Repaired Tetralogy of Fallot: Ratio of Right Ventricular Volume to Left Ventricular Volume as a Marker of Right Ventricular Dilatation 1

Transcription

1 Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at Original Research n Cardiac Imaging Mateusz piewak, MD, PhD Łukasz A. Małek, MD, PhD Joaa Petryka, MD, PhD Łukasz Mazurkiewicz, MD, PhD Konrad Werys, MS El bieta K. Biernacka, MD, PhD Mirosław Kowalski, MD, PhD Piotr Hoffman, MD, PhD Marcin Demkow, MD, PhD Jolanta Mi ko, MD, PhD Witold Ru yłło, MD, PhD Repaired Tetralogy of Fallot: Ratio of Right Ventricular Volume to Left Ventricular Volume as a Marker of Right Ventricular Dilatation 1 Purpose: Materials and Methods: To compare indexed right ventricular (RV) end-diastolic volume (RVEDVi) and the ratio of RV volume to left ventricular (LV) volume (RV/LV ratio) in prediction of significant pulmonary regurgitation (PR) after tetralogy of Fallot (TOF) repair and to assess sex differences in the RV/LV ratio. The ethics committee approved this retrospective singlecenter study, and patients or their parents or guardians signed written informed consent. RVEDVi, RV/LV ratio, and PR were measured with the use of magnetic resonance imaging in 155 consecutive patients with repaired TOF (mean age, 29.2 years [standard deviation]; 98 [63.2%] male and 57 [36.8%] female patients). PR fraction of 20% or greater was considered significant. The capability of the RVEDVi and that of the RV/LV ratio for prediction of significant PR were compared by using logistic regression analysis and receiver operating characteristic curve analysis. Results: RVEDVi was significantly higher in male (162.8 ml/m ) than in female (138.2 ml/m ) patients (P =.001). Conversely, the RV/LV ratio was similar in both sexes ( [male] vs [female], P =.13) both in the entire cohort and after excluding patients with significant ( 30 mm Hg) RV outflow tract gradient and/ or other residual hemodynamic abnormalities (P =.63). Receiver operating characteristic analysis revealed better discrimination of significant ( 20%) from insignificant (, 20%) PR with the use of the RV/LV ratio than with RVEDVi (area under the receiver operating characteristic curve, [model 4] vs [model 1], P =.01). In multivariate analysis, the only independent predictor of PR fraction was the RV/LV ratio. 1 From the Departments of Coronary Artery Disease and Structural Heart Diseases (M.Ś., J.P.), Interventional Cardiology and Angiology (Ł.A.M.), Cardiomyopathy (Ł.M.), and Radiology (K.W., J.M.), Cardiac Magnetic Resonance Unit, Institute of Cardiology, ul Alpejska 42, Warsaw, Poland; and Departments of Congenital Heart Diseases (E.K.B., M.K., P.H.) and Coronary Artery Diseases (M.D.), Institute of Cardiology, Warsaw, Poland; and Institute of Cardiology, Warsaw, Poland (W.R.). Received January 18, 2012; revision requested March 1; final revision received March 7; accepted March 14; final version accepted March 22. Supported by the Foundation for Polish Science. Address correspondence to M.Ś. ( Conclusion: The RV/LV ratio is more accurate than the RVEDVi in differentiation of significant from insignificant PR. After TOF repair, female and male patients have similar RV/LV ratios despite significant differences in RVEDVi between the sexes. q RSNA, 2012 Supplemental material: /suppl/doi: /radiol /-/dc1 q RSNA, radiology.rsna.org n Radiology: Volume 265: Number 1 October 2012

2 Pulmonary regurgitation (PR) is recognized as the most common complication in patients with repaired tetralogy of Fallot (TOF) (1 3). It is associated with progressive right ventricular (RV) dilatation and impaired RV systolic function (1 6). Moreover, researchers in previous studies (7 11) have demonstrated that PR affects not only the RV, but also the left ventricle (LV). In recent years, cardiac magnetic resonance (MR) imaging has emerged as a reliable tool for evaluating patients with repaired TOF (1 3). It enables an accurate and reproducible volumetric analysis of both ventricles and accurate assessment of interventricular interaction, providing information on changes in ventricular size and function after pulmonary valve replacement (PVR) and is considered as the reference Advances in Knowledge The ratio of right ventricular (RV) volume to left ventricular (LV) volume (RV/LV ratio) was found to be superior to the indexed RV end-diastolic volume (RVEDVi) in discriminating significant ( 20%) from insignificant (,20%) pulmonary regurgitation (PR) in patients with repaired tetralogy of Fallot (TOF). With the RV/LV ratio, moderate PR was differentiated from severe PR, whereas with RVEDVi, they were not differentiated. Adding the indexed LV end-diastolic volume (LVEDVi) to RVEDVi was of incremental value in predicting the severity of PR, and incorporating RVEDVi and LVEDVi into one parameter, namely the RV/LV ratio, provided a discriminatory power similar to and a better model fit than RVEDVi plus LVEDVi. After TOF repair, women and men had a similar RV/LV ratio despite significant differences in RVEDVi between the sexes. standard for quantitative assessment of PR (1 13). PVR timing remains a subject of debate (1 6,14,15). The criteria for PVR include RV dilatation as one of the factors in the decision about the need for reintervention. There is, however, no consensus in regard to the optimal cutoff point for RV enlargement for deciding who should undergo surgery for PVR. Furthermore, while most of the researchers use RV volume indexed for body surface area as a marker of RV dilatation, others use the ratio of the RV volume to LV volume (RV/LV ratio) (3 6,14 16). The ratio incorporates in one parameter the deleterious effects of PR on the RV (dilatation) and the LV (compression). Therefore, it can be hypothesized that increasing severity of PR will be reflected by a higher RV/LV ratio. Moreover, it has been recently shown that the reference values for indexed RV end-diastolic volume (RVED- Vi) in patients with repaired TOF are different for male and female patients, making unisex threshold levels for PVR unsuitable in clinical practice (17). Accordingly, the purpose of our study was to compare RVEDVi and the RV/LV ratio in the prediction of significant PR after TOF repair and to assess sex differences in the RV/LV ratio. Materials and Methods Study Population Approval of the local ethics committee was obtained. Each patient and/or parents or guardians gave written informed consent for the cardiac MR study. The study was designed and conducted according to the Standards for Reporting of Diagnostic Accuracy (18). Implication for Patient Care Clinical decision making on the basis of the RV/LV ratio in patients with repaired TOF may be better suited to assess the severity of PR and its consequences; this result should be confirmed or refuted in further prospective studies. This retrospective single-center study included 10 healthy volunteers (five [50%] male and five [50%] female subjects; mean age, 26.7 years 6 4.1[standard deviation]) and 155 consecutive patients (98 [63.2%] male and 57 [36.8%] female patients; mean age, 29.2 years ) who had undergone TOF repair and had undergone a cardiac MR study between June 2008 and mid-september To achieve a homogeneous patient population, individuals with pulmonary atresia and ventricular septal defect were excluded. In patients who underwent more than one cardiac MR study, as part of scheduled follow-up assessment, only the index study was included in the analysis. Selected data from smaller study samples have been analyzed previously (19,20). To avoid altering the RV/LV ratio by conditions other than PR, we created a highly selected subset of patients (subgroup 1) by excluding a patient when at least one of the following conditions was present: significant valvular incompetence for other valves than the Published online before print /radiol Content code: Radiology 2012; 265:78 86 Abbreviations: AUC = area under the receiver operating characteristic curve LV = left ventricle LVEDVi = indexed LV end-diastolic volume PR = pulmonary regurgitation PRF = PR fraction PVR = pulmonary valve replacement RV = right ventricle RVEDVi = indexed RV end-diastolic volume RV/LV ratio = ratio of RV volume to LV volume RVOT = RV outflow tract TOF = tetralogy of Fallot Author contributions: Guarantors of integrity of entire study, M.., J.M.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, M..; clinical studies, M.., Ł.A.M. J.P., Ł.M., K.W., E.K.B., M.K., P.H.; statistical analysis, M.Ś.; and manuscript editing, M.., Ł.A.M., J.M., W.R. Potential conflicts of interest are listed at the end of this article. Radiology: Volume 265: Number 1 October 2012 n radiology.rsna.org 79

3 pulmonary valve (eg, tricuspid regurgitation), residual ventricular septal defect, or a significant RV outflow tract (RVOT) obstruction, thereby avoiding the confounding effects of these lesions (1 3,20). Analyses on diagnostic accuracy of the RV/LV ratio in the prediction of PR severity grade were restricted to subgroup 1. Cardiac MR Imaging Cardiac MR imaging studies were performed with the use of a 1.5-T imager (Avanto; Siemens, Erlangen, Germany). Balanced steady-state free precession breath-hold cine images were acquired in the LV long axis and short axis, covering both ventricles from the base to the apex, with typical imaging parameters as follows: effective repetition time msec/echo time msec, 33 54/1.2; flip angle, ; section thickness, 8 mm; gap, 1.6 mm; in-plane image resolution, mm to mm; and temporal resolution, phases per cardiac cycle. Pulmonary flow was calculated from retrospectively gated free-breathing phase-contrast images, with typical imaging parameters as follows: 30 47/2; flip angle, 30 ; and section thickness, 5 mm. The imaging plane for a flow-sensitive gradient-echo sequence was prescribed perpendicularly to the main pulmonary artery or conduit and located at the midpoint of the blood vessel. Velocity encoding was set at cm/sec to avoid an aliasing artifact. Image Analysis RV end-diastolic volume and LV end-diastolic volume were calculated with the use of dedicated software (Mass 6.2.1; Medis, Leiden, the Netherlands) on the basis of manual tracing of endocardial borders at end diastole. Similarly, manual delineation of endocardial borders at end systole resulted in LV and RV end-systolic volumes. Finally, LV and RV ejection fractions and masses were calculated. All volume and mass parameters were indexed for body surface area. Considering numerous trabeculations in the RV, and blurred and ambiguous blood-myocardial boundaries because of partial volume effects, which result in worse reproducibility and are more time consuming (12), we decided to include trabeculations and papillary muscles in the blood pool and exclude them from mass calculations. The RV/LV ratio was obtained by dividing RV end-diastolic volume by LV end-diastolic volume. Severe RV dilatation was defined as either RVEDVi exceeding 170 ml/m 2 or the RV/LV ratio being greater than 2.0 (3,14 16,21). For flow calculations, a semiautomatic vessel edge detection algorithm with an operator correction (Argus; Siemens, Erlangen, Germany) was used. PR fraction (PRF) was defined as a percentage calculated by dividing backward flow by forward flow. Phasecontrast velocity mapping served as a reference standard for determining the severity of PR. PR was stratified according to PRF in three grades: mild (PRF, 20%), moderate (PRF = 20% 39%), and severe (PRF 40%) regurgitation (10,13). For the purpose of the statistical analysis, mild PR was considered as insignificant PR, and moderate and severe grades of PR were grouped and considered as significant PR (PRF 20%) (21). Imaging analyses were performed by five independent observers (M.., Ł.A.M., J.P., Ł.M., J.M.) during the regular clinical assignment of each physician to the cardiac MR unit. The readers experience in cardiac MR studies ranged from 1 to 9 years. This approach, with many independent observers instead of consensus interpretation, is considered most suitable in diagnostic performance studies and reflects a real-life clinical scenario better than consensus readings (22,23). As in normal clinical settings, the observers were not blinded to the patient s sex. To assess intraobserver variability of RVEDVi, indexed LV end-diastolic volume (LVEDVi) and the RV/LV ratio measurements, one observer (M.., with 3 years of experience) performed a second reading in a blinded fashion after 3 months or longer from an initial assessment (to minimize recall bias) in a group of 40 randomly selected patients. In addition, in a random sample of 10 patients, measurements were performed by two independent observers (M.., with 3 years of experience, and Ł.A.M., with 4 years of experience), who were blinded to the counterpart s results, and the interobserver variability was calculated. Echocardiography A standard echocardiographic study was performed in all patients. Peak instantaneous RVOT gradient was calculated with the use of the Bernoulli equation, on the basis of the maximal velocity across RVOT determined with the use of continuous-wave Doppler imaging. Significant RVOT obstruction was defined as a peak RVOT gradient of 30 mm Hg or greater (2). In addition, for all patients, details about additional hemodynamic abnormalities (the presence of ventricular septal defect and more than mild regurgitation at the valve other than the pulmonic valve) were recorded. Statistical Analysis All continuous variables were tested for normal distribution by using the Kolmogorov-Smirnov test. Continuous data are presented as means 6 standard deviations or medians with interquartile ranges and were compared with the use of either the Student t test or the Ma-Whitney test, where appropriate. Categorical variables are summarized as absolute numbers and percentages and were compared by using the x 2 test. Correlation was tested by using the Pearson correlation coefficient. Logistic regression was used to determine significant predictors of PR severity grade. First, the predictive value of RVEDVi was determined (model 1). In the next step, we added LVEDVi to the model (model 2). Finally, data on the RV/LV ratio were incorporated (model 3). The final selection of the variable with the best discriminative value was achieved with the use of stepwise analysis (model 4), with a respective P value of less than.05 used for entry and a P value of greater than.1 used for removal from the model. Model fit was assessed by using the Hosmer-Lemeshow goodness-of-fit test 80 radiology.rsna.org n Radiology: Volume 265: Number 1 October 2012

4 (24). The capability of each model for discrimination of significant (PRF 20%) from insignificant (PRF, 20%) PR was estimated by using the area under the receiver operating characteristic curve (AUC). The incremental value was defined as the difference in AUCs. AUCs were compared according to DeLong et al (25). Multivariate prediction of PRF was assessed by using multiple linear regression in a similar fashion as multivariate logistic regression was performed. Receiver operating characteristic curve analysis was performed to determine the best cutoff value of RVEDVi and the RV/LV ratio for prediction of PR severity (26). We calculated conventional measures of the diagnostic accuracy of the best cutoffs (27). The analysis of variance was used to compare the RV/LV ratio and RVEDVi among patients with mild, moderate, and severe PR. The Bonferroni correction was applied when multiple comparisons were performed, with P,.0167 ( ) considered to indicate a significant difference (23,28). For intra- and interobserver variability, the Bland-Altman analysis was used (22). A two-sided P value of less than.05 was considered to denote a significant difference. Statistical analysis was performed with the use of statistical software (MedCalc , MedCalc, Mariakerke, Belgium; SAS 9.2, SAS Institute, Cary, NC). Results Baseline Characteristics There were 211 studies performed in 165 patients with repaired TOF. Ten patients were excluded from the analysis (Fig E1 [online]). The remaining subjects (n = 155) constituted the study group (Fig E1 [online]). Baseline characteristics of the study population are presented in Table E1 (online). There were no significant differences in demographic and clinical parameters between men and women. Table 1 summarizes the findings of the imaging studies. Table 2 demonstrates the reproducibility of the measurements of RVEDVi, LVEDVi, Table 1 Results of Imaging Studies A: Hemodynamics Data in Patients Parameter All Patients (n = 155) Men (n = 98) Women (n = 57) P Value* RVEDVi (ml/m 2 ) RVESVi (ml/m 2 ) RVEF (%) RVMi (g/m 2 ) LVEDVi (ml/m 2 ) LVESVi (ml/m 2 ) LVEF (%) ,.0001 LVMi (g/m 2 ) ,.0001 RV/LV ratio PRF (%) Significant PR (%) 61.9 (96/155) 62.2 (61/98) 61.4 (35/57).95 Peak RVOT gradient (mm Hg) Significant RVOT gradient (%) 32.7 (50/153) 24.5 (24/98) 47.3 (26/55).007 B: Hemodynamics Data in Control Subjects All Healthy Volunteers Parameter (n = 10) Men (n = 5) Women (n = 5) P Value* RVEDVi (ml/m 2 ) LVEDVi (ml/m 2 ) RV/LV ratio Note. Unless otherwise indicated, data are means 6 standard deviations. All parameters but peak RVOT gradient were derived from cardiac MR imaging. Peak RVOT gradient was derived from echocardiographic study on the basis of the maximal velocity across RVOT determined with the use of continuous-wave Doppler imaging. LVEF = LV ejection fraction, LVESVi = indexed LV end-systolic volume, LVMi = indexed LV mass, RVEF = RV ejection fraction, RVESVi = indexed RV end-systolic volume, RVMi = indexed RV mass. * Men versus women. Numbers in parentheses were used to calculate the percentages. Data were available in 153 patients. No differences were observed (P =.9) in a subgroup limited to individuals with peak RVOT gradient less than 30 mm Hg. Table 2 Intra- and Interobserver Variability of Cardiac MR Measurements Variable Intraobserver Variability Mean Difference (%) 6 SD Limits of Agreement (%) and the RV/LV ratio. There was a significant correlation between RVEDVi and the RV/LV ratio (r = 0.72, P,.0001) (Fig 1). PRF did not correlate with RV ejection fraction in the entire group (r = 0.06, P =.29) or in subgroup 1 (r = 20.16, P =.15). Mean Difference (%) 6 SD Interobserver Variability Limits of Agreement (%) RVEDVi , , 23.0 LVEDVi , , 13.9 RV/LV ratio , , 18.8 Note. Mean difference is expressed as the percentage of the average of two measurements. SD = standard deviation. Comparison of the RV/LV Ratio between Male and Female Patients RVEDVi was significantly higher in male patients than in female patients (Table 1). Contrary to that, the RV/LV ratio was similar in both sexes. The differences in the RV/LV ratio remained Radiology: Volume 265: Number 1 October 2012 n radiology.rsna.org 81

5 Figure 1 Table 3 Logistic Regression Models for Prediction of Significant PR (PRF 20%) insignificant after excluding patients with a significant RVOT gradient and/ or other hemodynamic abnormalities (for subgroup 1, [male patients] vs [female patients]; P =.63). Furthermore, when the patients were classified according to quartiles with respect to the RV/LV ratio, a similar proportion of male and female patients was observed in each quartile (P =.72) (Fig E2a [online]). In contrast, in higher quartiles of RVEDVi, Figure 1: Correlation between RVEDVi and RV/LV ratio. Solid line = RV/LV ratio based cutoff for severe RV dilatation (. 2.0). Dashed line = RVEDVi-based cutoff for severe RV dilatation (. 170 ml/m 2 ). Of 51 patients with RVEDVi greater than 170 ml/m 2, there were 18 individuals with RV/LV ratio of 2.0 or less. Among 44 patients with RV/LV ratio greater than 2.0, 11 patients had RVEDVi of 170 ml/m 2 or less. Factors Model 1 Model 2 Model 3 Model 4* Predictors in models RVEDVi 1.04 (1.02, 1.09 (1.04, 0.93 (0.82, ;.0001) 1.13;.0001) 1.04;.23) LVEDVi (0.82, 1.11 (0.93, ;.0008) 1.33;.21) RV/LV ratio (1.20, 15.9;.02) 2.2 (1.52, 3.19;.0001) Model performance AUC (0.750, 0.920) (0.848, 0.975) (0.866, 0.983) (0.859, 0.980) Hosmer-Lemeshow 5.96, 8 (.65) 15.3, 8 (.053) 9.7, 8 (.28) 7.37, 8 (.49) goodness-of-fit test * Stepwise logistic regression analysis of model 3. Data are odds ratios, and the numbers in parentheses are the 95% confidence intervals and P values. For each 0.1 increment in the RV/LV ratio, referring only to the odds ratios of 4.39 and 2.2. Numbers in parentheses are the 95% confidence intervals. Data are the x 2 statistic and degrees of freedom, and numbers in parentheses are P values. A high P value is indicative of a good fit of the data with the model. there was a higher proportion of male patients than female patients (P =.004 for trend) (Fig E2b [online]). In control subjects, RVEDVi and LVEDVi were higher in men than in women (P =.03 and.02, respectively), but the RV/LV ratio was similar in both sexes (P =.54) (Table 1). The RV/LV ratio was greater in patients than in healthy volunteers ( vs , P =.0001). The RV/LV ratio in control subjects was in agreement with the values derived from the normal reference values published previously (29 32). Prediction of the Severity of PR RVEDVi combined with LVEDVi (model 2) had better accuracy than RVEDVi alone (model 1) in the discrimination of significant (PRF 20%) from insignificant (PRF, 20%) PR (P =.01) (Table 3, Fig 2). When the RV/LV ratio was added (model 3), the resulting AUC demonstrated no significant difference in discriminative value compared with the AUC for RVEDVi plus LVEDVi (model 2) (P =.3). The stepwise multivariate logistic regression analysis (model 4) revealed that only the RV/LV ratio was a significant predictor of PR severity grade. The RV/LV ratio alone (model 4) offered better discriminatory performance than did the RVEDVi (model 1) (P =.01) (Table 3, Fig 2). The Hosmer-Lemeshow test showed a P value of.053, which indicated borderline calibration of model 2. Models 1, 3, and 4 fitted well with the observed data (a nonsignificant P value for all the models [Table 3]). Multiple linear regression showed that addition of LVEDVi to RVEDVi improved the prediction of PRF both in subgroup 1 (Table 4) and in the entire study cohort (Table E2 [online]). When the RV/LV ratio was incorporated in a stepwise model, it remained the only independent predictor of PRF. Performance characteristics of RVEDVi and the RV/LV ratio cutoff values for detection of PRF of 20% or greater are presented in Table 5. The scatterplots of RVEDVi and the RV/LV ratio versus three grades of PR severity are demonstrated in Figure 3. There was a significant difference in the RV/LV ratio between patients with moderate versus severe PR (P =.0018), although there was no significant difference in RVEDVi between these groups (P =.09). Discussion There are two main findings in our study. First, we demonstrated the superiority of the RV/LV ratio to the RVEDVi 82 radiology.rsna.org n Radiology: Volume 265: Number 1 October 2012

6 in discriminating significant (PRF 20%) from insignificant (PRF, 20%) PR. Second, we proved that female patients and male patients had similar RV/LV ratios after TOF repair despite significant differences in RVEDVi. These observations may have an important clinical effect. The most common method of expressing RV size in cardiac MR studies is ventricular volume indexed for body surface area (milliliters per square meter). It has been recently shown that female patients with TOF have a lower RVEDVi than do male patients (17). We confirmed these findings but, at the same time, we demonstrated also that the RV/LV ratio was similar in both sexes. Higher RVEDVi in men and similar RV/LV ratios in both sexes indicate that the RV/LV ratio is a useful marker of RV dilatation for male patients and female patients, potentially suitable for clinical decision making irrespective of the patient s sex. While researchers in most studies and guidelines concerning PVR in patients with repaired TOF use RVEDVi as a marker of severe RV dilatation, only a few authors (3 6,14,15) have recommended the use of the RV/LV ratio. As suggested by Sarikouch et al (17), use of a unisex threshold level that is based on RVEDVi results in exposing women to a relatively more severe RV dilatation before the intervention is performed. Indeed, similar mean values of the RV/LV ratio in both sexes together with similar proportion of male patients and female patients in each quartile in the current study indicated that irrespective of the threshold level used, the RV/LV ratio based cutoff values would result in a similar proportion of men and women referred for surgery. Moreover, we showed that RV/LV ratio based and RVEDVi-based criteria for severe RV dilatation were not interchangeable, because there were patients who met one criterion but not the other. Investigators in previous studies demonstrated that LV volume was significantly influenced by PR (11,33). Coincident with RV dilatation, there is a shift of the interventricular septum Figure 2 Figure 2: Graph shows receiver operating characteristic curves of four models in prediction of significant PR (PRF 20%). Model 1 = RVEDVi; model 2 = RVEDVi and LVEDVi; model 3 = RVEDVi, LVEDVi, and RV/LV ratio; and model 4 = RV/LV ratio. P =.01 for model 2 versus model 1, P =.3 for model 3 versus model 2, P =.37 for model 4 versus model 3, P =.01 for model 4 versus model 1. Model 4 resulted in a significantly larger AUC than did model 1, indicating improved discriminatory performance of the RV/LV ratio compared with that of RVEDVi. According to Hosmer and Lemeshow (24), RVEDVi resulted in excellent discrimination (0.8 AUC, 0.9), whereas the RV/LV ratio had outstanding discrimination (AUC 0.9). = Numbers in parentheses are the 95% confidence intervals. Table 4 Linear Regression Models for Prediction of PRF in Subgroup 1 Factors Model 1 Model 2 Model 3 Model 4* Predictors in models Intercept RVEDVi 0.18 (0.03, 0.29 (0.02, 0.06 (0.10,.54)...,.0001, 0.55),.0001) LVEDVi (0.07, (0.19,.47)...,.0001) RV/LV ratio (8.6,.01) 25.7 (2.3,,.0001, 0.79) Model performance R Adjusted R * Stepwise logistic regression analysis of model 3. Data are regression coefficients, and the numbers in parentheses are the standard error and the P value except where otherwise specified. Data are regression coefficients, and the numbers in parentheses are the standard error, the P value, and the Pearson correlation coefficient r. toward the left side, causing changes in LV geometry and decreased LVEDVi. Moreover, an increase in LVEDVi has been observed after successful restoration of pulmonary valve competence with either the surgical or the percutaneous approach (4,6,34). RVEDVi reflects the effect of PR solely on the RV, without taking into consideration detrimental effects of PR on the LV. Use of RVEDVi, therefore, may lead to misinterpretation of the clinical importance of PR, as the effect on the LV is ignored. Radiology: Volume 265: Number 1 October 2012 n radiology.rsna.org 83

7 Figure 3 Figure 3: (a, b) Scatterplots of RV/LV ratio (a) and RVEDVi (b) versus three grades of PR severity at phase-contrast velocity mapping. Dashed lines = receiver operating characteristic based cutoff values for prediction of significant (moderate to severe) PR (PRF 20%). Solid lines = cutoff values for severe RV dilatation. With the RV/LV ratio, moderate PR was differentiated from severe PR (P =.0018), whereas with RVEDVi, they were not differentiated (P =.09). On a, none of the patients with the RV/LV ratio below 1.41 (dashed line) had significant PR. With the RV/LV ratio greater than 2.0 (solid line), all patients with mild (insignificant) PR could be excluded. On b, among patients with RVEDVi below ml/m 2 (dashed line), there were individuals with significant (moderate to severe) PR. Among patients with RVEDVi greater than 170 ml/m 2 (solid line), there were individuals with mild PR. Table 5 Conventional Measures of Diagnostic Accuracy of the RV/LV Ratio and RVEDVi for Detection of Significant PR (PRF 20%) Statistic Data We showed that LVEDVi added incremental value to RVEDVi in the prediction of PR severity. Incorporating RVEDVi and LVEDVi into one parameter, namely the RV/LV ratio, may be more useful clinically. The RV/LV ratio provided a discriminatory power similar RV/LV Ratio* RVEDVi 95% Confidence 95% Confidence Intervals Data Intervals Sensitivity 53/53 (100.0) 93.2, /53 (94.3) 84.4, 98.8 Specificity 19/25 (76.0) 54.9, /25 (64.0) 42.5, 82.0 Positive predictive value 53/59 (89.8) 79.2, /59 (84.7) 73.0, 92.8 Negative predictive value 19/19 (100.0) 82.2, /19 (84.2) 60.4, 96.4 Accuracy 72/78 (92.3) 84.0, /78 (84.6) 74.7, 91.8 * The cutoff was greater than Optimal cutoffs for prediction of significant PR were determined with receiver operating characteristic analysis. The cutoff was greater than ml/m 2. Optimal cutoffs for prediction of significant PR were determined with receiver operating characteristic analysis. Data are the numbers used to calculate the percentages. Numbers in parentheses are percentages. to and a better model fit (calibration) than the combination of RVEDVi plus LVEDVi. According to Hosmer and Lemeshow (24), model performance should be assessed by considering both calibration and discrimination. The P values in model 3 (insignificant for RVEDVi and LVEDVi but significant for the RV/LV ratio) suggest that, while the RV/LV ratio was an independent predictor of PRF of 20% or greater, either of the other terms could be omitted with no significant loss. This was confirmed by the stepwise analysis (model 4). Considering PR as a continuous variable (expressed as PRF), we proved that LVEDVi provided additional information to that provided by RVEDVi in prediction of PRF, as well as that the RV/LV ratio proved to be an independent predictor of PRF. Moreover, we demonstrated that, with the RV/LV ratio, moderate PR was differentiated from severe PR, whereas with RVEDVi, they were not differentiated. It has been shown that percutaneous pulmonary valve implantation leads to the RV/LV ratio reduction (34,35). The ratio, however, remained abnormal despite successful elimination of stenosis and/or regurgitation. This, therefore, led to a conclusion that the use of absolute or indexed RV volumes without considering the effect on LV dynamics may not be the best way of making decisions in the management of chronic PR (34). 84 radiology.rsna.org n Radiology: Volume 265: Number 1 October 2012

8 Further studies are needed in patients with repaired TOF to elucidate whether clinical decisions will be better guided on the basis of the RV/LV ratio rather than on the basis of the RVEDVi only. Because the severity of PR, as demonstrated in our study and in studies by other researchers, does not correlate with RV ejection fraction (36 40), the consequences of the lesion caot be easily evaluated by the increase in RV volume and the decrease in RV ejection fraction. Therefore, there is a constant need for searching for new parameters that reflect the effect of PR on the heart better than RVEDVi and/or RV ejection fraction do. Some authors postulate the use of corrected (ie, corrected for PR measuring the effective forward flow) RV ejection fraction to evaluate true RV performance (13,39,41,42). We believe that the RV/LV ratio may be of clinical utility, and we have provided evidence to support this hypothesis. The authors of prior studies (43 49) demonstrated a high predictive value of the RV/LV ratio (measured on computed tomography scans) in patients with acute pulmonary embolism and proved that the ratio was a simple marker of RV dysfunction. The cardiac MR imaging derived RV/LV ratio may become clinically useful in predicting prognosis and in decision-making about the optimal timing of PVR. Therefore, its potential role in various patient populations, including patients with congenital heart diseases, should be explored further. Our study had several limitations. Our study was retrospective, with all the inherent limitations of the methods used. The findings of our study may call for further, prospective research. Furthermore, it remains unknown whether there is a relationship between the RV/LV ratio and exercise capacity. We examined patients after TOF repair, and the findings of the current study caot be extrapolated to other patient populations. With normal values of the RV/LV ratio, simultaneous severe dilatation of the RV and LV, as for example in patients with dilated cardiomyopathy, is not excluded. Moreover, RV and LV systolic and diastolic functions in relation to the RV/LV ratio should be explored in further studies. In conclusion, the RV/LV ratio is a potential useful marker of cardiac dysfunction in patients after TOF repair. In contrast to the RVEDVi, the RV/LV ratio is independent of sex and is more accurate for stratifying PR in clinically useful grades. Further studies are required to demonstrate the potential clinical effect for guiding treatment options in patients after TOF repair. Disclosures of Potential Conflicts of Interest: M.. No potential conflicts of interest to disclose. Ł.A.M. No potential conflicts of interest to disclose. J.P. No potential conflicts of interest to disclose. Ł.M. No potential conflicts of interest to disclose. K.W. No potential conflicts of interest to disclose. E.K.B. No potential conflicts of interest to disclose. M.K. No potential conflicts of interest to disclose. P.H. No potential conflicts of interest to disclose. M.D. No potential conflicts of interest to disclose. J.M. No potential conflicts of interest to disclose. W.R. No potential conflicts of interest to disclose. References 1. Baumgartner H, Bonhoeffer P, De Groot NM, et al. ESC guidelines for the management of grown-up congenital heart disease (new version 2010). Eur Heart J 2010;31(23): Warnes CA, Williams RG, Bashore TM, et al. ACC/AHA 2008 Guidelines for the Management of Adults with Congenital Heart Disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (writing committee to develop guidelines on the management of adults with congenital heart disease). Circulation 2008;118(23):e714 e Silversides CK, Kiess M, Beauchesne L, et al. Canadian Cardiovascular Society 2009 Consensus Conference on the management of adults with congenital heart disease: outflow tract obstruction, coarctation of the aorta, tetralogy of Fallot, Ebstein anomaly and Marfan s syndrome. Can J Cardiol 2010;26(3):e80 e Oosterhof T, van Straten A, Vliegen HW, et al. Preoperative thresholds for pulmonary valve replacement in patients with corrected tetralogy of Fallot using cardiovascular magnetic resonance. Circulation 2007;116(5): Therrien J, Provost Y, Merchant N, Williams W, Colman J, Webb G. Optimal timing for pulmonary valve replacement in adults after tetralogy of Fallot repair. Am J Cardiol 2005;95(6): Buechel ER, Dave HH, Kellenberger CJ, et al. Remodelling of the right ventricle after early pulmonary valve replacement in chil- dren with repaired tetralogy of Fallot: assessment by cardiovascular magnetic resonance. Eur Heart J 2005;26(24): Niezen RA, Helbing WA, van der Wall EE, van der Geest RJ, Rebergen SA, de Roos A. Biventricular systolic function and mass studied with MR imaging in children with pulmonary regurgitation after repair for tetralogy of Fallot. Radiology 1996;201(1): Davlouros PA, Kilner PJ, Hornung TS, et al. Right ventricular function in adults with repaired tetralogy of Fallot assessed with cardiovascular magnetic resonance imaging: detrimental role of right ventricular outflow aneurysms or akinesia and adverse right-toleft ventricular interaction. J Am Coll Cardiol 2002;40(11): Kempny A, Diller GP, Orwat S, et al. Right ventricular-left ventricular interaction in adults with tetralogy of Fallot: a combined cardiac magnetic resonance and echocardiographic speckle tracking study. Int J Cardiol 2012;154(3): Muzzarelli S, Ordovas KG, Caavale G, Meadows AK, Higgins CB. Tetralogy of Fallot: impact of the excursion of the interventricular septum on left ventricular systolic function and fibrosis after surgical repair. Radiology 2011;259(2): Kuehne T, Saeed M, Gleason K, et al. Effects of pulmonary insufficiency on biventricular function in the developing heart of growing swine. Circulation 2003;108(16): Kilner PJ, Geva T, Kaemmerer H, Trindade PT, Schwitter J, Webb GD. Recommendations for cardiovascular magnetic resonance in adults with congenital heart disease from the respective working groups of the European Society of Cardiology. Eur Heart J 2010;31(7): van Straten A, Vliegen HW, Hazekamp MG, et al. Right ventricular function after pulmonary valve replacement in patients with tetralogy of Fallot. Radiology 2004;233(3): Frigiola A, Tsang V, Bull C, et al. Biventricular response after pulmonary valve replacement for right ventricular outflow tract dysfunction: is age a predictor of outcome? Circulation 2008;118(14 suppl):s182 S Frigiola A, Tsang V, Nordmeyer J, et al. Current approaches to pulmonary regurgitation. Eur J Cardiothorac Surg 2008;34(3): ; discussion van Huysduynen BH, van Straten A, Swee CA, et al. Reduction of QRS duration after pulmonary valve replacement in adult Fallot patients is related to reduction of right ventricular volume. Eur Heart J 2005;26(9): Radiology: Volume 265: Number 1 October 2012 n radiology.rsna.org 85

9 17. Sarikouch S, Koerperich H, Dubowy KO, et al. Impact of gender and age on cardiovascular function late after repair of tetralogy of Fallot: percentiles based on cardiac magnetic resonance. Circ Cardiovasc Imaging 2011;4(6): Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD Initiative. Radiology 2003;226(1): Spiewak M, Biernacka EK, Małek LA, et al. Quantitative assessment of pulmonary regurgitation in patients with and without right ventricular tract obstruction. Eur J Radiol 2011;80(2):e164 e Spiewak M, Biernacka EK, Małek LA, et al. Right ventricular outflow tract obstruction as a confounding factor in the assessment of the impact of pulmonary regurgitation on the right ventricular size and function in patients after repair of tetralogy of Fallot. J Magn Reson Imaging 2011;33(5): Wald RM, Redington AN, Pereira A, et al. Refining the assessment of pulmonary regurgitation in adults after tetralogy of Fallot repair: should we be measuring regurgitant fraction or regurgitant volume? Eur Heart J 2009;30(3): Bankier AA, Levine D, Halpern EF, Kressel HY. Consensus interpretation in imaging research: is there a better way? Radiology 2010;257(1): Levine D, Bankier AA, Halpern EF. Submissions to Radiology: our top 10 list of statistical errors. Radiology 2009;253(2): Hosmer DW, Lemeshow S. Assessing the fit of the model. In: Hosmer DW, Lemeshow S, eds. Applied logistic regression. 2nd ed. New York, NY: Wiley, 2000; DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44(3): Obuchowski NA. Receiver operating characteristic curves and their use in radiology. Radiology 2003;229(1): Langlotz CP. Fundamental measures of diagnostic examination performance: usefulness for clinical decision making and research. Radiology 2003;228(1): Tello R, Crewson PE. Hypothesis testing II: means. Radiology 2003;227(1): Maceira AM, Prasad SK, Khan M, Peell DJ. Reference right ventricular systolic and diastolic function normalized to age, gender and body surface area from steady-state free precession cardiovascular magnetic resonance. Eur Heart J 2006;27(23): Maceira AM, Prasad SK, Khan M, Peell DJ. Normalized left ventricular systolic and diastolic function by steady state free precession cardiovascular magnetic resonance. J Cardiovasc Magn Reson 2006;8(3): Alfakih K, Plein S, Thiele H, Jones T, Ridgway JP, Sivananthan MU. Normal human left and right ventricular dimensions for MRI as assessed by turbo gradient echo and steadystate free precession imaging sequences. J Magn Reson Imaging 2003;17(3): Lorenz CH, Walker ES, Morgan VL, Klein SS, Graham TP Jr. Normal human right and left ventricular mass, systolic function, and gender differences by cine magnetic resonance imaging. J Cardiovasc Magn Reson 1999;1(1): Haddad F, Doyle R, Murphy DJ, Hunt SA. Right ventricular function in cardiovascular disease, part II: pathophysiology, clinical importance, and management of right ventricular failure. Circulation 2008;117(13): Coats L, Khambadkone S, Derrick G, et al. Physiological consequences of percutaneous pulmonary valve implantation: the different behaviour of volume- and pressure-overloaded ventricles. Eur Heart J 2007;28(15): Coats L, Khambadkone S, Derrick G, et al. Physiological and clinical consequences of relief of right ventricular outflow tract obstruction late after repair of congenital heart defects. Circulation 2006;113(17): Geva T, Sandweiss BM, Gauvreau K, Lock JE, Powell AJ. Factors associated with impaired clinical status in long-term survivors of tetralogy of Fallot repair evaluated by magnetic resonance imaging. J Am Coll Cardiol 2004;43(6): Samyn MM, Powell AJ, Garg R, Sena L, Geva T. Range of ventricular dimensions and function by steady-state free precession cine MRI in repaired tetralogy of Fallot: right ventricular outflow tract patch vs. conduit repair. J Magn Reson Imaging 2007;26(4): Meadows J, Powell AJ, Geva T, Dorfman A, Gauvreau K, Rhodes J. Cardiac magnetic resonance imaging correlates of exercise capacity in patients with surgically repaired tetralogy of Fallot. Am J Cardiol 2007;100(9): Henkens IR, van Straten A, Schalij MJ, et al. Predicting outcome of pulmonary valve replacement in adult tetralogy of Fallot patients. A Thorac Surg 2007;83(3): Oosterhof T, Mulder BJ, Vliegen HW, de Roos A. Corrected tetralogy of Fallot: delayed enhancement in right ventricular outflow tract. Radiology 2005;237(3): Vliegen HW, van Straten A, de Roos A, et al. Magnetic resonance imaging to assess the hemodynamic effects of pulmonary valve replacement in adults late after repair of tetralogy of Fallot. Circulation 2002;106(13): Oosterhof T, Mulder BJ, Vliegen HW, de Roos A. Cardiovascular magnetic resonance in the follow-up of patients with corrected tetralogy of Fallot: a review. Am Heart J 2006;151(2): van der Meer RW, Pattynama PM, van Strijen MJ, et al. Right ventricular dysfunction and pulmonary obstruction index at helical CT: prediction of clinical outcome during 3-month follow-up in patients with acute pulmonary embolism. Radiology 2005;235(3): Ghaye B, Ghuysen A, Willems V, et al. Severe pulmonary embolism: pulmonary artery clot load scores and cardiovascular parameters as predictors of mortality. Radiology 2006;239(3): Doğan H, Kroft LJ, Huisman MV, van der Geest RJ, de Roos A. Right ventricular function in patients with acute pulmonary embolism: analysis with electrocardiography-synchronized multi detector row CT. Radiology 2007;242(1): Kang DK, Thilo C, Schoepf UJ, et al. CT signs of right ventricular dysfunction: prognostic role in acute pulmonary embolism. JACC Cardiovasc Imaging 2011;4(8): Becattini C, Agnelli G, Vedovati MC, et al. Multidetector computed tomography for acute pulmonary embolism: diagnosis and risk stratification in a single test. Eur Heart J 2011;32(13): Quiroz R, Kucher N, Schoepf UJ, et al. Right ventricular enlargement on chest computed tomography: prognostic role in acute pulmonary embolism. Circulation 2004;109(20): Mansencal N, Joseph T, Vieillard-Baron A, et al. Diagnosis of right ventricular dysfunction in acute pulmonary embolism using helical computed tomography. Am J Cardiol 2005;95(10): radiology.rsna.org n Radiology: Volume 265: Number 1 October 2012