Related Experiment Videos
Model-based determination of cut-off values for left ventricular hypertrophy from echocardiographic myocardial mass
D Morvan1, J L Golmard, M Komajda
1Service de Cardiologie, Hôpital Pitié-Salpêtrière, Paris, France.
Insights
A new non-linear model accurately calculates echocardiographic left ventricular myocardial mass distribution in normal subjects. This model helps establish reliable cut-off values for diagnosing left ventricular hypertrophy, accounting for various patient covariates.
Area of Science:
- Cardiology
- Medical Imaging
- Biostatistics
Background:
- Left ventricular myocardial mass (LVMM) is crucial for diagnosing left ventricular hypertrophy (LVH).
- Current LVMM cut-off values are critical and influenced by covariates like sex, age, and body surface area.
- Establishing experimental distributions for LVMM is challenging due to the difficulty in collecting sufficient normal subject data.
Purpose of the Study:
- To propose a novel non-linear model for calculating echocardiographic LVMM distribution in normal subjects.
- To develop a method for establishing reliable LVMM cut-off values, considering patient covariates.
- To improve the diagnostic accuracy of LVH detection using echocardiography.
Main Methods:
- A non-linear model was developed using personal and literature data.
- The model assumes a bivariate normal distribution for left ventricular diameters and uses the Devereux & Reicheck formula.
- Gaussian assumptions were validated using skewness tests; the model was adapted for myocardial mass index distribution.
Main Results:
- The model demonstrated excellent agreement with experimental data for LVMM probability density functions.
- Calculated probability density functions were successfully used to define LVH cut-off values at selected false-positive ratios.
- The model's ability to provide covariate-matched cut-off values was confirmed.
Conclusions:
- The proposed non-linear model offers an accurate method for determining normal LVMM distribution via echocardiography.
- This model facilitates the establishment of precise, covariate-adjusted LVH diagnostic criteria, even with small control groups.
- The findings enhance the clinical utility of echocardiography for LVH assessment by providing more individualized cut-off values.
Abstract:
1. The left ventricular myocardial mass is a measurement that is easy to obtain by echocardiography. It is currently used for the definition of left ventricular hypertrophy, but cut-off values are often critical, since they depend on covariates of left ventricular myocardial mass such as sex, age, body surface area, physical training, blood pressure, etc. As it is very difficult in any laboratory to obtain a sufficient number of normal subjects for the establishment of left ventricular myocardial mass experimental distributions, we propose a non-linear model for the calculation of echocardiographic left ventricular myocardial mass distribution in normal subjects, from personal and literature data. left ventricular myocardial mass probability density function was computed from the following two assumptions: the joint distribution of the internal and external left ventricular diameters is assumed to be bivariate normal, and the relation between left ventricular myocardial mass and ventricular diameters is given by the formula of Devereux & Reicheck (Devereux, R. B. & Reicheck, N. Circulation 1977; 55, 613-8). 2. The Gaussian assumption was tested by using skewness tests. The model was further developed for the myocardial mass index distribution. The calculated probability density functions were compared with experimental data and showed very good agreement. Furthermore, they were used to define cut-off values of left ventricular hypertrophy at selected false-positive ratios. Finally, since left ventricular myocardial mass may vary under normal conditions with co-variates, the model may provide co-variate-matched cut-off values for any, even small, series of non-diseased control subjects.