Measuring Atrioventricular Valve Regurgitation in Fontan Patients Using Cardiac MRI and the Association with Clinical Outcomes: A FORCE Study. (RF_FR_387)

Presenting Author(s)

• YL

  Yu Li, MD, MSc

  Resident Physician
  UPMC Children's Hospital of Pittsburgh

Primary Author(s)

• YL

  Yu Li, MD, MSc

  Resident Physician
  UPMC Children's Hospital of Pittsburgh

Co-Author(s)

• MS

  Mary Schiff, PhD, MPH

  Senior Statistician
  Children's Hospital of Pittsburgh of UPMC
• 

  Laura Olivieri, MD

  Associate Professor of Pediatrics
  Children's Hospital of Pittsburgh of UPMC
• 

  Vivek Muthurangu, MD

  Professor
  University College London, United Kingdom
• 

  Timothy Slesnick, MD

  Pediatric Cardiologist
  Children's Healthcare of Atlanta
• AC

  Adam B. Christopher, MD

  Assistant professor
  UPMC Children's Hospital of Pittsburgh
• VM

  Victor Morell, MD

  Professor of Pediatric Cardiothoracic Surgery
  UPMC Children's Hospital of Pittsburgh
• MC

  Mario Castro Medina, MD

  Assistant Professor of Cardiothoracic Surgery
  Children's Hospital of Pittsburgh of UPMC
• 

  Yue-Hin Loke, MD

  Associate Professor
  Children's National Medical Center
• 

  Rahul H. Rathod, MD, MBA

  Associate Chair of Cardiology
  Boston Children's Hospital
• JK

  Jacqueline Kreutzer, MD

  Professor
  Children's Hospital of Pittsburgh
• 

  Tarek Alsaied, MD

  Assistant Professor of Pediatrics
  Children's Hospital of Pittsburgh

Background: In patients who undergo the Fontan procedure, atrioventricular valve regurgitation (AVVR) is negatively associated with ventricular function and long-term survival. Echocardiographic surveillance of AVVR is more readily available but is more subjective compared to CMR measurements. Limited data exist to compare the modes of AVVR measurements in Fontan patients. This study aimed to compare CMR and echo AVVR measurements and investigate their associations with long-term clinical outcomes using data from the multi-institutional Fontan Outcome Registry Using CMR Examination (FORCE).

Methods: We identified 3,360 post-Fontan CMR measurements of AVVR using at least one of the four methods: (1) calculated by antegrade aortic flow and machine learning generated stroke volume, (2) calculated by manual CMR antegrade aortic flow and stroke volume, (3) CMR regurgitant fraction (RF) from the report using 2D AV valve flow and aortic antegrade flow, and (4) impression from CMR reports which combines the above methods. The AVVR was graded “none to mild” if RF was 0-20%, “moderate” if RF was 20-40%, and “severe” if >40%. Each CMR was matched with an echo performed within two years. Kappa statistics were used to study the concordance between CMR and echo measurements. A composite outcome was identified as the earliest occurrence of death, listing or receiving a heart transplant, or atrial arrhythmia. Multivariable Cox proportional hazard models were used to compare the associations between AVVR measurements using echo and four CMR methods and the clinical outcome.

Results: The median age at CMR was 14.2 years [IQR 10.2, 18.9] (42% female). Patients identified to have severe AVVR were less likely to have a left dominant ventricle, lower ejection fraction, and higher indexed EDV, ESV, and ventricular mass (Table 1). Concordances between CMR methods and echo grading of AVVR were only slight to moderate, and were higher using CMR reported RF (kappa = 0.46) and CMR reported impression (kappa = 0.56) (Table 2). AVVR determined by CMR machine learning and calculated methods were less significantly associated with the clinical outcomes. The composite outcome was significantly associated with AVVR severity using echo, CMR reported RF, and CMR reported impression methods (Figure 1). CMR RF from the report included the fewest patients but showed the greatest association with the clinical outcome, as patients who had moderate (HR=2.25, 95% CI: 1.30-3.90, p=0.004) and high (HR=7.28, 95% CI: 3.84-13.81, p< 0.001) AVVR had significantly increased hazard of the clinical outcome, after covariate adjustment.

Conclusion: AVVR was identified using four different CMR methods. The reported impression method, which incorporated all CMR measures, correlated the most with the echo report. AVVR by CMR reported RF, which uses AVV 2D flow, revealed the greatest risk of long-term adverse outcomes. The calculation using stroke volume (both manual and machine learning) and aortic antegrade flow has the least correlation with echocardiogram and clinical outcomes.

Comparison of the Association Between AV Valve Regurgitation and Composite Outcome Using CMR and Echocardiographic Methods. aNone to mild AVVR was used as the reference level. bAdjusted for sex, dominant ventricular morphology, and Fontan type.