MR evaluation of left ventricular volumes and function: threshold-based 3D segmentation versus short-axis planimetry

Kai Nassenstein, Armin de Greiff, Peter Hunold
Investigative Radiology 2009, 44 (10): 635-40

OBJECTIVES: Our study aimed to evaluate a three-dimensional (3D) threshold-based, region-growing segmentation algorithm for the assessment of left ventricular (LV) volumes in cardiac MRI.

MATERIALS AND METHODS: Two-dimensional cine steady-state free precession short-axis scans (TR 3 milliseconds, TE 1.5 milliseconds, FA 60 degrees , section thickness 8 mm) covering the entire LV without slice gap were acquired in 44 consecutive patients on a 1.5 Tesla MR system (Magnetom Avanto, Siemens, Germany). LV volumes were assessed with an in-house written 3D threshold-based, interslice region-growing segmentation algorithm, and were compared with manual tracing and short-axis planimetry with semiautomatic contour detection (Argus software, Siemens, Germany).

RESULTS: Three-dimensional threshold-based segmentation was feasible in 41 of 44 patients. Results of threshold-based segmentation were not statistically significantly different to manual tracing for the assessment of end-diastolic (EDV: 104.5 +/- 32.3 mL vs. 104.7 +/- 33.26 mL, P > 0.05), end-systolic (ESV: 34.5 +/- 17.2 mL vs. 33.8 +/- 17.45 mL, P > 0.05), and stroke (SV: 69.9 +/- 25.0 mL vs. 70.9 +/- 25.0 mL, P > 0.05) volumes. Bland-Altmann plots showed systematic lower EDV (-12.0 +/- 6.7 mL), ESV (-4.8 +/- 3.7 mL), and SV (-7.3 +/- 7.6 mL), and a systematic higher ejection fraction (0.4% +/- 5.0%) for threshold-based segmentation compared with short-axis planimetry with semiautomatic contour detection. Threshold-based segmentation showed an excellent reproducibility with low absolute differences between 2 consecutive analyses (EDV: 0.2 +/- 2.2 mL; ESV: 0.5 +/- 1.9 mL; SV: 0.6 +/- 3.4 mL; EF: 0.5% +/- 2.1%). Evaluation times were significantly shorter for threshold-based segmentation (60 +/- 16 vs. 278 +/- 31 seconds/669 +/- 157 seconds, P < 0.0001).

CONCLUSION: Automated threshold-based segmentation of the left ventricle allows fast and reproducible assessment of LV volumes in cardiac MRI. Exclusion of papillary muscles and myocardial trabeculations from LV volumes leads to systematically lower LV volumes compared with short axis planimetry with semi-automatic contour detection.

Full Text Links

Find Full Text Links for this Article


You are not logged in. Sign Up or Log In to join the discussion.

Related Papers

Remove bar
Read by QxMD icon Read

Save your favorite articles in one place with a free QxMD account.


Search Tips

Use Boolean operators: AND/OR

diabetic AND foot
diabetes OR diabetic

Exclude a word using the 'minus' sign

Virchow -triad

Use Parentheses

water AND (cup OR glass)

Add an asterisk (*) at end of a word to include word stems

Neuro* will search for Neurology, Neuroscientist, Neurological, and so on

Use quotes to search for an exact phrase

"primary prevention of cancer"
(heart or cardiac or cardio*) AND arrest -"American Heart Association"