NetForum uses cookies to ensure that we give you the best experience on our website. If you continue to use the site, we'll assume that you are happy to receive these cookies on the NetForum website. Read about our cookies.
NetForum Community
Learn. Share. Optimize.
Log in | Sign up now | Submit content | Contact
Go to similar content

Tips for delayed enhancement imaging

Application Tip
de Kok, Wendy Philips Healthcare Philips Global

Delayed enhancement cardiac MR imaging (DE-CMR) is often used to evaluate infarcted myocardium to see if enough viable tissue is available for revascularization.

 

DE-CMR is based on the fact that infarctions enhance strongly after administering gadolinium-based contrast agent. After the contrast agent perfused into the myocardium, it will rapidly wash-out again from healthy or viable myocardium. Contrast agent wash-out  from infarcted or necrosed myocardium progesses much slower. This results in a contrast difference between viable and non-viable myocardium on T1-weighted sequences after a certain delay time.

This document provides tips to optimize delayed enhancement imaging.

How to obtain the optimal contrast

A DE-CMR image is only diagnostic, if contrast between viable and non-viable tissue is sufficient .

The standard sequence for DE-CMR is a cardiac triggered TFE-sequence with an inversion prepulse. The inversion delay time (TI) is chosen so that a contrast difference between viable and non-viable myocardium is visualized.

 

T1-relaxation of non-viable myocardium progresses faster due to the higher concentration of contrast agent. This will lead to brighter signal for the non-viable myocardium in a real image, for a relatively large range of inversion delay times (see Figure 1).

 

However, modulus images are usually obtained, in which the choice of inversion delay time is more critical for obtaining optimal image contrast between viable and non-viable myocardium. (see Figure 2 and 3).

 

If TI is too short, the negative magnetization of viable tissue may equal the positive magnetization of non-viable tissue, resulting in images with equal signal intensity for both tissues.

 

Click on a figure to enlarge

 

Fig. 1. IR_TFE real Sufficient contrast between scar and normal for any TI that is indicated with the orange rectangle.Fig. 2. IR_TFE modulus Optimal contrast between scar and normal if TI is set to null the signal of normal myocardium.Fig. 3. IR_TFE modulus No contrast difference between scar and normal if TI is too short.
Fig. 1. IR_TFE real
Fig. 2. IR_TFE modulus
Fig. 3. IR_TFE modulus
Sufficient contrast between scar and normal for any TI that is indicated with the orange rectangle.
Optimal contrast between scar and normal if TI is set to null the signal of normal myocardium.
No contrast difference between scar and normal if TI is too short.

 

The optimal prepulse delay time depends on the contrast agent dose, and on the time between the contrast injection and the acquisition of the delayed enhancement scan.

 

The time between injection and scan is usually 10 to 30 minutes. This time can be spent on acquisition of other scans. (If possible, plan scans that benefit from having a contrast agent in the blood pool, like coronary artery imaging).

 

Note: PSIR (Phase Sensitive Inversion Recovery, or autoviability) generates corrected real images, that are virtually insensitive for prepulse delay timing. PSIR will be available in R2.6.

How to determine the prepulse delay time; the method

Based on experience, one could determine the optimal prepulse delay time by trial-and-error: the IR-TFE sequence can usually be obtained in one breath-hold and could just be repeated with various inversion delay times to find the optimal contrast.

 

Another method is to use Interactive scanning, where the inversion delay time can be modified on the fly.

 

More common however, is to acquire a single-slice multi-phase sequence with a shared inversion prepulse. One inversion prepulse is applied after which multiple low resolution TFE-images are acquired at different phases in the cardiac cycle. The images of each cardiac phase will show T1-weighted contrast that is based on the delay time between the inversion prepulse and the acquisition of the image. The trigger delay time that is annotated in each image represents the inversion delay time. This method is known as Look-Locker. (see Figure 4)

Fig. 4. Look-Locker Look-Locker sequence shows change in T1-weighting over the cardiac phases due to varying inversion delay time.Fig. 5. Intensity display Signal intensity display of a ROI drawn in the myocardium.
Fig. 4. Look-Locker
Fig. 5. Intensity display
Look-Locker sequence shows change in T1-weighting over the cardiac phases due to varying inversion delay time.
Signal intensity display of a ROI drawn in the myocardium.

 

The optimal inversion delay time to null the signal of normal myocardium can be selected by visual inspection. Additionally, a region of interest (ROI) can be created and copied over all phases, to display a signal intensity graph.

 

Workflow:

  • Select a ROI type from the image's right mouse menu.
  • Draw ROI in myocardial septum.
  • Select "copy all" from the menu at the ROI identifier.
  • Set running attribute to "phases".
  • Select "statistics, intensity" from the menu at the ROI identifier.

 

The time scale of the intensity display can be zoomed in under "settings..."

Table of signal intensity values at each cardiac phase can be opened from the intensity display.

How to optimize the Look-Locker sequence

The Look-Locker sequence was originally designed to determine T1-values of tissue. T1-relaxation curves are known to be disturbed by RF pulses, and the original Look-Locker sequence was designed to minimize disturbance of T1-relaxation. This was done by using TFE-EPI with a very low TFE factor (1 or 2), a small flip angle and a relatively long TR, to avoid saturation of the magnetization.

 

With DE-CMR, the Look-Locker technique is used to determine prepulse delay time, and it is less important to find the most accurate T1-value. The sequence can therefore be adapted to improve image quality and to optimize TI determination for the specific DE-CMR scan.

 

The optimal inversion delay time is usually between 200-300 ms, depending on the delay after contrast injection. It is therefore sufficient to acquire phases only around this time after the shared prepulse.

 

Suggestions to improve image quality and reliability of TI-determination:

  • Increase TFE-factor and reduce EPI-factor for reduced artifact levels.
  • Reduce TR for increased temporal resolution (at the cost of SNR...).
  • Use TFE prepulse delay (e.g. 150 ms) to skip phases in the first part of the RR-interval .
  • Use a large RR-window (first %) to reduce the no trigger period: check max trigger delay on info page (see Fgures 6 and 7).

 

Fig. 6 Schematic display of Look-Locker during the cardiac cycle, continuous acquisition of phases.Fig. 7 Schematic display of Look-Locker during the cardiac cycle, using a prepulse delay and an increased R-window to reduce the number of acquired phases.
Fig. 6
Fig. 7
Schematic display of Look-Locker during the cardiac cycle, continuous acquisition of phases.
Schematic display of Look-Locker during the cardiac cycle, using a prepulse delay and an increased R-window to reduce the number of acquired phases.

How to choose the optimal prepulse delay time - practical tips

The signal intensity of normal myocardium will gradually decrease over the phases, as the negative magnetization relaxes back to equilibrium. Its signal intensity will be zero when the magnetization crosses zero, and then it will gradually increase again as the positive magnetization increases in size.

 

The temporal resolution of the Look-Locker sequence must be sufficient to detect the moment of minimal signal of normal myocardium. Note that this zero-crossing may occur between two phases of the Look-Locker sequence. The temporal resolution should be 40 ms at maximum, but preferrably shorter, and is determined by the phase interval (that is displayed on the info page).

 

If it is suspected that zero-crossing occurred between two phases, it is advised to choose the timing of the phase that was acquired directly after zero-crossing, just to avoid that the TI is too short. (see Figures 8 and 9)

 

Note: If TI is too long, normal myocardium will not be completely black on the modulus image. WW/WL can be changed in this case.

Fig. 8. TI too short A too short inversion delay time leads to nearly equal signal intensity of myocardium and (contrast-rich) blood pool.Fig. 9. TI right Sufficiently long inversion delay time results in large image contrast between myocardium and (contrast-rich) blood pool.
Fig. 8. TI too short
Fig. 9. TI right
A too short inversion delay time leads to nearly equal signal intensity of myocardium and (contrast-rich) blood pool.
Sufficiently long inversion delay time results in large image contrast between myocardium and (contrast-rich) blood pool.

 

T1-relaxation is a continuous process, and the optimal TI will become longer if more time has passed since the injection of the contrast agent. To compensate for increased T1-relaxation between the inspection of the Look-Locker and the acquisition of the first delayed enhancement scan, it is advised to add 20-30 ms to the estimated TI.

 

The same applies to the acquisition of successive DE-CMR scans:

TI should be increased with a few ms for each next scan.

How to choose the TFE profile order

For delayed enhancement, is an inversion recovery prepared TFE sequence (IR_TFE_BH) is used. The TFE profile order can be set to either linear or low-high. Low-high profile order is required if additional prepulses like REST slabs or SPIR are added to the sequence.

 

A comparison between low-high and linear profile order showed that use of a linear profile order is advantageous to reduce edge enhancement and ringing effects in the final image, and to increase image contrast, even though image blurring might slightly increase.

 

The edge enhancement effects that sometimes appear with low-high profile order, are more pronounced if the inversion delay time is (too) short and will disappear with increasing TI.

 

It is therefore advised to use linear profile order if no additional prepulses are required.

(see Figures 10, 11 and 12)

 

Fig. 10. TI 240 low-high Edge enhancement artifacts in the middle of the myocardium at TI=240 ms and low-high profile order.Fig. 11. TI 240 linear Increased image contrast (blood pool/myocardium) and reduced edge enhancement effects at TI=240 ms and linear profile order.Fig. 12.  TI 270 low-high Reduced edge enhancement effects at longer TI (270 ms) and low-high profile order. Note: patient motion occurred
Fig. 10. TI 240 low-high
Fig. 11. TI 240 linear
Fig. 12. TI 270 low-high
Edge enhancement artifacts in the middle of the myocardium at TI=240 ms and low-high profile order.
Increased image contrast (blood pool/myocardium) and reduced edge enhancement effects at TI=240 ms and linear profile order.
Reduced edge enhancement effects at longer TI (270 ms) and low-high profile order. Note: patient motion occurred

How to define the TFE shot length

The TFE acquistion shot duration is determined by the TFE factor (contrast page), and is indicated in ms on the info page. The TFE shot length must match the resting period of the heart in diastole, to avoid cardiac motion blurring. The duration of the quiet period of the heart in diastole depends on the RR-interval, i.e. on the patient's heart rate.

In general: shot length must be reduced if heart rate increases and vice versa.

 

The protocol's standard TFE-factor (and shot length of ~100 ms) results in limited motion blurring as long as the heart rate is below +/- 80 bpm. The TFE factor should be reduced at higher heart rates.

 

Reducing the TFE factor further will narrow the point spread function and reduce image blurring, but will result in increased scan time (=breath-hold time). This may be done if longer breath-hold time is feasible.

 

How to choose scan mode - M2D versus 3D

DE-CMR can be acquired as M2D or 3D sequence, and both can be acquired in an acceptable breath-hold time. The list below scores the most important characteristics of both techniques, for breath-hold acquisitions:

 

In 3D acquisitions, it is important to acquire a 3D volume with sufficient coverage in one breath-hold. This is achieved by applying SENSE: Doubling the volume will double scan time and SNR. A SENSE factor of two reduces scan time and SNR again with a factor of two. The result is a 3D volume with twice the size in equal scan time and with equal SNR.

(see Figures 13 and 14 for 2D and 3D example)

 

If breath-holding is not feasible, respiratory compensation using a navigator can be applied.

To ensure optimal navigator respiratory compensation, it is advised to use SPIR fat suppression. TFE profile order must be low-high in that case.

 

Fig. 13. 2D DE-CMR High spatial resolution, but limited SNR.Fig. 14. 3D DE-CMR (1 slice) Increased SNR, at the cost of spatial resolution.
Fig. 13. 2D DE-CMR
Fig. 14. 3D DE-CMR (1 slice)
High spatial resolution, but limited SNR.
Increased SNR, at the cost of spatial resolution.

How to deal with heart rate variations - use multiple RR-intervals

The DE-CMR sequence uses prospective cardiac triggering. The trigger delay time (motion page) is set to mid-diastole to acquire TFE shots only during the resting period of the heart in diastole.

The trigger delay time is constant during the entire scan, and defines the time between the detection of the R-peak and the acquisition of the TFE shot. If heart rate variation (arrhythmia) occurs during the scan, timing of some TFE shots will be different from others, leading to cardiac motion blurring.

 

Heart rate variations usually average out over more RR-intervals and therefore it might be good to acquire one TFE shot every second RR-interval only. Set TFE shot interval to user defined, 2 beats (contrast page) to achieve this.

 

T1-relaxation is also affected by the duration of the RR-interval: it continues until the next R-peak is detected, after which all longitudinal magnetization is inverted again. Variations in RR-interval will result in different Mz and thus in contrast differences over the TFE shots.

Even with regular heart rates, results may be better with a TFE shot interval of two (or more) beats in patients with very high heart rates -- such as pediatric patients -- to allow for sufficient T1-relaxation (increasing SNR!) before the next inversion pulse is applied.

 

The optimal inversion delay time will change if the DE-CMR sequence is acquired over more beats. The Look-Locker sequence can also be acquired over more beats by entering a heart rate that is lower than the patient's actual heart rate (e.g. ~50%), making sure that every second R-top is not detected as a new trigger. Don't forget to enter the patient's actual heart rate again before starting the DE-CMR scan!! (This work-around is no longer necessary in R2.6, where user-defined shot interval is available for multi-phase scans as well).

 

Note: Arrhythmia rejection can be enabled.

Fig. 15 Shot interval shortest.Fig. 16 Shot interval 2 beats for improved SNR and CNR due to increased T1-relaxation.
Fig. 15
Fig. 16
Shot interval shortest.
Shot interval 2 beats for improved SNR and CNR due to increased T1-relaxation.

How to detect edema - T2 weighting

The DE-CMR exam can be extended with additional T2-weighted sequences, to detect possible edema in the myocardium. In acute infarction, edema might provide a means of differentiating acute and chronic myocardial infarction.

 

T2-weighted black blood sequences can be used for this purpose, or alternatively use a balanced TFE sequence with T2-prep pulses. These T2-prep pulses suppress signal of normal myocardium in which the edema shows brighter. (see Figure 17 for example in healthy volunteer)

 

As image contrast is affected by gadolinium-based contrast agent, it is recommended to acquire this sequence pre-contrast.

Fig. 17. 2D_BTFE using T2-prep
Fig. 17. 2D_BTFE using T2-prep


This content has been made possible by NetForum Community.
Share this on: Share your link in twitter Share your link in facebook Share your link on LinkedIn Print Rate this article: Log in to vote

 
Rating:
Votes:
5
Views:
12831
Added:
Feb 28, 2008

Rate this:
Log in to vote
 

Application Tip
Achieva 1.5T, Achieva 3.0T, Achieva 3.0T X-series
Release 1, Release 2, Release 2.5
Nova, Nova Dual, Pulsar, Quasar, Quasar Dual
Cardiac, DE-CMR, Function Non-Cine
 

Clinical News
Best Practices
Case Studies
Publications and Abstracts
White Papers
Web seminars and Presentations
ExamCards
Protocols
Application Tips and FAQ
Training
Try an Application
Business News
Case Studies
White Papers
Web Seminars and Presentations
Utilization Services
Contributing Professionals
Contributing Institutions
Become a Contributor