Incongruent Rejection Appearance by Biopsy and CMR: Adding Value by Adding [Parametric] Values

Description of Clinical Presentation: A 17 year old with heart transplant presented prior to a two-year surveillance visit with glucose abnormalities. Echocardiogram revealed systolic dysfunction with elevated BNP (1547pg/L). Methylprednisolone and milrinone were started and he underwent biopsy. Pathology demonstrated 1R for acute cellular rejection and 0 for antibody mediated rejection (AMR), thus the cause of dysfunction remained unclear. Since biopsy was unrevealing, IVIG was given for possible myocarditis. CMR was requested for evaluation of rejection versus another etiology.

Diagnostic Techniques and Their Most Important Findings: CMR on a 1.5T scanner revealed LVEF 32% and RVEF 31%, bilateral pleural effusions, and hypointensity within the myocardium on bSSFP imaging. Given recent biopsy, inflammation was expected along the septal wall. However, there were several T2 "hotspots" in the LV particularly in the inferior and free walls (T2 of 54-72ms). Significant elevation in T1 and ECV was found in the septum as expected but also in anterior and free walls (ECV of 34-46%). LGE along RV septum was expected but no LV LGE was noted (Fig 1).

He stabilized and was discharged on Sacubitril/Valsartan and Empagliflozin however 4 weeks later he presented to the ED with concerns for rejection, requiring re-induction with ATG and steroids. A repeat CMR (Fig 2) showed similar dysfunction, ECV, and T2 values as prior and troponin was elevated to 36ng/L.

Learning Points from this Case: CMR was extremely useful for supporting rejection with negative biopsy. The incongruent findings confounded the assessment and more aggressive anti-rejection treatment could have been initiated on initial presentation. While biopsy is considered the gold-standard, in our case the gold-standard failed and CMR was integral to evaluation. We considered that RV inflammation would be associated with recent biopsy, findings in the LV made rejection much more concerning. The elevation in T2 (54-72ms) was localized to hotspots, concerning for rejection and was normal in the inferior (42ms) and lower lateral wall (46ms). ECV and T2 mapping have been evaluated extensively in rejection surveillance. Increased ECV is a useful measure of fibrosis as it is usually elevated from excessive collagen.¹

ECV should be near normal or only mildly elevated in the absence of rejection in a transplanted heart as rejection is best detected with a cutoff of 26.8% to 31.1%² whereas our case demonstrated significant elevations far from the biopsy site. T2 mapping has the highest accuracy, followed by T1, ECV, and LGE² and we felt the degree of elevation was consistent with rejection and lack of LGE was consistent with its limited utility. We were mindful that ECV has high sensitivity, but overall performance can be limited by low specificity,² which may have contributed to treatment delay and consideration of alternate diagnoses given the negative biopsy. T2 mapping was most helpful as T2 values > 56ms have been associated AMR,³ and, specifically in children, a T2 cutoff of 54.5ms can detect rejection regardless of T1 values.⁴

Figure 1: Panels demonstrating bilateral pleural effusions (A), and abnormal hypodense appearance of myocardium along right ventricular aspect of intraventricular septum (B, arrows) both on bSSFP imaging, and significant late gadolinium enhancement in the right ventricle, parietal band, and RV septal surface without presence in the LV (C, D, arrows) on LGE imaging.
Figure 2: Panels demonstrating persistent and increased bilateral pleural effusions (A) and abnormal hypodense appearance of myocardium has increased along right ventricular aspect of intraventricular septum (B, arrow) on bSSFP imaging, and late gadolinium enhancement in the right ventricle has improved in the RV septum but may be appearing along the LV free wall subepicardium (C, D, arrows) on LGE imaging.