Diagnostic performance of black-blood versus gray-blood delayed enhancement imaging for the detection of myocardial infarction (RF_TH_152)

Presenting Author(s)

• 

  David C. Wendell, PhD

  Senior Research Associate
  Duke University Medical Center

Primary Author(s)

• 

  David C. Wendell, PhD

  Senior Research Associate
  Duke University Medical Center

Co-Author(s)

• HK

  Han Kim, MD

  Associate Professor of Medicine
  Duke University Medical Center
• EJ

  Elizabeth Jenista, PhD

  Research Scholar
  Duke Cardiovascular Magnetic Resonance Center
• 

  Clerio F. Azevedo, MD, PhD

  Cardiologist
  Minneapolis Heart Institute at Abbott Northwestern Hospital
• FA

  Fawaz Alenezi, MD

  Assistant Professor of Medicine
  Duke University Medical Center
• CC

  Céleste Chevalier, MD

  Postdoctoral fellow
  Duke University Medical Center
• SD

  Stephen Darty, BSc

  Cardiovascular MRI technologist
  Duke Cardiovascular Magnetic Resonance Center
• GG

  George Gamoneda, RT

  MR Technologist
  Duke University Hospital
• NM

  Nestor Mena, RT

  MR Technologist
  Duke University Hospital
• WR

  Wolfgang Rehwald, PhD

  Staff Scientist
  Siemens Healthineers
• EC

  Enn-Ling Chen, PhD

  Research Professor
  Duke Cardiovascular Magnetic Resonance Center
• MP

  Michele Parker, MSc

  Statistician / Business Manager
  Duke Cardiovascular Magnetic Resonance Center
• RK

  Raymond J. Kim, MD

  Professor of Medicine and Radiology
  Duke University Medical Center

Figure 1: Comparison of signal intensities between FIDDLE and BN-DE imaging in a control patient without MI. Cine imaging and conventional DE-CMR images (left) show region of analysis (orange line). Note BN-DE image has a suppressed blood pool signal compared to DE-CMR (e.g “gray-blood”). However, normal myocardium has lower image intensity (e.g. is black). For FIDDLE, the cavity signal is black whereas normal myocardium has higher image intensity (e.g is “gray”).

Figure 2: Comparison of signal intensities between FIDDLE and BN-DE in two patients with MI. A) MI has high conspicuity for both FIDDLE and BN-DE, although blood pool-to-infarct CNR is higher for FIDDLE. B) MI has divergent conspicuity between techniques, where FIDDLE easily detects the MI while in BN-DE the blood pool-to-infarct CNR is low since both blood pool and infarct are gray.

Figure 3: Blood pool inhomogeneity is seen in BN-DE images (center) which are absent for DE-CMR (left) and FIDDLE (right).


Background: Recent advancements in delayed-enhancement imaging (DE) have focused on suppressing the signal from the blood pool and increasing the conspicuity of subendocardial infarction(1-4). Flow-Independent Dark-blood DeLayed-Enhancement (FIDDLE) which combines magnetization preparation, inversion recovery, and PSIR produces “black-blood” DE images(5) whereas Blood-Nulled Delayed-Enhancement (BN-DE) where the inversion time is set to null the blood pool provides “gray-blood” PSIR images (Figure 1A)(4). A limitation of BN-DE is when the T1 of blood and infarct are similar.  In this situation both will appear gray which will reduce the conspicuity of infarction. This study compares the diagnostic performance of FIDDLE with BN-DE in patients with MI.

Methods: 61 patients with clinically documented MI (positive troponins and clear infarct related artery on cath) and 15 controls (lowest Framingham risk category for CAD) were prospectively enrolled. FIDDLE and BN-DE images were acquired interleaved 10 minutes after gadolinium administration. Images were evaluated in a masked fashion for the presence and transmurality of MI. Contrast-to-noise ratio (CNR) between infarct and LV blood pool was calculated.

Results:

FIDDLE demonstrated better diagnostic performance than BN-DE for detecting MI with a sensitivity of 100% versus 90% (p=0.03) and accuracy of 100% versus 89% (p< 0.01). Specificity was also higher for FIDDLE, however did not reach statistical significance (100% v 87%; p=0.5). For subendocardial infarcts (≤25% transmural), the sensitivity of BN-DE dropped to 81% (p=0.03) and accuracy to 85% (p< 0.01). Figure 2 shows typical images in two patients with MI showing that contrast between blood pool and infarction is consistently high for FIDDLE but can be low for BN-DE. Not surprisingly, the MI-to-blood-pool CNR was higher for FIDDLE (57.8±37.6) compared to BN-DE (7.9±19.7; p< 0.0001). Figure 3 demonstrates that blood pool homogeneity can occasionally be poor for BN-DE. This is likely due to the short inversion time required for BN-DE imaging which may accentuate small differences in apparent T1 due to flow effects, B1 inhomogeneity, etc. 

Conclusion: FIDDLE provides improved diagnostic performance compared to BN-DE particularly in subendocardial infarcts. Gray blood techniques such as BN-DE may not detect infarcts when the T1 of blood and infarct are similar. 

Figure 1: Comparison of signal intensities between FIDDLE and BN-DE imaging in a control patient without MI. Cine imaging and conventional DE-CMR images (left) show region of analysis (orange line). Note BN-DE image has a suppressed blood pool signal compared to DE-CMR (e.g “gray-blood”). However, normal myocardium has lower image intensity (e.g. is black). For FIDDLE, the cavity signal is black whereas normal myocardium has higher image intensity (e.g is “gray”).

Figure 2: Comparison of signal intensities between FIDDLE and BN-DE in two patients with MI. A) MI has high conspicuity for both FIDDLE and BN-DE, although blood pool-to-infarct CNR is higher for FIDDLE. B) MI has divergent conspicuity between techniques, where FIDDLE easily detects the MI while in BN-DE the blood pool-to-infarct CNR is low since both blood pool and infarct are gray.

Figure 3: Blood pool inhomogeneity is seen in BN-DE images (center) which are absent for DE-CMR (left) and FIDDLE (right).