1887
banner image
No data available.
Please log in to see this content.
You have no subscription access to this content.
No metrics data to plot.
The attempt to load metrics for this article has failed.
The attempt to plot a graph for these metrics has failed.
Automatic phase determination for retrospectively gated cardiac CT
Rent:
Rent this article for
USD
10.1118/1.1791351
/content/aapm/journal/medphys/31/12/10.1118/1.1791351
http://aip.metastore.ingenta.com/content/aapm/journal/medphys/31/12/10.1118/1.1791351

Figures

Image of FIG. 1.
FIG. 1.

This flowchart depicts the proposed procedure of cardiac CT volume reconstruction including the automatic determination of stable cardiac phases.

Image of FIG. 2.
FIG. 2.

This figure shows the principle of a gated helical CT acquisition. The source moves on a helical path , and an ECG signal is recorded in parallel. Based on a defined cardiac phase , the acquired data are gated resulting in a virtually interrupted helix. A volume is reconstructed with the gated data, representing the heart at the defined phase.

Image of FIG. 3.
FIG. 3.

This figure shows templates within two volumes at different cardiac phases. Between the templates at different phases, the similarity is calculated. From the calculated value, information about the motion at the phase at the spatial position is obtained.

Image of FIG. 4.
FIG. 4.

The template shift in direction is shown within this figure. For each template position , the similarity is calculated. With this technique, spatially resolved information of the cardiac motion is derived. The information can be related to each cardiac cycle according to the position of the template at . In this figure, the templates at and overlap spatially.

Image of FIG. 5.
FIG. 5.

A schematic plot of the dynamic phantom at a fixed time is shown in the upper sketch (a). The dynamic frustum is of high contrast with time-dependent radii and . It is surrounded by a water cylinder . In the lower sketch (b), the averaging within axial slices is shown. The average density of the reconstructed axial slice is calculated on a circle with variable radius .

Image of FIG. 6.
FIG. 6.

From the average of the similarity curves of all templates, shown in the upper plot (a), minima of global motion information can be derived. With the proposed algorithm, stable phases are determined at 12.4% and 62.4% of the motion cycle. In the motion map of the phantom [lower plot (b)], two valleys of minimal motion are shown. The space-dependent minima determined from the motion map are marked with the white crosses. The ideal minimum curve is illustrated with the overlaid red curves.

Image of FIG. 7.
FIG. 7.

In this figure, the phantom is shown in cross-sectional slices at two phases of minimal global motion . Images (a) and (b) represent the ideal phantom at 12.4%, respectively, 62.4% of the motion cycle, whereas image (c) and (d) represent reconstructions of the two phases of minimal global motion.

Image of FIG. 8.
FIG. 8.

This figure shows the average densities of the phantom reconstructions at the two phases of minimal global motion. The average densities have been determined on circles [see Fig. 5(b)] with radii [ on the axis] within the axial slices ( position along the axis) of the volumes. Plot (a) corresponds to the reconstruction at 12.4%, whereas plot (b) corresponds to 62.4% of the motion cycle. The ideal radius of the frustum is given by the black line.

Image of FIG. 9.
FIG. 9.

In this figure, the phantom is shown in cross-sectional slices with slice-specific optimal phases corresponding to the two valleys of minimal motion. Images (a) and (b) show the ideal phantom. Images (c) and (d) have been reconstructed with the ideal space-dependent stable cardiac phase [red curves in the motion map; see Fig. 6(b)]. Images (e) and (f) have been reconstructed with the space-dependent stable cardiac phases derived from the motion map (white cross marks in the motion map). The images in the left column correspond to valley 1 and the images in the right column correspond to valley 2 of the motion map.

Image of FIG. 10.
FIG. 10.

In this figure, the average densities of the phantom reconstructions with slice-specific gating according to the motion map are shown. Similar to Fig. 8, the average densities have been calculated on circles in the axial plane. The ideal space-dependent object radius of the frustum is given by the black line. Plot (a) represents the reconstruction at valley of minimal motion 1, whereas plot (b) represents the reconstruction at valley 2.

Image of FIG. 11.
FIG. 11.

From the average of the similarity curves of all templates, shown in the upper plot (a), minima of global motion information can be derived, e.g., at 39.7% RR and 76.9% RR. Maxima of global motion information can be determined at 16% RR and 58% RR. The motion map of case A is shown in the lower plot (b).

Image of FIG. 12.
FIG. 12.

Axial slices of case A at are shown in this figure . The images are reconstructed at every 10% RR, at two motion maxima and at the stable global phases derived with the automatic technique. The arrow marks the mitral valve.

Image of FIG. 13.
FIG. 13.

Sagittal slices of case A are shown in this figure , reconstructed at every 10% RR, at two motion maxima and at the stable global phases derived with the automatic technique. Two by-pass arteries are marked with white arrows. The mitral and the aortic valve are marked with dashed arrows. The solid black arrow marks the interventricular septum.

Image of FIG. 14.
FIG. 14.

From the average of the similarity curves of all templates, shown in the upper plot (a), minima of global motion information can be derived, e.g., at 36.0% RR and 67.6% RR. Maxima of global motion information can be determined at 22% RR and 48% RR. The motion map of case B is shown in the lower plot (b).

Image of FIG. 15.
FIG. 15.

Axial slices of case B at are shown in this figure , reconstructed at every 10% RR, at two motion maxima and at the stable global phases derived with the automatic technique.

Image of FIG. 16.
FIG. 16.

Sagittal slices of case B are shown in this figure , reconstructed at every 10% RR, at two motion maxima and at the stable global phases derived with the automatic technique. Black arrows mark coronary arteries. The myocardial border is tagged with the white arrow.

Image of FIG. 17.
FIG. 17.

From the average of the similarity curves of all templates, shown in the upper plot (a), minima of global motion information can be derived, e.g., at 4.3% RR and 46.7% RR. Maxima of global motion information can be determined at 28% RR and 92% RR. The motion map of case C is shown in the lower plot (b).

Image of FIG. 18.
FIG. 18.

Axial slices of case C at are shown in this figure , reconstructed at every 10% RR, at two motion maxima and at the stable global phases derived with the automatic technique.

Image of FIG. 19.
FIG. 19.

Coronal slices of case C are shown in this figure , reconstructed at every 10% RR, at two motion maxima and at the stable global phases derived with the automatic technique. The two solid black arrows mark coronary arteries. The myocardial border is marked with the dashed black arrow.

Image of FIG. 20.
FIG. 20.

Two coronal images at systole of case A are shown . The upper image has been reconstructed with a global stable phase position of 39.7% RR cycle. For the lower image, the information of the motion map [see Fig. 11(b)] has been used for reconstruction. For each cardiac cycle, respectively, every position, an appropriate stable phase has been determined from the motion map. The black arrows mark areas with potentially fewer artifacts.

Tables

Generic image for table
TABLE I.

Scan and reconstruction parameter for the simulation study and three different patient data sets are shown in this table.

Generic image for table
TABLE II.

This table shows the global optimal phase points for systole and diastole in per RR cycle for all cases. The values have been obtained by using the mean absolute difference (MAD) and the correlation coefficient (CC) for calculating the similarity. Low-resolution volumes with different spatial resolution and a variable number of phases (50 and 25) have been used for the analysis. The standard deviation (StDev) for all different phases is given for each case.

Loading

Article metrics loading...

/content/aapm/journal/medphys/31/12/10.1118/1.1791351
2004-11-22
2014-04-20
Loading

Full text loading...

This is a required field
Please enter a valid email address
752b84549af89a08dbdd7fdb8b9568b5 journal.articlezxybnytfddd
Scitation: Automatic phase determination for retrospectively gated cardiac CT
http://aip.metastore.ingenta.com/content/aapm/journal/medphys/31/12/10.1118/1.1791351
10.1118/1.1791351
SEARCH_EXPAND_ITEM