, or other (i.e., masses, functional and flow analysis). Any additional information provided by CMRI and not available from the echocardiogram was dichotomized as ‘‘significant’’ or ‘‘not significant’’ using previously published methods for determining the yield of a new imaging technique [6], in which significant findings were defined as new CMRI information that affected the surgical management plan or would have been life threatening if unknown. New findings that did not affect surgical management were defined as not significant. The significance of the additional findings obtained from CMRI was independently categorized by two investigators (K.M., J.T.J.) and adjudicated by a third investigator (L.M.) as needed. For the QA group, new information described at surgery and not obtained by preoperative echocardiography was defined as ‘‘significant’’ or ‘‘not significant’’ using the same definitions. The feasibility of CMRI to detect the clinically significant echocardiographic inaccuracy was independently determined by two investigators experienced in the use of CMRI. The echocardiographic findings and the CMRI findings were collected via chart review and compared with the findings listed in the operative report. Because the goal for the ‘‘significant’’ or ‘‘not significant’’ categorization was consensus, intra- and interobserver reliability was not addressed.
Pediatr Cardiol
Statistical Analysis For the analysis, indication categories for the CMRI and QA groups were chosen as predictors of the outcome for each group. The outcome was defined as the incidence of significant new information provided by either CMRI or intraoperative assessment that was not obtained by preoperative echocardiography. Continuous data were expressed as means ± standard deviations or medians with ranges, as appropriate, and categorical data were tabulated. For the CMRI group, univariate logistic regression was used to assess the association between each indication category for CMRI and the odds of significant new findings compared with the remaining subjects. Likewise, the QA group was evaluated using univariate logistic regression for each indication category and the significance of new information provided by intraoperative assessment that could have been diagnosed by CMRI. Because confounding of significance by anything other than the indication category was unlikely and the results by univariate analysis were not significant, a
Table 1 Group characteristics CMRI group (n = 137) n (%)
QA group (n = 85) n (%)
Males
79 (58)
38 (45)
Median age at CMRI: days (range)
5 (0–89)
NA
Mean weight at CMRI (kg) Indication categories
3.16 ± 0.73
NA
Aortic arches
59 (43)
8 (9)
Pulmonary artery anatomies
41 (30)
2 (2)
Pulmonary or systemic vein anatomies
25 (18)
10 (12)
Septation defects
0 (0)
26 (31)
Coronary arteries
0 (0)
15 (18)
Valve abnormalities
0 (0)
22 (26)
Others
11 (9)
0 (0)
19 (14)
2 (2)
Heterotaxy syndromes
CMRI, cardiac magnetic resonance imaging; QA, quality assurance; NA, not applicable
Table 2 Odds of significant new cardiac magnetic resonance imaging (CMRI) information by univariate logistic regression
multivariate analysis was not performed. A two-sided p value lower than 0.05 was considered statistically significant. The statistical analyses were performed using Stata release 12 (StataCorp. 2011, Stata Statistical Software: Release 12; College Station, TX, USA).
Results CMRI Group The CMRI group consisted of 137 subjects with a median age of 5 days (range, 0–89 days) (Table 1). In this study, CMRI was performed after one echocardiogram for 39 % (53/137) of the subjects, after two echocardiograms for 35 % (48/137) of the subjects, and after three or more echocardiograms for 26 % (36/137) of the subjects. New CMRI information was noted for 76 % (104/137) of the CMRI group, with 69 % (72/104) of these new findings significant and altering the surgical plan for 53 % (72/137) of the group with significant findings. Of the 19 subjects with heterotaxy, CMRI yielded significant findings for 37 % (7/19). The univariate logistic regression analysis showed no difference between the odds of significant new information from CMRI for the indication categories (Table 2). For each category, the proportion of those with the specified indication who had a significant finding was compared with the proportion of those without the specified indication who had a significant finding. All the subjects were sedated and intubated for CMRI, with 66 % (90/137) undergoing intubation solely for the purpose of obtaining the CMRI. Complications occurred for 5 % (7/137) of the subjects, including one subject who had a bradycardic event treated with chest compressions and epinephrine with abortion of the CMRI. The remaining six subjects had transient vital sign changes responding to fluid bolus or ventilation changes, with successful completion of the CMRI. QA Group The diagnosis for 51 % (43/85) of the 85 subjects with significant new information from intraoperative assessment
Indication category
Proportion of those with specified indication with significant finding n (%)
Aortic arches
29/59 (49)
Pulmonary arteries
26/41 (63)
Systemic and pulmonary veins Heterotaxy syndromes
Proportion of those without specified indication with significant finding n (%)
Odds ratio (95 % CI)
p Value
43/78 (55)
0.8 (0.4–1.5)
0.5
46/96 (48)
1.9 (0.9–4.0)
0.1
10/25 (40)
62/112 (55)
0.5 (0.2–1.3)
0.2
7/19 (37)
65/118 (55)
0.5 (0.2–1.3)
0.1
123
Pediatr Cardiol Table 3 Description of anatomic categories and proportion of findings that could have been diagnosed by cardiac magnetic resonance imaging (CMRI) had they occurred before surgery Indication category (n)
CMRI could have determined diagnosis n (%)
Aortic arches (8)
8 (100)
Heterotaxy syndromes (2)
2 (100)
Systemic and pulmonary veins (10)
9 (90)
Septation defects (26)
18 (69)
Pulmonary arteries (2)
1 (50)
Valves (22)
4 (15)
Coronaries (15)
0 (0)
could have been determined by a preoperative CMRI. The characteristics of the QA group are listed in Table 1. Both subjects with heterotaxy had significant new information obtained from intraoperative assessment that could have been diagnosed by CMRI. In contrast, only 15 % of the valvar findings could have been diagnosed by CMRI and were limited to distinguishing supravalvar from valvar stenosis. None of the coronary artery anomalies detected by intraoperative assessment were likely to be diagnosed with current CMRI techniques in the young infants (Table 3). The odds that CMRI could have identified significant new anatomic information from intraoperative assessment was similar among the indication categories (Table 4). For each category, the proportion of those who had the specified indication with a significant finding was compared with the proportion of those without the specified indication who had a significant finding.
Discussion Our study demonstrated that CMRI in young infants with CHD is useful in surgical planning. After preoperative echocardiography, CMRI provided additional anatomic
Table 4 Odds of significant new intraoperative information that could have been observed by CMRI using univariate logistic regression
Indication category
Proportion of those with specified indication with significant finding n (%)
Proportion of those without specified indication with significant finding n (%)
Odds ratio (95 % CI)
p Value
Aortic arches Pulmonary arteries
6/8 (75) 1/2 (50)
55/77 (71) 60/83 (72)
1.2 (0.2–6.4) 0.4 (0.02–6.4)
0.8 0.5
Systemic and pulmonary veins
10/11 (91)
51/74 (69)
4.5 (0.5–37.3)
0.1
Septation defects
12/18 (67)
49/67 (73)
0.7 (0.1–2.2)
0.6
55/81 (68)
3.4 (0.5–Inf)
0.2
Valves
123
information that altered the surgical plan for more than half of the subjects. For intubated infants, CMRI was safe, with few and largely minor complications that allowed completion of the imaging study. Other studies have shown that CMRI provides visualization of structures not well seen by echocardiography but does not evaluate the impact on the surgical plan [7]. Studies investigating the usefulness of CMRI for infants have been limited largely to children with heterotaxy syndrome. Niwa et al. [5] found that both cardiac and extracardiac anatomy could be well defined by CMRI for patients with heterotaxy syndrome, but the importance of CMRI for surgical management was not evaluated. In contrast, Geva et al. [4] prospectively evaluated 14 patients with heterotaxy syndrome using both echocardiography and CMRI and reported the management plan before and after CMRI. The surgical plan changed for 28 % of the subjects based on CMRI findings. This estimate of surgical management effect is similar to our findings in the subpopulation of subjects with heterotaxy. In addition to providing valuable information for surgical planning, CMRI was safe for the intubated infants with CHD. The complication rate in this study that intubated all infants was 5 %, similar to previously published reports of 4 to 5 % [3]. This complication rate is notably lower than that reported for cardiac angiography (8–24 %) in similarly sized infants [8, 12]. An added advantage of CMRI is avoidance of exposure to ionizing radiation. Our QA data further support the utility of CMRI in the preoperative setting for young children with CHD. Similar to our study, previous reports have described the utility of CMRI for evaluating the aortic arch, pulmonary arteries, and systemic and pulmonary veins [2, 7, 10]. The current CMRI technology does not add significant additional information in the evaluation of coronary arteries or the details of the valve components. Many centers use computed tomography (CT) for evaluation of anatomic cardiac and extracardiac anomalies in this age group. Compared
4/4 (100)
Pediatr Cardiol
with CMRI, cardiac CT is faster and better at evaluating coronary arteries, albeit with the addition of ionizing radiation risk. At our institution, CMRI is the preferred imaging method for evaluation of CHD, and cardiac CT is performed sparingly, especially for evaluation of coronary artery anomalies, and comparison of CMRI and cardiac CT could not be performed.
Study Limitations This study was limited by its retrospective design. We were unable to determine specific cardiac diagnoses that would indicate the need for routine preoperative CMRI because of our small sample size. We also could not determine in retrospect whether a more detailed performance or careful interpretation of echocardiography should have been able to identify the cardiac findings noted by CMRI. Unrecognized confounding cannot be excluded. However, because our findings did not show statistical significance by indication category, it is most likely that we did not have enough power to detect a difference between the groups.
Conclusions Preoperative CMRI after echocardiography resulted in new information that altered the surgical plan for more than half of the young infants. In addition, CMRI could have identified more than 50 % of the significant intraoperative findings not detected using preoperative echocardiography. For intubated infants, CMRI was safe, with few and largely minor complications. Further studies to determine standard preoperative criteria for the use of CMRI in infants with CHD may help to define appropriate cost-effective use of this diagnostic method. Acknowledgment This investigation was supported by the University of Utah Study Design and Biostatistics Center, with funding in part from the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health, through grant 8UL1TR000105 (formerly UL1RR025764).
References 1. Beier UH, Jelnin V, Jain S, Ruiz CE (2006) Cardiac computed tomography compared to transthoracic echocardiography in the management of congenital heart disease. Catheter Cardiovasc Interv 68:441–449 2. Chung KJ, Simpson IA, Newman R, Sahn DJ, Sherman FS, Hesselink JR (1988) Cine magnetic resonance imaging for evaluation of congenital heart disease: role in pediatric cardiology compared with echocardiography and angiography. J Pediatr 113:1028–1035 3. Fogel MA, Weinberg PM, Parave E, Harris C, Montenegro L, Harris MA, Concepcion M (2008) Deep sedation for cardiac magnetic resonance imaging: a comparison with cardiac anesthesia. J Pediatr 152:534–539, 539 e1 4. Geva T, Vick GW III, Wendt RE, Rokey R (1994) Role of spin echo and cine magnetic resonance imaging in presurgical planning of heterotaxy syndrome: comparison with echocardiography and catheterization. Circulation 90:348–356 5. Niwa K, Uchishiba M, Aotsuka H, Tateno S, Tashima K, Fujiwara T, Matsuo K (1994) Magnetic resonance imaging of heterotaxia in infants. J Am Coll Cardiol 23:177–183 6. O’Leary PW, Hagler DJ, Seward JB, Tajik AJ, Schaff HV, Puga FJ, Danielson GK (1995) Biplane intraoperative transesophageal echocardiography in congenital heart disease. Mayo Clin Proc 70:317–326 7. Prakash A, Torres AJ, Printz BF, Prince MR, Nielsen JC (2007) Usefulness of magnetic resonance angiography in the evaluation of complex congenital heart disease in newborns and infants. Am J Cardiol 100:715–721 8. Rhodes JF, Asnes JD, Blaufox AD, Sommer RJ (2000) Impact of low body weight on frequency of pediatric cardiac catheterization complications. Am J Cardiol 86(1275–1278):A9 9. Sharma S, Anand R, Kanter KR, Williams WH, Dooley KJ, Jones DW, Vincent RN (1992) The usefulness of echocardiography in the surgical management of infants with congenital heart disease. Clin Cardiol 15:891–897 10. Simpson IA, Chung KJ, Glass RF, Sahn DJ, Sherman FS, Hesselink J (1988) Cine magnetic resonance imaging for evaluation of anatomy and flow relations in infants and children with coarctation of the aorta. Circulation 78:142–148 11. Tsai-Goodman B, Geva T, Odegard KC, Sena LM, Powell AJ (2004) Clinical role, accuracy, and technical aspects of cardiovascular magnetic resonance imaging in infants. Am J Cardiol 94:69–74 12. Vitiello R, McCrindle BW, Nykanen D, Freedom RM, Benson LN (1998) Complications associated with pediatric cardiac catheterization. J Am Coll Cardiol 32:1433–1440
123
ORIGINAL ARTICLE
Yield of Cardiac Magnetic Resonance Imaging as an Adjunct to Echocardiography in Young Infants With Congenital Heart Disease Joyce T. Johnson • Kimberly M. Molina Molly McFadden • L. LuAnn Minich • Shaji C. Menon
•
Received: 18 November 2013 / Accepted: 25 March 2014 Ó Springer Science+Business Media New York 2014
Abstract Echocardiography provides adequate preoperative imaging for most young infants with congenital heart disease (CHD). When anatomic details require further clarification, cardiac magnetic resonance imaging (CMRI) may be useful but adds the risks of sedation or general anesthesia for a vulnerable population. This study aimed to determine the safety of CMRI and its yield of additional significant information for this population. The study identified all infants age 90 days or younger with preoperative echocardiography and a CMRI from the period 2002–2012. Indications, complications, and imaging results were collected. The additional CMRI information was defined as ‘‘significant’’ if it altered surgical management or ‘‘not significant’’ if it did not. Associations between indications for CMRI and the likelihood of new significant findings were sought. For 137 infants (58 % male), CMRI was performed at a median age of 5 days (range, 0–89 days). The CMRI yielded additional information for 76 % (104/137) of the patients. The additional findings were significant for 69 % (72/104) of these patients. The incidence of significant new findings was similar among indication categories. All the infants were intubated. Complications occurred for 5 % of the patients, including one subject with a bradycardic event that prevented completion of the exam and six patients with transient vital sign changes that allowed exam completion. More than 50 % of young infants with CHD who
J. T. Johnson (&) K. M. Molina L. L. Minich S. C. Menon Division of Cardiology, Department of Pediatrics, University of Utah and Primary Children’s Hospital, 100 N. Mario Capecchi Drive, Salt Lake City, UT 84113, USA e-mail: [email protected]; [email protected] M. McFadden Division of Epidemiology, Department of Medicine, University of Utah and Primary Children’s Hospital, Salt Lake City, UT, USA
underwent preoperative CMRI had new findings affecting surgical management. Among these patients, CMRI-associated complications were few and predominantly minor for intubated infants. Further studies to determine standard preoperative criteria for the use of CMRI for infants with CHD may help to define appropriate cost-effective use of this diagnostic method. Keywords Surgical planning Cardiac magnetic resonance imaging Congenital heart disease Infant
Introduction Echocardiography is the imaging method of choice for defining the preoperative anatomy and physiology of infants with congenital heart disease (CHD) [9]. For cases in which echocardiography fails to provide sufficient information, cardiac catheterization has been the traditional alternative approach despite the associated risks and exposure to ionizing radiation [8, 12]. Cardiac magnetic resonance imaging (CMRI) is an alternative noninvasive imaging method that avoids these risks [11]. Although acoustic windows in infants usually allow an excellent echocardiographic definition of cardiac anatomy and function, this technology has limitations. Noncardiac chest anomalies, lung disease, and extracardiac lesions may interfere with echocardiographic imaging. The yield of additional preoperative information that affects surgical management, the potential to identify anatomy missed by echocardiography, and the complications associated with CMRI have not been studied systematically. Previous studies have evaluated the yield of new imaging methods and their effect on surgical management and have provided guidance for their incorporation into practice [1, 6]. We used this methodology with the aim of determining the
123
Pediatr Cardiol
added clinical benefit and risks of CMRI for preoperative planning for young infants with CHD. We hypothesized that CMRI for preoperative evaluation of infants with CHD will safely provide additional information that affects surgical planning. In addition, we aimed to evaluate a group of patients without preoperative CMRI but with findings by intraoperative assessment that differed from echocardiographic findings to assess further the ideal indications for CMRI.
Methods Study Population In this retrospective cohort study performed to evaluate the utility of CMRI in young infants with CHD, we assessed two groups. The CMRI group was identified by searching the Pediatric CMRI Database for the period 2002–2012 to identify infants age 90 days or younger with CHD who had preoperative echocardiography and CMRI. Infants who had undergone cardiac surgery before CMRI were excluded from the study. The CMRI group allowed us to determine the utility of CMRI for patients who had the exam. We used a quality assurance (QA) group to determine whether surgically relevant information not obtained by preoperative echocardiography could have been identified by CMRI had it occurred. We searched the Pediatric Echocardiography Quality Assurance Database for all existing years (2005–2012) to identify surgically relevant discrepancies between preoperative echocardiographic diagnoses and intraoperative assessment at the first cardiac surgery. Those who had a CMRI before surgery and children older than 4 years at the time of surgery were excluded from the QA group because all surgically relevant information would likely have manifested by the age of 4 years. The QA group had no patients who also were in the CMRI group. The study was approved by the Primary Children’s Hospital Privacy Board and the University of Utah Institutional Review Board under a waiver of consent. Medical Record Data Collection The medical records of both groups were reviewed for demographic information, diagnoses, date and type of surgical procedures and imaging studies, and surgical descriptions of the anatomy. Complications and indications for intubation in the CMRI group were recorded. Imaging Data Collection For the CMRI group, the echocardiographic and CMRI images and reports from both imaging studies were reviewed for echocardiographic diagnoses, indications for
123
CMRI, CMRI diagnoses, details of anatomy, and function/ flow analysis. All studies were performed using a GE Medical Systems (Milwaukee, WI, USA) 1.5-T Signa LX Echospeed MRI scanner. Torso or cardiac-phased array coils were chosen according to the patient’s body size. For patients under general anesthesia, breathholding was performed to limit motion artifact. No changes were made in the scanner or sequences used at our center during the study period. The common imaging sequences used for structural and functional evaluation of structural heart disease in this cohort included a three-plane localizer; double- and tripleinversion recovery fast spin echo (FSE); electrocardiogram-triggered, segmented k-space, fast-spoiled, gradient-recalled cine sequences (SSFP); free-breathing cine phase contrast flow measurements (40 phases per cardiac cycle) with retrospective cardiac gating; multi-slab 3D whole heart SSFP T2 prep navigator sequence; and inversion recovery-prepared gradient-echo sequence 10 to 20 min after an intravenous bolus of 0.2-mmol/kg gadolinium-based contrast. The images were transferred to a work station for postprocessing (MASS; Medis, Leiden, The Netherlands). The subjects were grouped by indication into the following categories based on the anatomic area involved: pulmonary arteries, pulmonary and systemic veins, aortic arch, septation defects, coronary artery anatomy, valves, or other (i.e., masses, functional and flow analysis). Any additional information provided by CMRI and not available from the echocardiogram was dichotomized as ‘‘significant’’ or ‘‘not significant’’ using previously published methods for determining the yield of a new imaging technique [6], in which significant findings were defined as new CMRI information that affected the surgical management plan or would have been life threatening if unknown. New findings that did not affect surgical management were defined as not significant. The significance of the additional findings obtained from CMRI was independently categorized by two investigators (K.M., J.T.J.) and adjudicated by a third investigator (L.M.) as needed. For the QA group, new information described at surgery and not obtained by preoperative echocardiography was defined as ‘‘significant’’ or ‘‘not significant’’ using the same definitions. The feasibility of CMRI to detect the clinically significant echocardiographic inaccuracy was independently determined by two investigators experienced in the use of CMRI. The echocardiographic findings and the CMRI findings were collected via chart review and compared with the findings listed in the operative report. Because the goal for the ‘‘significant’’ or ‘‘not significant’’ categorization was consensus, intra- and interobserver reliability was not addressed.
Pediatr Cardiol
Statistical Analysis For the analysis, indication categories for the CMRI and QA groups were chosen as predictors of the outcome for each group. The outcome was defined as the incidence of significant new information provided by either CMRI or intraoperative assessment that was not obtained by preoperative echocardiography. Continuous data were expressed as means ± standard deviations or medians with ranges, as appropriate, and categorical data were tabulated. For the CMRI group, univariate logistic regression was used to assess the association between each indication category for CMRI and the odds of significant new findings compared with the remaining subjects. Likewise, the QA group was evaluated using univariate logistic regression for each indication category and the significance of new information provided by intraoperative assessment that could have been diagnosed by CMRI. Because confounding of significance by anything other than the indication category was unlikely and the results by univariate analysis were not significant, a
Table 1 Group characteristics CMRI group (n = 137) n (%)
QA group (n = 85) n (%)
Males
79 (58)
38 (45)
Median age at CMRI: days (range)
5 (0–89)
NA
Mean weight at CMRI (kg) Indication categories
3.16 ± 0.73
NA
Aortic arches
59 (43)
8 (9)
Pulmonary artery anatomies
41 (30)
2 (2)
Pulmonary or systemic vein anatomies
25 (18)
10 (12)
Septation defects
0 (0)
26 (31)
Coronary arteries
0 (0)
15 (18)
Valve abnormalities
0 (0)
22 (26)
Others
11 (9)
0 (0)
19 (14)
2 (2)
Heterotaxy syndromes
CMRI, cardiac magnetic resonance imaging; QA, quality assurance; NA, not applicable
Table 2 Odds of significant new cardiac magnetic resonance imaging (CMRI) information by univariate logistic regression
multivariate analysis was not performed. A two-sided p value lower than 0.05 was considered statistically significant. The statistical analyses were performed using Stata release 12 (StataCorp. 2011, Stata Statistical Software: Release 12; College Station, TX, USA).
Results CMRI Group The CMRI group consisted of 137 subjects with a median age of 5 days (range, 0–89 days) (Table 1). In this study, CMRI was performed after one echocardiogram for 39 % (53/137) of the subjects, after two echocardiograms for 35 % (48/137) of the subjects, and after three or more echocardiograms for 26 % (36/137) of the subjects. New CMRI information was noted for 76 % (104/137) of the CMRI group, with 69 % (72/104) of these new findings significant and altering the surgical plan for 53 % (72/137) of the group with significant findings. Of the 19 subjects with heterotaxy, CMRI yielded significant findings for 37 % (7/19). The univariate logistic regression analysis showed no difference between the odds of significant new information from CMRI for the indication categories (Table 2). For each category, the proportion of those with the specified indication who had a significant finding was compared with the proportion of those without the specified indication who had a significant finding. All the subjects were sedated and intubated for CMRI, with 66 % (90/137) undergoing intubation solely for the purpose of obtaining the CMRI. Complications occurred for 5 % (7/137) of the subjects, including one subject who had a bradycardic event treated with chest compressions and epinephrine with abortion of the CMRI. The remaining six subjects had transient vital sign changes responding to fluid bolus or ventilation changes, with successful completion of the CMRI. QA Group The diagnosis for 51 % (43/85) of the 85 subjects with significant new information from intraoperative assessment
Indication category
Proportion of those with specified indication with significant finding n (%)
Aortic arches
29/59 (49)
Pulmonary arteries
26/41 (63)
Systemic and pulmonary veins Heterotaxy syndromes
Proportion of those without specified indication with significant finding n (%)
Odds ratio (95 % CI)
p Value
43/78 (55)
0.8 (0.4–1.5)
0.5
46/96 (48)
1.9 (0.9–4.0)
0.1
10/25 (40)
62/112 (55)
0.5 (0.2–1.3)
0.2
7/19 (37)
65/118 (55)
0.5 (0.2–1.3)
0.1
123
Pediatr Cardiol Table 3 Description of anatomic categories and proportion of findings that could have been diagnosed by cardiac magnetic resonance imaging (CMRI) had they occurred before surgery Indication category (n)
CMRI could have determined diagnosis n (%)
Aortic arches (8)
8 (100)
Heterotaxy syndromes (2)
2 (100)
Systemic and pulmonary veins (10)
9 (90)
Septation defects (26)
18 (69)
Pulmonary arteries (2)
1 (50)
Valves (22)
4 (15)
Coronaries (15)
0 (0)
could have been determined by a preoperative CMRI. The characteristics of the QA group are listed in Table 1. Both subjects with heterotaxy had significant new information obtained from intraoperative assessment that could have been diagnosed by CMRI. In contrast, only 15 % of the valvar findings could have been diagnosed by CMRI and were limited to distinguishing supravalvar from valvar stenosis. None of the coronary artery anomalies detected by intraoperative assessment were likely to be diagnosed with current CMRI techniques in the young infants (Table 3). The odds that CMRI could have identified significant new anatomic information from intraoperative assessment was similar among the indication categories (Table 4). For each category, the proportion of those who had the specified indication with a significant finding was compared with the proportion of those without the specified indication who had a significant finding.
Discussion Our study demonstrated that CMRI in young infants with CHD is useful in surgical planning. After preoperative echocardiography, CMRI provided additional anatomic
Table 4 Odds of significant new intraoperative information that could have been observed by CMRI using univariate logistic regression
Indication category
Proportion of those with specified indication with significant finding n (%)
Proportion of those without specified indication with significant finding n (%)
Odds ratio (95 % CI)
p Value
Aortic arches Pulmonary arteries
6/8 (75) 1/2 (50)
55/77 (71) 60/83 (72)
1.2 (0.2–6.4) 0.4 (0.02–6.4)
0.8 0.5
Systemic and pulmonary veins
10/11 (91)
51/74 (69)
4.5 (0.5–37.3)
0.1
Septation defects
12/18 (67)
49/67 (73)
0.7 (0.1–2.2)
0.6
55/81 (68)
3.4 (0.5–Inf)
0.2
Valves
123
information that altered the surgical plan for more than half of the subjects. For intubated infants, CMRI was safe, with few and largely minor complications that allowed completion of the imaging study. Other studies have shown that CMRI provides visualization of structures not well seen by echocardiography but does not evaluate the impact on the surgical plan [7]. Studies investigating the usefulness of CMRI for infants have been limited largely to children with heterotaxy syndrome. Niwa et al. [5] found that both cardiac and extracardiac anatomy could be well defined by CMRI for patients with heterotaxy syndrome, but the importance of CMRI for surgical management was not evaluated. In contrast, Geva et al. [4] prospectively evaluated 14 patients with heterotaxy syndrome using both echocardiography and CMRI and reported the management plan before and after CMRI. The surgical plan changed for 28 % of the subjects based on CMRI findings. This estimate of surgical management effect is similar to our findings in the subpopulation of subjects with heterotaxy. In addition to providing valuable information for surgical planning, CMRI was safe for the intubated infants with CHD. The complication rate in this study that intubated all infants was 5 %, similar to previously published reports of 4 to 5 % [3]. This complication rate is notably lower than that reported for cardiac angiography (8–24 %) in similarly sized infants [8, 12]. An added advantage of CMRI is avoidance of exposure to ionizing radiation. Our QA data further support the utility of CMRI in the preoperative setting for young children with CHD. Similar to our study, previous reports have described the utility of CMRI for evaluating the aortic arch, pulmonary arteries, and systemic and pulmonary veins [2, 7, 10]. The current CMRI technology does not add significant additional information in the evaluation of coronary arteries or the details of the valve components. Many centers use computed tomography (CT) for evaluation of anatomic cardiac and extracardiac anomalies in this age group. Compared
4/4 (100)
Pediatr Cardiol
with CMRI, cardiac CT is faster and better at evaluating coronary arteries, albeit with the addition of ionizing radiation risk. At our institution, CMRI is the preferred imaging method for evaluation of CHD, and cardiac CT is performed sparingly, especially for evaluation of coronary artery anomalies, and comparison of CMRI and cardiac CT could not be performed.
Study Limitations This study was limited by its retrospective design. We were unable to determine specific cardiac diagnoses that would indicate the need for routine preoperative CMRI because of our small sample size. We also could not determine in retrospect whether a more detailed performance or careful interpretation of echocardiography should have been able to identify the cardiac findings noted by CMRI. Unrecognized confounding cannot be excluded. However, because our findings did not show statistical significance by indication category, it is most likely that we did not have enough power to detect a difference between the groups.
Conclusions Preoperative CMRI after echocardiography resulted in new information that altered the surgical plan for more than half of the young infants. In addition, CMRI could have identified more than 50 % of the significant intraoperative findings not detected using preoperative echocardiography. For intubated infants, CMRI was safe, with few and largely minor complications. Further studies to determine standard preoperative criteria for the use of CMRI in infants with CHD may help to define appropriate cost-effective use of this diagnostic method. Acknowledgment This investigation was supported by the University of Utah Study Design and Biostatistics Center, with funding in part from the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health, through grant 8UL1TR000105 (formerly UL1RR025764).
References 1. Beier UH, Jelnin V, Jain S, Ruiz CE (2006) Cardiac computed tomography compared to transthoracic echocardiography in the management of congenital heart disease. Catheter Cardiovasc Interv 68:441–449 2. Chung KJ, Simpson IA, Newman R, Sahn DJ, Sherman FS, Hesselink JR (1988) Cine magnetic resonance imaging for evaluation of congenital heart disease: role in pediatric cardiology compared with echocardiography and angiography. J Pediatr 113:1028–1035 3. Fogel MA, Weinberg PM, Parave E, Harris C, Montenegro L, Harris MA, Concepcion M (2008) Deep sedation for cardiac magnetic resonance imaging: a comparison with cardiac anesthesia. J Pediatr 152:534–539, 539 e1 4. Geva T, Vick GW III, Wendt RE, Rokey R (1994) Role of spin echo and cine magnetic resonance imaging in presurgical planning of heterotaxy syndrome: comparison with echocardiography and catheterization. Circulation 90:348–356 5. Niwa K, Uchishiba M, Aotsuka H, Tateno S, Tashima K, Fujiwara T, Matsuo K (1994) Magnetic resonance imaging of heterotaxia in infants. J Am Coll Cardiol 23:177–183 6. O’Leary PW, Hagler DJ, Seward JB, Tajik AJ, Schaff HV, Puga FJ, Danielson GK (1995) Biplane intraoperative transesophageal echocardiography in congenital heart disease. Mayo Clin Proc 70:317–326 7. Prakash A, Torres AJ, Printz BF, Prince MR, Nielsen JC (2007) Usefulness of magnetic resonance angiography in the evaluation of complex congenital heart disease in newborns and infants. Am J Cardiol 100:715–721 8. Rhodes JF, Asnes JD, Blaufox AD, Sommer RJ (2000) Impact of low body weight on frequency of pediatric cardiac catheterization complications. Am J Cardiol 86(1275–1278):A9 9. Sharma S, Anand R, Kanter KR, Williams WH, Dooley KJ, Jones DW, Vincent RN (1992) The usefulness of echocardiography in the surgical management of infants with congenital heart disease. Clin Cardiol 15:891–897 10. Simpson IA, Chung KJ, Glass RF, Sahn DJ, Sherman FS, Hesselink J (1988) Cine magnetic resonance imaging for evaluation of anatomy and flow relations in infants and children with coarctation of the aorta. Circulation 78:142–148 11. Tsai-Goodman B, Geva T, Odegard KC, Sena LM, Powell AJ (2004) Clinical role, accuracy, and technical aspects of cardiovascular magnetic resonance imaging in infants. Am J Cardiol 94:69–74 12. Vitiello R, McCrindle BW, Nykanen D, Freedom RM, Benson LN (1998) Complications associated with pediatric cardiac catheterization. J Am Coll Cardiol 32:1433–1440
123
