Feasibility of 4D Flow Hemodynamics as an Indicator of RV Dysfunction in Patients with Transposition of the Great Arteries (RF_TH_175)

Presenting Author(s)

• BC

  Brendan T. Crabb, MD

  Resident Physician
  University of California, San Diego

Primary Author(s)

• BC

  Brendan T. Crabb, MD

  Resident Physician
  University of California, San Diego

Co-Author(s)

• SW

  Sophie Y. Wong, MD

  Resident Physician
  University of California, San Diego
• PR

  Parker Rushworth, MD

  Fellow
  University of California, Irvine
• LA

  Laith Alshawabkeh, MD, MSc

  Associate Professor
  University of California, San Diego
• DV

  Daniela Valdez-Jasso, PhD

  Associate Professor
  University of California, San Diego
• 

  Albert Hsiao, MD, PhD

  Professor
  UC San Diego

Mean 4D Flow waveforms (A) and boxplots of 4D Flow hemodynamic metrics (B) for each d-TGA surgical repair subtype and normal controls. Arterial switch patients were characterized by increased RV-TTP, and a trend towards increased RVLV-dTTP relative to morphologically normal patients. Atrial switch patients were characterized by reduced RV-FWHM and increased RVLV-dTTP relative to morphologically normal patients.

4D Flow hemodynamic metrics for arterial switch patients (n=13) correlate strongly with the neopulmonary stenosis pressure gradient. These findings suggest that the altered RV systolic waveforms for d-TGA patients may be due to differences in afterload.


Background: Transposition of the great arteries (TGA) is a rare, but serious congenital heart defect defined by errant embryologic segmentation of the outflow tract. Surgical management of dextro-TGA (d-TGA) has evolved in the last fifty years, favoring arterial switch over atrial switch to facilitate using the morphologic left ventricle as the systemic ventricle. MRI is the reference standard for surveillance of patients with TGA, particularly for detecting complications of d-TGA repair and right ventricular (RV) failure. However, we lack a sensitive and reliable benchmark for RV dysfunction. We therefore sought to determine whether hemodynamic measurements from 4D Flow MRI might serve as a benchmark of RV dysfunction.

Methods: With IRB approval and a waiver of informed consent, we retrospectively analyzed a single-institutional convenience sample of 40 adult patients with either surgically repaired d-TGA or morphologically normal hearts, who had MRI including 4D Flow as part of routine clinical care between 2015 and 2023. Conventional measures of RV function including end-diastolic volume (EDV) and ejection fraction (EF) were tabulated. In addition, synchronous pulmonic and aortic blood flow waveforms from 4D Flow were collected. These were used to measure full-width half max of systolic ejection (RV-FWHM), time to peak flow (RV-TTP), and difference in time to peak flow (RVLV-dTTP) between the ventricular flow waveforms. We assessed for differences between subgroups using paired two-sample t-tests. In patients with arterial switch, we further explored the relationship between these metrics and the severity of neopulmonary outflow stenosis using Spearman’s rank correlation (r).

Results: Of the 40 patients analyzed, there were 14 atrial switch d-TGA, 13 arterial switch d-TGA, and 13 morphologically normal patients. Atrial switch patients demonstrated increased RV EDV relative to normal (322±133 vs 147±57 mL, p< 0.001), but no difference in RV EF (60±15 vs 59±4, p=0.76). Atrial switch patients also demonstrated diminished RV-FWHM and increased RVLV-dTTP relative to normal (198±38 vs 231±32 ms, p< 0.05; 56±24 vs 18±30 ms, p< 0.01, respectively). Arterial switch patients demonstrated increased RV-TTP (211±29 vs 178±38 ms, p< 0.05) and a trend towards an increased RVLV-dTTP (41±27 vs 18±30 ms, p=0.06). Arterial switch patients had neopulmonary stenosis of varying severity, ranging from 3 to 45 mmHg, which correlated strongly with RVLV-dTTP (r=0.70, p< 0.001).

Conclusion: Patients with repaired d-TGA exhibit considerably altered RV systolic waveforms compared to morphologically normal hearts, which we attribute to differences in ventricular afterload. Furthermore, the proposed metric of RVLV-dTTP, for which concurrent measurements are made possible by 4D Flow, exhibited a strong correlation with neopulmonary stenosis, suggesting it may be a sensitive indicator of afterload. Further investigation of the application of RVLV-dTTP to other causes of RV afterload where non-invasive measurements may not be possible, such as pulmonary hypertension, is warranted

Mean 4D Flow waveforms (A) and boxplots of 4D Flow hemodynamic metrics (B) for each d-TGA surgical repair subtype and normal controls. Arterial switch patients were characterized by increased RV-TTP, and a trend towards increased RVLV-dTTP relative to morphologically normal patients. Atrial switch patients were characterized by reduced RV-FWHM and increased RVLV-dTTP relative to morphologically normal patients.

4D Flow hemodynamic metrics for arterial switch patients (n=13) correlate strongly with the neopulmonary stenosis pressure gradient. These findings suggest that the altered RV systolic waveforms for d-TGA patients may be due to differences in afterload.

Mean 4D Flow waveforms (A) and boxplots of 4D Flow hemodynamic metrics (B) for each d-TGA surgical repair subtype and normal controls. Arterial switch patients were characterized by increased RV-TTP, and a trend towards increased RVLV-dTTP relative to morphologically normal patients. Atrial switch patients were characterized by reduced RV-FWHM and increased RVLV-dTTP relative to morphologically normal patients.

4D Flow hemodynamic metrics for arterial switch patients (n=13) correlate strongly with the neopulmonary stenosis pressure gradient. These findings suggest that the altered RV systolic waveforms for d-TGA patients may be due to differences in afterload.