Diagnostic Accuracy of Cardiac MRI for Evaluating Transplant Rejection

A systematic review and meta-analysis found T2 mapping had the highest noninvasive diagnostic accuracy for evaluating acute cardiac transplant rejection. The study findings were published in JACC: Cardiovascular Imaging.

Researchers from Cedars-Sinai Medical Center in the United States searched publication databases through April 1, 2020, for 1.5-T cardiac magnetic resonance imaging (CMR) diagnostic evaluation among patients who underwent cardiac graft. A total of 10 studies published between 2010 and 2019 with single-center, prospective or single-center, retrospective designs were included.

The studies assessed T2 mapping, late gadolinium enhancement (LGE), native T1 mapping, and extracellular volume fraction (ECV). All but 1 study used the International Society for Heart and Lung Transplantation 2005 criteria for rejection.

The average age of the patient cohorts ranged between 34 and 71 years, and 67% were men.

Native T1 was assessed among 267 endomyocardial biopsy (EMB)-CMR pairs. Abnormal T1 ranged between 1022 to 1090 ms. The area under the hierarchical modeling-based summary receiver-operative characteristic (HSROC) curve (AUC) for T1 was 0.84 (95% CI, 0.81-0.87) with a sensitivity of 84.6%, specificity of 70.1%, positive likelihood ratio of 2.83, negative likelihood ratio of 0.22, and diagnostic odds ratio of 12.85.

T2 mapping was analyzed among 276 EMB-CMR pairs. The range of cutoffs for T2 was 48.7 to 59 ms. The HSROC AUC for T2 mapping was 0.92 (95% CI, 0.89-0.94) with a sensitivity of 86.5%, specificity of 85.9%, positive likelihood ratio of 6.14, negative likelihood ratio of 0.16, and diagnostic odds ratio of 39.14.

ECV was evaluated among 132 EMB-CMR pairs and the ECV cutoff range was 26.8% to 31.1%. The HSROC AUC for ECV was 0.78 (95% CI, 0.74-0.81) with a sensitivity of 91.3%, specificity of 67.6%, positive likelihood ratio of 2.82, negative likelihood ratio of 0.13, and diagnostic odds ratio of 21.86.

The studies using LGE comprised 364 EMB-CMR pairs. The range of LGE evaluation was 8 to 10 minutes after contrast. The HSROC AUC for LGE was 0.56 (95% CI, 0.51-0.60) with a sensitivity of 50.1%, specificity of 60.2%, positive likelihood ratio of 1.26, negative likelihood ratio of 0.83, and diagnostic odds ratio of 1.51.

There was significant heterogeneity for the sensitivity and specificity of T1 and LGE and the specificity of T2 (I² range, 60.19%-92.09%; all P £.05).

The meta-analysis was limited, as 2 of the studies were at high risk for patient selection bias and half of studies had uncertain bias risk.

“Myocardial characterization by multiparametric CMR provides high diagnostic accuracy for the detection of acute rejection in heart transplantation,” the study authors noted. “Quantitative T2 assessment demonstrated the highest diagnostic accuracy, followed by quantification of T1 and ECV.”

Reference

Han D, Miller RJH, Otaki Y, et al. Diagnostic accuracy of cardiovascular magnetic resonance for cardiac transplant rejection: a meta-analysis. JACC Cardiovasc Imaging. Published online July 14, 2021. doi:10.1016/j.jcmg.2021.05.008