Myocardial creep-induced misalignment artifacts in PET/MR myocardial perfusion imaging

Elia von Felten1, Georgios Benetos1, Dimitri Patriki1

  • 1Department of Nuclear Medicine, Cardiac Imaging, University Hospital Zurich, University of Zurich, Ramistrasse 100, CH-8091, Zurich, Switzerland.

Abstract

Insights

Misalignment between PET and MR attenuation correction maps is common in hybrid PET/MR myocardial perfusion imaging. Myocardial creep during adenosine stress significantly impacts image alignment, potentially causing false positives.

Area of Science:

  • Cardiovascular Imaging
  • Medical Physics
  • Radiology

Background:

  • Positron emission tomography (PET) and magnetic resonance (MR) imaging are advanced modalities for myocardial perfusion imaging (MPI).
  • Hybrid PET/MR scanners offer integrated data acquisition but face challenges with spatial alignment between PET and MR-derived attenuation correction (AC) maps.
  • Misalignment can introduce artifacts, potentially leading to misinterpretation of myocardial perfusion.

Purpose of the Study:

  • To assess the impact of adenosine on the spatial alignment of MR-derived AC maps and PET datasets during MPI.
  • To evaluate the frequency of misalignment and its effect on quantitative perfusion metrics.
  • To investigate the role of myocardial creep in causing misalignment during adenosine stress MPI.

Main Methods:

  • Twenty-eight volunteers underwent adenosine stress and rest 13N-ammonia MPI on a hybrid PET/MR scanner.
  • MRAC maps were generated using Dixon sequences; PET images were reconstructed with and without manual co-registration adjustments.
  • Summed rest, stress, and difference scores (SRS, SSS, SDS) were compared between shifted and non-shifted PET images; myocardial creep was quantified.

Main Results:

  • Realignment was necessary for 89.3% of stress and 42.9% of rest PET datasets.
  • Non-shifted images showed significantly higher SRS, SSS, and SDS compared to shifted images (p < 0.05).
  • The difference in SDS correlated significantly with myocardial creep (r=0.541, p=0.005).

Conclusions:

  • Misalignment between MRAC and PET datasets is frequent during adenosine stress MPI on hybrid PET/MR devices.
  • Myocardial creep appears to be a significant contributor to this misalignment.
  • Careful review and correction of PET/MRAC data are recommended to avoid false-positive findings in MPI.