Left atrial surface cut point to detect trapezoidal shape for characterisation of atrial anatomical remodeling.

Background: Complete geometrical and shape characterization of left atrium has not been performed. Ellipse formula has been proved to underestimate the real LA volume. The aim of the study is to analyze the relation between LA area and shape in order to predict their value in the assessment of the severity of anatomical remodeling.

Methods: 216 consecutive patients (pts) aged 53 +/- 27 years were included. The following parameters were assessed: LA dimensions (LAd=M-mode, parasternal long axis, LAt and LA1 are the measurements of short- and long-axis in apical four chamber view), LA surface in apical four chamber view (LAs). A new measurement was introduced, the basal dimension of the LA (LAb) as the maximal transverse distance at the base of LA apical four chamber view. LA measurements were calculated at end-systole (maximal). Trapezoidal LA shape was defined if transverse dimension < basal dimension.

Results: LAs ranged 10.5-54.5 cm2. Trapezoidal LA was found in 149 pts. ROC curve for prediction trapezoidal LA showed a 0.89 area under the curve. The analysis demonstrates a cutoff value of 22.3 cm2 for LAs to detect trapezoidal shape with a sensitivity of 85% and specificity of 82.5%. The simple regression analysis demonstrated a statistically significant linear correlation between LAd and LAs (r2=0.63, p<0.0001) but LA1 was better correlated with LAs (r2=0.7, p<0.0001). Trapezoid LA with atrialization of the pulmonary veins and predominant dilatation of bazal atrium than annular side may explain underestimation of LA volume using ellipse formula.

Conclusion: LAs is a reliable parameter and may be the best choice to estimate LA dilatation; LAs is related to shape remodeling; LA1 is a better parameter than LAd for estimating LA dilatation and complete characterization of LA remodeling should include shape definition and LAb.