Ventricular Volume and Systolic Function Assessment by Cardiac Multidetector Computed Tomography

Open access:

The copyright in this work belongs to Radcliffe Medical Media. Only articles clearly marked with the CC BY-NC logo are published with the Creative Commons by Attribution Licence. The CC BY-NC option was not available for Radcliffe journals before 1 January 2019. Articles marked ‘Open Access’ but not marked ‘CC BY-NC’ are made freely accessible at the time of publication but are subject to standard copyright law regarding reproduction and distribution. Permission is required for reuse of this content.

Accurate quantification of ventricular volumes and function is important for the assessment and management of patients with suspected ischemic and non-ischemic cardiomyopathies. 2D echocardiography is widely used for qualitative and semi-quantitative assessment of cardiac size and function. However, 2D echocardiography has limited accuracy, particularly in patients with regional variability in left ventricular (LV) function or remodeled hearts. 3D echocardiography has improved accuracy, but—like 2D echocardiography—is limited by poor acoustic windows and suboptimal endocardial definition in a significant minority of patients. Radionuclide ventriculography and cine ventriculography have also been used for assessment of LV ejection fraction (EF) in select patients, but are limited by poor spatial resolution and geometric assumptions, respectively. Cardiac magnetic resonance imaging (MRI), with its wide field of view and high spatial (1–2mm) and temporal (20–40ms) resolution, is currently the gold standard for accurate and reproducible assessment of ventricular volumes in both clinical and research settings.1,2 However, MRI has restricted availability, is costly, and there are many patients in whom MRI-incompatible devices prohibit its use.

A growing number of patients undergo cardiac multidetector computed tomography (MDCT) for non-invasive detection of coronary artery disease or other indications. Secondary simultaneous assessment of LV volumes and function, although not routinely performed, is generally possible and may provide incremental information. In addition, cardiac MDCT may have an important primary role in the assessment of patients in whom precise quantification of LV volumes and function is needed, but who have a contraindication to MRI. Cardiac MDCT has a spatial resolution (0.4–0.7mm) that is superior to MRI but a temporal resolution (83–200ms) that is inferior, despite utilization of multisegment reconstruction and dual-source/detector technology. Cardiac MDCT also has the advantage that the entire data set is rapidly acquired in a single breath-hold (7–15ms). In contrast, cardiac MRI typically takes 20–30 minutes to generate multiple short-axis and long-axis slices of the cardiac chambers. However, unlike cardiac MRI, cardiac MDCT requires administration of iodinated contrast material and is associated with a significant radiation dose for the patient (8–15mSv).

Accuracy of Multidetector Computed Tomography for Volumetric Analysis

Long before the advent of MDCT, electron beam CT (EBCT) was successfully utilized for the assessment of LV volumes and systolic function.3 Despite its excellent temporal resolution (50–100ms), EBCT is limited by modest spatial resolution (3mm non-overlapping axial slices) and restricted availability. In addition, prospective electrocardiogram (ECG) gating makes it particularly susceptible to changes in heart rate or arrhythmia.

Cardiac MDCT with retrospective ECG gating generates a large data set covering the entire cardiac cycle from which end-systolic and end-diastolic images are derived. Four- to 64- slice MDCT has been shown to compare favorably with cardiac MRI for assessment of LV volumes and EF.2,4–9 In clinical practice, at least 16 detector rows are required in order to allow scan acquisition within a single breath-hold (<16 seconds). Studies have generally demonstrated excellent correlation between MDCT and MRI for cardiac volume and function assessment, with closer agreement between MDCT and MRI than for other modalities such as 2D echocardiography, single photon emission computed tomography (SPECT), and cine ventriculography. Cardiac MDCT appears to consistently underestimate LVEF by 1–7% compared with cardiac MRI, possibly related to the inferior temporal resolution of cardiac MDCT. Predictably, the use of beta-blockers to slow the heart rate in patients undergoing MDCT appears to affect volume and EF measurements.8

Although generally robust, studies of MDCT have reported that up to 10% of LV segments are not evaluable by CT due to poor contrast opacification or motion artifact.10 Few studies have evaluated MDCT for quantification of right ventricular (RV) volumes and function.10–12 Nevertheless, it appears to compare favorably with assessment by MRI. The RV cavity may not be adequately filled with contrast in many cardiac MDCT scans, which are usually timed for maximal contrast opacification in the left side of the heart and coronary arteries. This may be particularly problematic for scanners with more detector rows (64- versus 16- slice), where the larger scan volume and faster acquisition time necessitates more rapid passage of contrast through the heart. The routine use of a saline flush following contrast injection to prevent streaking of contrast in the right heart also reduces contrast opacification in the RV. One study showed that 25% of RV segments were not evaluable due to poor contrast opacification.10 However, if contrast administration is timed correctly, RV volume assessment can be achieved successfully in all patients.12

How to Perform Volumetric Calculations by Multidetector Computed Tomography

Currently, volumetric and function analysis by cardiac MDCT can be performed by manual tracing of borders, semi-automated software, or fully automated systems. The process starts with the creation of CT volume data sets at 10% intervals through the cardiac cycle. From these, end-systolic and end-diastolic data sets are manually chosen by the technologist or reader. Manual or automatic generation of short-axis multiplanar reformation (MPR) slices of the LV and RV from base to apex allows calculation of end-diastolic and end-systolic volumes using Simpson’s method (see Figure 1).

These slices are typically 8–10mm thick with no gap. Tracing of the endocardial borders is then performed for the individual slices with subsequent calculation of the volume using automated software or manually using the following formula: volume = (area disc 1 + area disc 2 + ....+ area disc n) x slice thickness, where n equals the total number of slices. Care must be taken to ensure correct definition of the ventricular basal segments and to ensure that papillary muscles and trabeculations are included in the ventricular cavity volume. Simultaneous tracing of the epicardial border and calculation of total LV ventricular volume allows calculation of LV mass and regional LV thickening. EF is derived from the following formula:
EF = (LV end-diastolic volume – LV end-systolic volume)/ LV end-diastolic volume

Manual or semi-automated calculation of LV volumes using Simpson’s method by both CT and MRI is accurate and has been validated in numerous published studies. However, the post-processing and calculations are time-consuming, taking anywhere from 10 to 25 minutes to complete. A variation of Simpson’s method has been evaluated using axial slices instead of short-axis slices through the ventricles, but it does not appear to be as accurate.13 In addition, a number of fully or semi-automated software programs for detection of ventricular and atrial volumes are commercially available or in development. These software programs use a variety of border detection, density threshold, or shape variability algorithms to automatically segment the cardiac chambers and vessels. Comparison of some of these automated programs to manual calculation suggests comparable accuracy with significant reductions in analysis time, albeit in small numbers of patients.14–17 Atrial volume and function can be calculated using Simpson’s method, although the area length formula is used more frequently. Automated software tools for calculation of atrial volume are available, but await validation.14–17


In conclusion, ventricular volume and systolic function assessment by MDCT is accurate and reproducible. Ongoing development of semi-automated and automated software will allow easier calculation of these measurements in patients undergoing cardiac MDCT for evaluation of the coronary arteries or other indications, providing additive and potentially valuable information. In addition, there appears to be a primary role for MDCT for accurate assessment of cardiac volumes and function in the growing number of patients with cardiac devices where MRI is contraindicated.


  1. Lima JA, Desai MY, J Am Coll Cardiol, 2004;44:1164–71.
    Crossref | PubMed
  2. Van der Vleuten PA, Willems TP, Gotte MJ, et al., Acta Radiol, 2006;47:1049–57.
    Crossref | PubMed
  3. Lipton MJ, Radiol Clin North Am, 1985;23:613–26.
  4. Belge B, Coche E, Pasquet A, et al.,Eur Radiol, 2006;16:1424–33.
    Crossref | PubMed
  5. Dewey M, Muller M, Eddicks S, et al., J Am Coll Cardiol, 2006;48:2034–44.
    Crossref | PubMed
  6. Heuschmid M, Rothfuss JK, Schroeder S, et al., Eur Radiol, 2006;16:551–9.
    Crossref | PubMed
  7. Juergens KU, Grude M, Maintz D, et al., Radiology, 2004;230:403–10.
    Crossref | PubMed
  8. Schlosser T, Mohrs OK, Magedanz A, et al.,Acta Radiol, 2007;48:30–35.
    Crossref | PubMed
  9. Yamamuro M, Tadamura E, Kubo S, et al., Radiology, 2005;234:381–90.
    Crossref | PubMed
  10. Raman SV, Shah M, McCarthy B, et al., Am Heart J, 2006;151:736–44.
    Crossref | PubMed
  11. Koch K, Oellig F, Oberholzer K, et al., Eur Radiol, 2005;15:312–18.
    Crossref | PubMed
  12. Raman SV, Cook SC, McCarthy B, et al., Am J Cardiol, 2005;95:683–6.
    Crossref | PubMed
  13. Juergens KU, Seifarth H, Maintz D, et al., AJR Am J Roentgenol, 2006;186:S371–8.
    Crossref | PubMed
  14. Iler MA, Gomez de Diego JJ, Pollono PM, et al., Eur Heart J, 2006;27(Suppl. 1):147.
  15. Montaudon M, Laffon E, Berger P, et al., Eur Radiol, 2006;16:2341–9.
    Crossref | PubMed
  16. Muhlenbruch G, Das M, Pohl C, et al., Eur Radiol, 2006;16:1117–23.
    Crossref | PubMed
  17. Schlosser T, Pagonidis K, Herborn CU, et al., AJR Am J Roentgenol, 2005;184:765–73.
    Crossref | PubMed