Murilo Foppa, Hospital de Clinicas de Porto Alegre
8 May 2015
We read with great interest the extensive review from Kawel-Boehm et al.[1] with proposed normal values for cardiovascular magnetic resonance (CMR) imaging. This is an important and essential step towards standardization of CMR acquisition and measurement protocols.
It should be highlighted that normality definition is complex. Ideally, cut-offs should be based on the best separation between those with from those without a predefined disease or condition (diagnostic), or at risk for that (prognostic), but this is frequently not feasible. An alternative method is to define what is abnormal as a statistic deviation from the distribution of a supposedly normal population. This method is also likely more robust to identify alterations in unexpected or unclear conditions. The criticism for this approach, however, is how to correctly identify a representative sample from the normal population. Two strategies are usually employed: to identify and study a convenient sample of normal individuals selected based on clinical criteria, or to use a large and representative sample, powered enough to statistically dilute individual impact and to narrow confidence intervals. This strategy permits to describe the distributions across relevant subgroups, with the additional benefit of allowing investigation of multiple determinants simultaneously.
In the review by Kawel-Boehm et al.[1] we can infer that most of currently available evidence for CMR reference values is based in the alternative method, the deviation from normal strategy. Authors assessed potential impact of CMR technique and the biological variation of the well-recognized physiological determinants of cardiac size/function, namely sex, body-size and age. Of note, however, many of the reference values analyzed were selected from studies with small samples of normal individuals, which might impair generalizability, especially if we consider the large reported variations related to demographics.
For example, regarding left ventricular volumes, there are 2 large community representative samples published. In MESA[2] , even though an older gradient-recalled echo sequence was used, values were not much different from the FHS-Offspring study[3], which used the modern and currently in use steady-state free precession sequence. Adding these studies to the analysis would allow a more confident definition of the true boundaries of the normal CMR-derived ventricular measurements.
In the future, data from planned or ongoing large national cohort studies [4, 5] and multinational registries[6] will offer a valuable tool for normative data. Meanwhile, there will be still a gap of unknown size between what we measure and what can be stated as abnormal.
Murilo Foppa, MD DSc*
Felipe Soares Torres, MD PhD**
*Cardiology and **Radiology Divisions
Hospital de Clinicas de Porto Alegre - Universidade Federal do Rio Grande do Sul
Ramiro Barcelos 2350. Room 2061
Porto Alegre – RS – Brasil 9035-003
mufoppa@hcpa.edu.br
References
1. Kawel-Boehm N, Maceira A, Valsangiacomo-Buechel ER, Vogel-Claussen J, Turkbey EB, Williams R, Plein S, Tee M, Eng J, Bluemke DA: Normal values for cardiovascular magnetic resonance in adults and children. J Cardiovasc Magn Reson 2015, 17:29.
2. Natori S, Lai S, Finn JP, Gomes AS, Hundley WG, Jerosch-Herold M, Pearson G, Sinha S, Arai A, Lima JAC, Bluemke DA: Cardiovascular function in multi-ethnic study of atherosclerosis: normal values by age, sex, and ethnicity. Am J Roentgenol 2006, 186(6 Suppl 2):S357–365.
3. Yeon SB, Salton CJ, Gona P, Chuang ML, Blease SJ, Han Y, Tsao CW, Danias PG, Levy D, O’Donnell CJ, Manning WJ: Impact of age, sex, and indexation method on MR left ventricular reference values in the framingham heart study offspring cohort. J Magn Reson Imaging 2014.
4. Petersen SE, Matthews PM, Bamberg F, Bluemke DA, Francis JM, Friedrich MG, Leeson P, Nagel E, Plein S, Rademakers FE, Young AA, Garratt S, Peakman T, Sellors J, Collins R, Neubauer S: Imaging in population science: cardiovascular magnetic resonance in 100,000 participants of UK Biobank - rationale, challenges and approaches. J Cardiovasc Magn Reson 2013, 15:46.
5. The German National Cohort: aims, study design and organization. Eur J Epidemiol 2014, 29:371–382.
6. Bruder O, Wagner A, Lombardi M, Schwitter J, Rossum A van, Pilz G, Nothnagel D, Steen H, Petersen S, Nagel E, Prasad S, Schumm J, Greulich S, Cagnolo A, Monney P, Deluigi CC, Dill T, Frank H, Sabin G, Schneider S, Mahrholdt H: European cardiovascular magnetic resonance (EuroCMR) registry – multi national results from 57 centers in 15 countries. J Cardiovasc Magn Reson 2013, 15:9.
CMR reference values: Are we ready for normals?
8 May 2015
We read with great interest the extensive review from Kawel-Boehm et al.[1] with proposed normal values for cardiovascular magnetic resonance (CMR) imaging. This is an important and essential step towards standardization of CMR acquisition and measurement protocols.
It should be highlighted that normality definition is complex. Ideally, cut-offs should be based on the best separation between those with from those without a predefined disease or condition (diagnostic), or at risk for that (prognostic), but this is frequently not feasible. An alternative method is to define what is abnormal as a statistic deviation from the distribution of a supposedly normal population. This method is also likely more robust to identify alterations in unexpected or unclear conditions. The criticism for this approach, however, is how to correctly identify a representative sample from the normal population. Two strategies are usually employed: to identify and study a convenient sample of normal individuals selected based on clinical criteria, or to use a large and representative sample, powered enough to statistically dilute individual impact and to narrow confidence intervals. This strategy permits to describe the distributions across relevant subgroups, with the additional benefit of allowing investigation of multiple determinants simultaneously.
In the review by Kawel-Boehm et al.[1] we can infer that most of currently available evidence for CMR reference values is based in the alternative method, the deviation from normal strategy. Authors assessed potential impact of CMR technique and the biological variation of the well-recognized physiological determinants of cardiac size/function, namely sex, body-size and age. Of note, however, many of the reference values analyzed were selected from studies with small samples of normal individuals, which might impair generalizability, especially if we consider the large reported variations related to demographics.
For example, regarding left ventricular volumes, there are 2 large community representative samples published. In MESA[2] , even though an older gradient-recalled echo sequence was used, values were not much different from the FHS-Offspring study[3], which used the modern and currently in use steady-state free precession sequence. Adding these studies to the analysis would allow a more confident definition of the true boundaries of the normal CMR-derived ventricular measurements.
In the future, data from planned or ongoing large national cohort studies [4, 5] and multinational registries[6] will offer a valuable tool for normative data. Meanwhile, there will be still a gap of unknown size between what we measure and what can be stated as abnormal.
Murilo Foppa, MD DSc*
Felipe Soares Torres, MD PhD**
*Cardiology and **Radiology Divisions
Hospital de Clinicas de Porto Alegre - Universidade Federal do Rio Grande do Sul
Ramiro Barcelos 2350. Room 2061
Porto Alegre – RS – Brasil 9035-003
mufoppa@hcpa.edu.br
References
1. Kawel-Boehm N, Maceira A, Valsangiacomo-Buechel ER, Vogel-Claussen J, Turkbey EB, Williams R, Plein S, Tee M, Eng J, Bluemke DA: Normal values for cardiovascular magnetic resonance in adults and children. J Cardiovasc Magn Reson 2015, 17:29.
2. Natori S, Lai S, Finn JP, Gomes AS, Hundley WG, Jerosch-Herold M, Pearson G, Sinha S, Arai A, Lima JAC, Bluemke DA: Cardiovascular function in multi-ethnic study of atherosclerosis: normal values by age, sex, and ethnicity. Am J Roentgenol 2006, 186(6 Suppl 2):S357–365.
3. Yeon SB, Salton CJ, Gona P, Chuang ML, Blease SJ, Han Y, Tsao CW, Danias PG, Levy D, O’Donnell CJ, Manning WJ: Impact of age, sex, and indexation method on MR left ventricular reference values in the framingham heart study offspring cohort. J Magn Reson Imaging 2014.
4. Petersen SE, Matthews PM, Bamberg F, Bluemke DA, Francis JM, Friedrich MG, Leeson P, Nagel E, Plein S, Rademakers FE, Young AA, Garratt S, Peakman T, Sellors J, Collins R, Neubauer S: Imaging in population science: cardiovascular magnetic resonance in 100,000 participants of UK Biobank - rationale, challenges and approaches. J Cardiovasc Magn Reson 2013, 15:46.
5. The German National Cohort: aims, study design and organization. Eur J Epidemiol 2014, 29:371–382.
6. Bruder O, Wagner A, Lombardi M, Schwitter J, Rossum A van, Pilz G, Nothnagel D, Steen H, Petersen S, Nagel E, Prasad S, Schumm J, Greulich S, Cagnolo A, Monney P, Deluigi CC, Dill T, Frank H, Sabin G, Schneider S, Mahrholdt H: European cardiovascular magnetic resonance (EuroCMR) registry – multi national results from 57 centers in 15 countries. J Cardiovasc Magn Reson 2013, 15:9.
Competing interests
Authors declare no conflicts of interest.