Abstract
Background: Heart transplantation is the treatment of choice in select patients with advanced heart failure. Since its introduction in the early 1970s, it has delivered significant benefits in survival and quality of life of these patients. However, despite significant progress in the selection of patients, technical aspects of the procedures and long-term outcomes, HT recipients face a significant risk of future complications, namely graft rejection and cardiac allograft vasculopathy, which warrant regular invasive monitoring with cardiac catheterisation. Stress perfusion cardiac magnetic resonance imaging (stress CMR) is recommended in the international guidelines in class I in patients with known or suspected coronary artery disease, but its role in post-HT surveillance is yet to be established. We sought to assess the safety and feasibility profile, as well as the diagnostic and prognostic yield and the overall clinical utility of quantitative stress CMR in HT recipients referred for assessment of cardiac allograft vasculopathy (CAV).
Methods: A retrospective study of all orthotopic HT recipients undergoing clinically indicated stress perfusion CMR with adenosine was conducted. The safety and feasibility profile was demonstrated by evaluating any adverse events during or immediately following adenosine infusion and the elicited change in haemodynamic parameters. The diagnostic endpoint was area under the curve (AUC) for angiography-defined graft vasculopathy. The utility endpoint was the net benefit purported by undergoing stress CMR as a gatekeeper to angiography at a prespecified threshold range (10-40%). The prognostic endpoint was time to occurrence of a major adverse cardiac event (unplanned cardiac hospitalization or death from any cause).
Results: 60 HT recipients were identified having stress CMR and angiography. Prevalence of CAV was 30%. Adenosine stress CMR was safe with no serious adverse events noted. Myocardial perfusion reserve index had an AUC for significant CAV of 0.79, with sensitivity 85.7%, specificity 70.3%, positive predictive value of 52.2% and negative predictive value of 92.9% at a Youden threshold of 2.1. A total of 11 patients experienced MACE at a median follow-up of 1.8 years. Myocardial perfusion reserve index, along with stress MBF, was a significant risk predictor. As a gatekeeper to invasive testing, stress CMR with quantitative perfusion showed superior decisional performance than stress CMR alone, allowing as much as 39% of patients to properly avoid angiography.
Conclusions: Stress CMR with quantitative perfusion showed excellent safety and feasibility in HT recipients, and superior accuracy, prognostic yield and decisional performance than stress CMR alone. These findings warrant prospective confirmation.
Outline
Stress cardiac magnetic resonance (CMR) perfusion imaging is recommended in international guidelines as class I in patients with known or suspected coronary artery disease1, and is a promising non-invasive technique for the detection of cardiac allograft vasculopathy (CAV) of cardiac transplantation recipients2. Yet, there are concerns regarding the use of vasodilators after heart transplantation3. Further, the clinical utility of stress CMR in cardiac transplantation recipients has not been formally investigated. Finally, the comparative diagnostic accuracy and prognostic yield of fully quantitative versus qualitative stress perfusion CMR in cardiac transplantation have not been previously established. Accordingly, we aimed to assess 1) the safety and feasibility profile, 2) the diagnostic and prognostic yield, and 3) overall clinical utility of adenosine stress CMR perfusion in HT recipients with suspected CAV. This thesis comprises three parts. In chapter 1, a literature review is presented on the contemporary use and outcomes of heart transplantation, and current strategies for quality optimisation in long-term CAV surveillance are briefly outlined. Results of a systematic review on use of stress CMR after heart transplantation are presented. In chapters 2–3, the methods and findings of the study are shown. In chapters 4–6, results are discussed and put into the broader context of research on post-transplantation surveillance; study limitations are addressed, and future developments proposed. Finally, the Appendix addresses some open queregarding the haemodynamic response to pharmacological vasodilation in the transplanted heart, and its implications for the measurement and clinical value of myocardial blood flow.
Introduction
Cardiac transplantation
Cardiac transplantation is the selected destination treatment in patients with advanced heart failure4. This procedure, introduced in the late 1960s and originally looked upon with skepticism and suspicion, offers unparalleled improvement in both prognosis and quality of life in patients with ongoing symptoms, its use increasing over time and with improving outcomes5 (Figure 1). Patients referred for cardiac transplantation are required to undergo a significant amount of preprocedural testing4 aimed at identifying potential contraindications. Patients deemed fit for transplantation typically undergo orthotopic heart transplantation, in which the native ventricles and all cardiac valves are excised, and anastomoses are created between the atria, aorta, and pulmonary valves of the donor and the recipient; in contrast, in patients undergoing heterotopic heart transplantation, the recipient heart is left in place and connected to the donor heart. As short-term survival improves (~85-90% at 1 year, and 75% at 5 years5), a growing number of patients are now followed up well beyond the 10th, 20th, or even 30th year since transplantation, which require extensive follow-up for the prevention of long-term complications.
Complications following cardiac transplantation
Notwithstanding the inherent risks of cardiac surgery, heart transplantation recipients are at risk of early- and late-onset complications, which will be briefly outlined (Figure 2). Cardiac allograft vasculopathy, a late-onset complication, will be discussed in detail in a separate section.
In the early days up until the first years following cardiac transplantation, the graft recipient is at risk of developing heart failure and acute rejection, which is the leading cause of death in this period. Immunosuppressive therapy which is required for control of the immune response leads to opportunistic infections which typically occur between 30 days and the first year from the procedure. Cardiac allograft vasculopathy is commonly found in patients at > 2 years from transplantation and is partially addressable with pharmacological treatment. Other late complications of cardiac transplantation include malignancies – mainly as a consequence of viral infections, chronic kidney disease (CKD), and the onset of diabetes, hypertension, dyslipidemia, depression and anxiety, gout, and osteoporosis6.
Primary graft dysfunction
This is defined as the failure of graft function in the first 24 hours after transplantation, which is not caused by concomitant rejection, pulmonary hypertension or procedural complications. Primary graft dysfunction has variable incidence (rates of 3-28% have been reported4) and is associated with poor short-term survival. It manifests itself with diastolic dysfunction and a restrictive pattern, associated with atrioventricular and aortic valve regurgitation which is typically asymptomatic.
Graft rejection
Rejection caused by activation of the recipient immune system, and categorised as antibody- or cell-mediated based on the predominant component in the immune response. While cell-mediated rejection occurs acutely in the first after transplantation, antibody-mediated rejection may be developed later on. Incidence of rejection is estimated at 13% in the first year since discharge, and it is recognized a major contributor to mortality7 in the first three years from transplantation. Detection of rejection requires patients to regularly undergo endomyocardial biopsy, and histological grading is based on criteria proposed by the International Society for Heart and Lung Transplantation (ISHLT)8.
Infections
As a consequence of the use of immunosuppressive therapy, infections are the first cause of death between 30 days and 1 year since transplantation9. After the first year, immunosuppression is typically gradually decreased, but infections remain among the top three causes of death in the long term, with an attributable mortality of > 10%7.
Other late complications
New onset of a malignancy following cardiac transplantation is estimated at 11% between 1 and 5 years from the procedure4 and is mainly related to viral infections from agents such as the human papilloma virus, Epstein-Barr virus, and human herpes virus 8. Other common late-onset complications include diabetes, hypertension, dyslipidemia, and renal dysfunction, as well as psychological disorders, gout, and osteoporosis6.
Cardiac allograft vasculopathy
Cardiac allograft vasculopathy (CAV) has been labelled the «Achilles heel» of cardiac transplantation10. It consists in the diffuse intimal thickening and luminal narrowing in the coronary tree, possibly leading to focal occlusion and heart failure, with no or limited development of new collateral vessels11. Patients with CAV are often asymptomatic due to graft denervation2,11, and both epicardial12 and microvascular disease13 were separately found to be prognostic. CAV is believed to be the result of an interplay of immunological and non-immunological factors which contribute to lipid deposition and proliferation of smooth muscle cells and macrophages in the intima2 of both epicardial vessels and the coronary microcirculation (Figure 3).
Risk of CAV is associated with graft age, with incidence estimated at 8% at 1 year and 47% at 10 years4. The Stanford classification for CAV is descriptive of the morphological features at angiography. It distinguishes discrete lesions (type A) from distal diffuse concentric narrowing in main vessels (type B) or distal branches (type C)11. CAV grading is based on criteria proposed by the Cardiac Transplant Research Database, a multicenter registry that collected data from 4,637 invasive angiograms across 39 institutions in the 1990s14. This grading has been embraced by the ISHLT10, and defines CAV as pictured in Figure 4.
Additionally, CAV should be managed in light of the clinical context. Rapidly progressing CAV in the first 2 years since transplantation is believed to be mainly driven by vasculitis and should be managed more aggressively due to its poor prognosis. Instead, a later (> 2 years) presentation in the presence of preserved graft function is typical of an indolent course. Finally, rapidly progressive CAV is defined as new-onset CAV within 1 year of a previous normal angiogram; this is also associated with poor prognosis10.
Due to the diffuse involvement of the coronary tree, revascularization is rarely an option in patients with significant (i.e. moderate or severe) CAV15 and is ultimately considered a palliative treatment with no survival benefit16. However, pharmacological treatment has been proven useful to hamper disease progression. Use of pravastatin reduced progression in intimal thickening in an open-label, randomized clinical trial17. Inhibitors of the mammalian target of rapamycin such as sirolimus have also shown to reduce CAV progression18.
Established and proposed strategies for CAV surveillance
Several single-modal or multimodal imaging strategies have been proposed for CAV detection and surveillance following cardiac transplantation. The benefits and drawbacks for each modality will be briefly outlined; a subsequent, specific section will discuss use of stress CMR imaging.
Invasive coronary angiography (ICA)
Due to the extensive evidence available for its use, ICA is considered the gold standard for CAV detection10. ICA is highly accessible to both pediatric and adult patients. However, routine, repeated use of invasive coronary angiography for CAV detection since the first years following cardiac transplantation has long been questioned15,19,20. Arguments against the use of invasive coronary angiography include the immediate risk of harm and radiation exposure, the palliative nature of revascularization – CAV being a diffuse, slowly progressing disease – along with the reported reduced sensitivity when compared with other techniques, notably intravascular ultrasound21. Despite these considerations, ICA remains the recommended procedure for CAV detection and surveillance.
Intravascular ultrasound (IVUS)
Use of IVUS has long been investigated in the transplant outpatient setting22, helping to shed light on CAV pathophysiology10. The 2010 ISHLT statement on CAV shared consensus that IVUS is most useful for its negative predictive value10. Additionally, despite recognizing its high sensitivity, the panel deemed use of IVUS as «investigational» and advocated that ICA should be instead used to guide patient management. However, IVUS has been showed to effectively select patients benefiting from subsequent routine invasive angiography23.
Dobutamine stress echocardiography
Echocardiography with dobutamine stress was found to be prognostic24 and predictive of CAV25 among heart transplant recipients, with impairment in strain parameters observed even in the presence of mild CAV26. However, when compared to CTCA against a reference of ICA, it was shown to be less sensitive27; later on, larger studies were conducted showing variable accuracy28–30 which culminated in a systematic review and meta-analysis finding highly variable sensitivity and variable specificity31. Authors concluded that the modality would be unlikely to yield enough positive and negative predictive value to effectively gatekeep patients to angiography.
Positron emission tomography and single photon emission CT
Positron emission tomography (PET) and single photon emission CT (SPECT) are perfusion imaging modalities variably implemented in the follow-up of cardiac transplantation. Quantification of MBF by PET was diagnostic for CAV32 and prognostic at mid-term follow-up33. A later study showed superior discriminative ability of MPR over stress MBF34. Similarly, SPECT was found to be diagnostic against ICA35, but later reports found very low (< 10%) sensitivity36 and overall unacceptable accuracy37, warranting against use of SPECT to gatekeep angiography. Finally, a SPECT system utilizing cadmium-zinc-telluride crystals, allowing for direct measurement of MBF and MPR, was recently validated against PET in heart transplant recipients38, but the clinical utility of such technological advance remains unestablished39.
CTCA
To date, several studies have assessed the diagnostic value of CTCA for CAV detection. Upon comparison with coronary MR angiography, it showed superior performance40, and protocols with low radiation burden were validated in transplant recipients41–43. In a longitudinal outcomes study, a strategy of CTCA appeared safe and feasible in retrospect44 and use of the modality gained momentum45,46, despite the lack of high-level evidence (e.g. randomized trials, prospective comparative trials). Upon diagnostic meta-analysis, CTCA yielded high accuracy, with sensitivity and specificity > 90% for 64-slice CTCA47. In a study on real-world implementation, CTCA granted excellent image quality in heart transplant recipients, however, in 25% of patients the scan was not deemed safe due mainly to renal impairment48. Notably, the literature on accuracy of CTCA mainly focused on epicardial CAV with ICA as the reference test, and no comparative studies assessed the modality against IVUS. A recent study49 evaluated the combination of anatomical assessment of CTCA with CT-based myocardial perfusion imaging, showing feasible microvascular evaluation with CT. Overall, the literature on CTCA for CAV showed safety and feasibility with some remaining concerns of real-world application. Additionally, use of CTCA for CAV surveillance would require a separate test for detecting allograft rejection.
CMR for CAV surveillance
We conducted a systematic literature review to appraise the investigation of CMR in the follow-up of heart transplant recipients (PROSPERO: CRD42023435640). Over the last two decades, several studies have evaluated CMR in the context of allograft rejection50–62 and CAV63–74. In total, 15 studies were identified regarding use of CMR in CAV (Table 1).
| Study | Sample size | CMR module(s) | Main Reference | Main finding |
|---|---|---|---|---|
| Muehling et al., 200375 | 27 | Stress perfusion (semi-quantitative) | ICA | MPR index significantly decreased in heart transplant recipients compared with healthy controls |
| Korosoglou et al., 200963 | 69 | Strain, stress perfusion (semi-quantitative) | ICA | MPR, diastolic strain rate have high accuracy* |
| Nunoda et al., 201040 | 22 | Whole-heart MR angiography | ICA | MR angiography has limited sensitivity |
| Colvin-Adams et al., 201164 | 68 | Stress perfusion (qualitative) | ICA | Visual assessment has limited sensitivity |
| Machida et al., 201267 | 38 | Cine, LGE | IVUS | CAV associated with LV diastolic dysfunction |
| Hussain et al., 201366 | 24 | Vessel wall LGE | IVUS | Vessel LGE associated with IVUS-derived intimal measures |
| Machida et al., 201365 | 46 | Cine, LGE | IVUS | LV peak filling rate accurate to detect CAV |
| Braggion-Santos et al., 201468 | 132 | Cine, LGE | ICA | LGE may guide risk stratification in suspected CAV |
| Miller et al., 201454 | 48 | Cine, stress perfusion (fully quantitative), parametric mapping | IVUS | MPR outperforms ICA for CAV detection on IVUS |
| Mirelis et al., 201570 | 8 | Stress perfusion (semi-quantitative) | IVUS | MPR index detects microvascular dysfunction |
| Chih et al., 201632 | 29 | Stress perfusion (semi-quantitative) | IVUS | MPR index inversely correlated with maximal intimal thickness, has high rule-out ability |
| Erbel et al., 201672 | 63 | Strain, stress perfusion (semi-quantitative) | IVUS, biopsy | MPR index correlated with microvessel luminal radius, detects high-risk patients with no overt CAV |
| Narang et al., 201876 | 20 | Stress perfusion (semi-quantitative) | ICA | MPR index associated with severity of CAV |
| DeSa et al., 202073 | 20 | Stress perfusion (semi-quantitative) | ICA | MPR index associated with presence of CAV |
| van Heeswijk et al., 202077 | 20 | LGE, parametric mapping | ICA, OCT, biopsy | Intima-media thickness ratio associated with increased fibrosis burden on biopsy and extracellular volume fraction on CMR |
| Abbasi et al., 202274 | 77 | Cine, parametric mapping | ICA | Patients with any CAV (grade 1-3) show reduced LV ejection fraction, increased global and segmental T2 values |
Over the course of 20 years, 15 studies evaluated the use of various CMR components to adjudicate presence or severity of CAV. Sample size refers to the number of heart transplant recipients enrolled to a study.
Methods
Design
This was a retrospective observational study of all outpatients under surveillance for CAV after orthotopic cardiac transplantation at a large tertiary center (Harefield Hospital, London, United Kingdom), in which stress CMR was clinically indicated by the referring cardiologist. The study included a cross-sectional comparison between stress CMR results and the closest angiographic study findings (ICA or, when unavailable, CTCA), as well as a longitudinal component with follow-up for major adverse cardiovascular events. This study received approval of the local institutional review board at Royal Brompton and Harefield Hospitals, Guys’ & St Thomas NHS Trust. Patient informed consent was waived due to the retrospective nature of the study, which did not affect patient management in any way.
CMR studies performed on orthotopic heart transplant recipients at Royal Brompton and Harefield hospitals between 2015 and 2023 were included if the patient had undergone a complete adenosine stress perfusion CMR study of diagnostic quality. N = 70 stress CMR studies were performed on 60 individual patients; for analyses other than the safety profile, only the index study was considered.
Study cases
All stress CMR studies conducted at Royal Brompton and Harefield Hospitals between 2015 and 2023 were reviewed, and orthotopic cardiac transplantation outpatients were included if they underwent a complete stress perfusion CMR study deemed of diagnostic quality, with adenosine as the vasodilator. For the safety analysis, all eligible studies were included, whereas for other analyses the index stress CMR study was included and all subsequent stress CMR studies discarded. For each study patient, baseline characteristics were collected, including age, sex, BMI, BSA, cause for transplantation, date of transplantation and graft ischemic time, along with presence of diabetes mellitus, hypertension, dyslipidemia, smoker status, chronic kidney disease, thyroid dysfunction, chronic obstructive pulmonary disease, tricuspid regurgitation, atrial fibrillation, atrial flutter or other atrial arrhythmias, history of venous thromboembolism, history of pacemaker/defibrillator implantation, autoimmune diseases, anxiety and depression at the time of the stress CMR study. Use of immunosuppressant or other cardiac medications at time of CMR was also noted.
Procedures and tests
CMR
The multi-parametric CMR (1.5 Tesla Siemens Sola scanner; Erlangen, Germany) protocol for CAV assessment comprised adenosine stress perfusion, rest perfusion, cine imaging for LV function and early (EGE) and late gadolinium enhancement (LGE). CMR data were analysed by level-2 or level-3 readers, using visual analysis according to the standard AHA/ACC 17-segment model. For visual interpretation of the perfusion study («qualitative» stress CMR), motion corrected images were utilized where available, and significant ischemia was reported in the presence of a stress-induced perfusion defect comprising at least 1.5 segments. An automated in-line quantitative perfusion mapping algorithm was utilized to derive global and segmental values for stress and rest myocardial blood flow (MBF) and myocardial perfusion reserve (MPR) («quantitative» stress CMR)78. Global values for stress and rest MBF in mL/min/g, and MPR were collected. Rest MBF and MPR were then indexed to the rest rate-pressure product (heart rate · systolic blood pressure, both collected at baseline). The rate-pressure product is an index of haemodynamic response which has been linked to myocardial oxygen consumption79 and is utilized in nuclear imaging studies to account for differences in haemodynamic conditions80}, particularly heterogeneous in the context of heart transplant81}. All CMR studies were anonymized and re-analyzed specifically for the present investigation, and readers were blinded to angiographic results or other clinical information. Analysis of CMR studies was conducted on Circle CVi42 software workstations (Circle Cardiovascular Imaging Inc., Calgary, Canada). For each study, left and right ventricular, end-diastolic and end-systolic volumes were measured, along with left ventricular, right ventricular and left atrial ejection fraction, and lateral mitral (MAPSE) and tricuspid (TAPSE) annular plane systolic excursion. Volumes were indexed to the body surface area, and LV mass was also measured. Cardiac output and cardiac index were also derived. Myocardial strain was measured with global longitudinal, radial, and circumferential strain. Presence of LGE was assessed visually and, for patients with any LGE, the LGE mass was measured with the 5-SD method. LGE mass was then indexed to LV mass.
Angiography
For each study case, the closest angiographic study was identified and its date and findings collected. For cases where ICA was not performed, the closest CTCA was utilized as a surrogate. CAV grading was accepted as reported or, where not formally reported, derived based on aforementioned ISHLT criteria10. Significant CAV was defined as CAV grade 2 or 3, in accordance with previous findings of no significant differences in morphofunctional features between patients with no or mild CAV74, and owing to the definition of CAV grade 3, where independently of the severity of epicardial lesions, a severe disease is adjudicated in patients with LV systolic dysfunction. For patients with CAV, lesion sites and, for ICA, stent deployment were noted.
Follow-up
The study prognostic endpoint was time to the first major adverse cardiovascular event (MACE), defined as unplanned cardiovascular hospitalization or death from any cause. Causes and dates for unplanned cardiovascular hospitalizations and death were ascertained from hospital electronic health records accessed until 31st December 2023.
Data collection
All data relevant to the present study were anonymized upon collection and uploaded on an encrypted electronic data capture server (REDCap, Vanderbilt University, Nashville, USA). Anonymized CMR data, including results from analysis of all CMR modules, were exported to .XML files. A custom-made Shiny app was developed for extracting data from .XML files and batch transfer them into database entries. Additional study data, including baseline characteristics, angiography results and at during follow-up, were searched in electronic health records and collected by authorized staff (clinical CMR fellows, Harefield Hospital). Baseline characteristics included transplant date and cause, type of heart transplant (orthotopic, heterotopic, or heart-lung), presence of diabetes mellitus, hypertension, dyslipidemia at the time of CMR; smoker status, chronic kidney disease, atrial fibrillation or other atrial arrhythmias. Blood tests at time of CMR, including creatinine and estimated glomerular filtration rate, total cholesterol, HDL cholesterol, and glycated haemoglobin, were also noted. For each study case, the closest-in-time circulating donor-specific antihuman leukocyte antigen antibodies (DSA) blood test was identified, and its date and results noted.
Statistics
Baseline characteristics were summarised using descriptive statistics. Wilcoxon rank sum test, Fisher’s exact test, Student’s, and Pearson’s Chi-squared test were used where appropriate. Associations between variables were assessed by linear or logistic regression. Statistical analysis was undertaken using R software, version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria) at a 5% significance level. Statistical packages utilized are detailed in the Appendix.
Diagnostic accuracy
Sensitivity, specificity, positive and negative predictive values (PPV, NPV), and area under the curve (AUC) were determined for qualitative stress CMR results and quantitative CMR parameters to detect significant CAV on index angiography (ICA or CTCA). Patients were included in the diagnostic analysis if they had complete data from both stress CMR and angiography (ICA or, when unavailable, CTCA). The Youden index was used to adjudicate proposed optimal thresholds for continuous stress CMR parameters to detect significant CAV on angiography.
Clinical utility
Decision curve analysis was undertaken to assess for clinical utility of qualitative and quantitative stress CMR across a pre-specified range of threshold probabilities (10% to 40%) considered representative of a sufficiently wide range of clinical scenarios82. The clinical utility of each test was expressed as 1) the net benefit of undergoing CMR rather than referring the patient directly to invasive angiography; and 2) the net reduction in unnecessary angiographies performed83, both calculated at each threshold probability in the pre-specified range. Explanation of decision curves and their analysis and interpretation is offered in the Results section. To account for model overfit, ten-fold cross-validation at 25 repeats was utilized.
Prognostic yield
Survival models were constructed based on CMR results and parameters, and Kaplan-Meier curves were used to estimate survival probabilities and compare MACE rates. The log-rank test was used to compare survival curves. Cox proportional hazards regression models were used to estimate hazard ratios (HRs) and 95%CIs for the association between test results and MACE. Restricted cubic splines were employed to model relationships between continuous variables and HRs.
Results
Out of 12,755 stress perfusion CMR studies performed at Royal Brompton and Harefield hospitals between 1st January 2015 and 31st December 2023, 70 (0.5%) were adenosine stress CMR studies performed in 60 outpatient heart transplant recipients (31.7% females) at a median of 18.278 years after transplantation (IQR 3.1-23.9); 25% of studies were conducted within the first three years post-transplantation. Median (IQR) time between CMR and angiography was 14.3 (25.4) months and time from CMR to serum testing for positivity to DSA was 36 (11-99) days.
Baseline characteristics
Median (IQR) patient age at CMR was 52 years (35 to 60), with 41 (68%) males. Main reason for cardiac transplantation was dilated cardiomyopathy (n = 26, 53%). Hypertension was highly prevalent at 45%, while 28% had a dyslipidemia and 25% of patients had diabetes mellitus. Very few were smokers (n = 2, 3.3%). Atrial fibrillation was reported in 6.6%. None of these characteristics differed significantly between patients with no or mild CAV and patients with moderate or severe CAV. Detailed characteristics of included patients are detailed in Table 2.
| Characteristic | Overall, N = 601 | CAV 0-1, N = 421 | CAV 2-3, N = 181 | p-value2 |
|---|---|---|---|---|
| Sex | 0.7 | |||
| Male | 41 (68%) | 28 (67%) | 13 (72%) | |
| Female | 19 (32%) | 14 (33%) | 5 (28%) | |
| Reason for transplantation | 0.4 | |||
| Congenital heart disease | 3 (6.1%) | 3 (8.3%) | 0 (0%) | |
| Myocarditis | 5 (10%) | 3 (8.3%) | 2 (15%) | |
| Familial Dilated CMP | 26 (53%) | 19 (53%) | 7 (54%) | |
| Hypertrophic CMP | 3 (6.1%) | 3 (8.3%) | 0 (0%) | |
| ACM | 0 (0%) | 0 (0%) | 0 (0%) | |
| Ischemic heart disease | 3 (6.1%) | 1 (2.8%) | 2 (15%) | |
| Amyloidosis | 1 (2.0%) | 1 (2.8%) | 0 (0%) | |
| Other Infiltrative CMP | 3 (6.1%) | 3 (8.3%) | 0 (0%) | |
| Post-partum CMP | 3 (6.1%) | 2 (5.6%) | 1 (7.7%) | |
| Arteritis | 1 (2.0%) | 1 (2.8%) | 0 (0%) | |
| Other | 0 (0%) | 0 (0%) | 0 (0%) | |
| Not specified | 1 (2.0%) | 0 (0%) | 1 (7.7%) | |
| Diabetes mellitus | 15 (25%) | 11 (26%) | 4 (22%) | >0.9 |
| Hypertension | 27 (45%) | 16 (38%) | 11 (61%) | 0.10 |
| Dyslipidemia | 17 (28%) | 10 (24%) | 7 (39%) | 0.2 |
| Smoker | 0.5 | |||
| Former | 2 (3.3%) | 1 (2.4%) | 1 (5.6%) | |
| Never | 58 (97%) | 41 (98%) | 17 (94%) | |
| Atrial fibrillation | 0.8 | |||
| None | 56 (93%) | 39 (93%) | 17 (94%) | |
| Paroxysmal | 2 (3.3%) | 1 (2.4%) | 1 (5.6%) | |
| Persistent or permanent | 2 (3.3%) | 2 (4.8%) | 0 (0%) | |
| Chronic kidney disease (eGFR < 60 ml/min) | 30 (79%) | 22 (79%) | 8 (80%) | >0.9 |
| 1 n (%) | ||||
| 2 Pearson’s Chi-squared test; Fisher’s exact test | ||||
Angiographic findings
Patients free from CAV on angiography were 24 (40%), leaving the majority of patients with mild to severe disease (n = 36, 60%), Prevalence of significant (e.g. ISHLT grade 2-3) CAV on index angiography was 30% (n = 18 patients). While 46.667% had ICA results available, 53.333% only had CTCA results. Patients referred to CTCA had significantly lower incidence of CAV (61% patients with no CAV on CTCA vs 14% on ICA, p < 0.001), indicating referral bias towards CTCA for low-risk patients. Still, of 28 patients with ICA available, 10 (35.7%) had non-significant CAV. Findings on ICA or CTCA for included patients are detailed in Table.
CMR findings
At CMR, mean (SD) left ventricular ejection fraction was 61.7% (12.5%), with seven patients (11.7%) having an ejection fraction <45%. Late gadolinium enhancement was reported in 23 (38.3%) patients, in most of which the scar burden was minimal at 2.91 (5.39) grams (2.25% as a fraction of LV mass). CMR findings are detailed in Table 3.
| Characteristic | Overall, N = 601 | CAV 0-1, N = 421 | CAV 2-3, N = 181 | p-value2 |
|---|---|---|---|---|
| BSA | 1.92 (1.82, 2.08) | 1.92 (1.82, 2.10) | 1.94 (1.84, 2.05) | 0.9 |
| LVEDVi (ml/m2) | 70 (61, 79) | 69 (60, 79) | 73 (64, 81) | 0.6 |
| LVESVi (ml/m2) | 25 (20, 32) | 23 (19, 30) | 33 (21, 42) | 0.059 |
| LVEF (%) | 64 (55, 71) | 66 (59, 71) | 58 (47, 64) | 0.023 |
| LVCI (l/min/m2) | 3.47 (3.09, 4.22) | 3.56 (3.10, 4.33) | 3.29 (3.09, 3.65) | 0.2 |
| RVEDVi (ml/m2) | 75 (68, 90) | 77 (69, 89) | 72 (66, 89) | 0.7 |
| RVESVi (ml/m2) | 34 (28, 42) | 35 (28, 41) | 33 (28, 44) | 0.9 |
| RVEF (%) | 57 (51, 61) | 57 (53, 62) | 56 (50, 60) | 0.2 |
| RVCI (l/min/m2) | 3.48 (2.66, 4.27) | 3.53 (2.96, 4.27) | 3.23 (2.58, 4.24) | 0.4 |
| MAPSE Lateral | 12.7 (9.7, 15.1) | 12.8 (10.2, 15.1) | 11.7 (9.0, 14.8) | 0.5 |
| TAPSE | 12.1 (8.0, 16.2) | 12.1 (8.5, 16.5) | 11.0 (6.8, 15.2) | 0.2 |
| LAX Global Longitudinal Strain | -12.7 (-16.0, -9.6) | -13.4 (-17.4, -10.6) | -12.1 (-13.6, -9.0) | 0.15 |
| LAX Global Radial Strain | 19 (15, 27) | 20 (16, 32) | 19 (14, 22) | 0.11 |
| LA EF (%) | 33 (21, 48) | 40 (25, 48) | 25 (13, 33) | 0.074 |
| Hypokinesia | 7 (12%) | 3 (7.1%) | 4 (22%) | 0.2 |
| Dyskinesia | 2 (3.3%) | 1 (2.4%) | 1 (5.6%) | 0.5 |
| Akinesia | 0 (0%) | 0 (0%) | 0 (0%) | |
| LGE presence | 25 (42%) | 16 (38%) | 9 (50%) | 0.4 |
| LGE mass (%) | 2.3 (1.2, 5.3) | 1.9 (1.2, 2.7) | 5.3 (1.1, 6.9) | 0.3 |
| 1 Median (IQR); n (%) | ||||
| 2 Wilcoxon rank sum test; Wilcoxon rank sum exact test; Fisher’s exact test; Pearson’s Chi-squared test | ||||
Stress CMR findings
Feasibility
Adenosine dosage of 140 µg/kg/min was utilized in the majority of studies (n = 37, 62%) while 210 µg/kg/min was reached in 8 patients (13%). During adenosine infusion, the haemodynamic response was highly variable (Figure 7). Heart rate increased from a mean ± SD of 82 ± 14 to 91 ± 18 beats per minute, while systolic blood pressure decreased from 132 ± 16 to 124 ± 14 mm Hg. Overall, 50.1% patients reported chest tightness or chest pain during infusion, 22.4% reported shortness of breath, 17.2% did not experience any symptoms and 10.3% had hot flush. Haemodynamic and symptom response were not dependent on adenosine dosage subgroup (all p-values > 0.5). Stress CMR findings are detailed in ?@tbl-stress.
Safety
No life-threatening adverse events, nor brief or prolonged atrioventricular block, or major arrythmias occurred during or immediately after adenosine infusion, with one patient having self-reversed sinus pauses. Mild, transient adverse effects were registered: three patients (4.5%) complained of headache, and another reported nausea. Figure 8 presents safety findings.
Adverse events during or following adenosine stress perfusion CMR among patients with orthotopic cardiac transplantation.
| Characteristic | Overall, N = 601 | CAV 0-1, N = 421 | CAV 2-3, N = 181 | p-value2 |
|---|---|---|---|---|
| Adenosine dosage (µg/kg/min) | 0.7 | |||
| 140 | 37 (62%) | 25 (60%) | 12 (67%) | |
| 180 | 15 (25%) | 12 (29%) | 3 (17%) | |
| 210 | 8 (13%) | 5 (12%) | 3 (17%) | |
| Peak systolic BP (mm Hg) | 124 (116, 132) | 123 (115, 132) | 124 (119, 128) | >0.9 |
| Peak HR (bpm) | 92 (81, 100) | 91 (79, 100) | 93 (83, 99) | 0.7 |
| Rest HR (bpm) | 79 (71, 90) | 80 (74, 90) | 75 (69, 88) | 0.5 |
| Change in HR (bpm) | 6 (0, 14) | 5 (0, 14) | 8 (3, 14) | 0.3 |
| Change in diastolic BP (mm Hg) | -9 (-16, 0) | -12 (-18, -2) | -2 (-8, 2) | 0.089 |
| Change in systolic BP (mm Hg) | -4 (-18, 3) | -7 (-18, 2) | 0 (-12, 3) | 0.2 |
| Symptoms during stress | 0.3 | |||
| Chest pain | 29 (50%) | 23 (56%) | 6 (35%) | |
| Hot flush | 6 (10%) | 5 (12%) | 1 (5.9%) | |
| No symptoms | 10 (17%) | 5 (12%) | 5 (29%) | |
| SoB | 13 (22%) | 8 (20%) | 5 (29%) | |
| Presence of inducible myocardial ischemia | 11 (18%) | 6 (14%) | 5 (28%) | 0.3 |
| Rest rate-pressure product | 10,664 (9,261, 12,029) | 10,772 (9,468, 12,085) | 9,794 (8,636, 11,662) | 0.2 |
| Stress rate-pressure product | 11,172 (10,092, 12,597) | 10,989 (9,779, 12,602) | 11,408 (10,458, 12,558) | 0.7 |
| Stress Myocardial Blood Flow | 2.3 (1.81, 2.8) | 2.3 (1.94, 3.1) | 2.1 (1.45, 2.5) | 0.069 |
| Stress Myocardial Perfusion Reserve Index | 2.3 (1.74, 3.1) | 2.4 (2.08, 3.3) | 1.8 (1.22, 2.1) | 0.001 |
| Rest Myocardial Blood Flow Index | 0.9 (0.81, 1.2) | 0.9 (0.78, 1.1) | 1.2 (0.84, 1.6) | 0.082 |
| Pulmonary transit time, centroid | 6.7 (5.5, 8.7) | 6.6 (5.5, 8.7) | 7.3 (5.6, 8.6) | 0.6 |
| Pulmonary transit time, peak to peak | 6.80 (5.70, 8.10) | 6.45 (5.68, 8.03) | 7.50 (6.70, 8.80) | 0.13 |
| 1 n (%); Median (IQR) | ||||
| 2 Fisher’s exact test; Wilcoxon rank sum test; Wilcoxon rank sum exact test | ||||
Quantitative perfusion CMR
Results from the in-line myocardial perfusion mapping algorithm were deemed diagnostic in 52 patients (86.7%). In the remaining 8 patients, sufficient quality in quantitative myocardial perfusion mapping was not achieved due to ECG mistriggering or failed breath-holding. Stress MBF index was normally distributed across the sample and its mean ± SD was 2.2 ± 0.7, while MPR index was not normally distributed and its median (IQR) was 2.0 (1.5 to 2.8). On logistic regression, stress MBF, MPR and their indices were associated with the presence of significant CAV at angiography (Table 4, Figure 9, and Supplemental Table 1).
| OR (95%CI) | P-value | |
|---|---|---|
| Ischemia on visual inspection | 2.24 (0.56, 8.75) | 0.24 |
| Rest MBF (mL/min/g) | 3.86 (0.6, 37.1) | 0.19 |
| Rest MBF index | 10.6 (1.8, 99.2) | 0.021 |
| Stress MBF (mL/min/g) | 0.35 (0.11, 0.87) | 0.043 |
| MPR | 0.21 (0.05, 0.61) | 0.012 |
| Stress MBF index | 0.31 (0.09, 0.82) | 0.033 |
| MPR index | 0.14 (0.03, 0.46) | 0.007 |
Diagnostic accuracy
Inducible ischemia on visual inspection
Inducible ischemia was not correlated to the presence of significant CAV on angiography. Visual ischemia had sensitivity of 27.8% and specificity of 85.4%, along with a positive predictive value of 45.5% and a negative predictive value of 72.9%.
MPR index
Overall AUC for significant CAV was 78.8%, with sensitivity 85.7%, specificity 70.3%, positive predictive value of 52.2% and negative predictive value of 92.9% at a threshold of 2.1 based on the Youden index.
Stress MBF
Stress MBF yielded an AUC of 66.6% with sensitivity 100%, specificity 36.8%, positive predictive value of 36.8% and negative predictive value of 100%.
Rest MBF index
AUC for the rest MBF index was 66.0% with sensitivity 91.9%, specificity 42.9%, positive predictive value of 66.7% and negative predictive value of 81.0%.
Comprehensive results are presented in Table 5. Visual assessment and a strategy based on MPR index were compared for reclassification across the range of pre-test probability of CAV (Figure 10). A quantitative approach based exclusively on the measurement of stress MPR index, where CAV would be adjudicated for MPR index ≥ 2.1, was effective at ruling-in and ruling-out CAV, while qualitative assessment of ischemia offered almost non-existent ruling-out capacity.
| AUC (95%CI) | Sensitivity | Specificity | PPV | NPV | |
|---|---|---|---|---|---|
| Ischemia on visual inspection | 56.6% (44.6-68.5) | 27.80% | 85.40% | 45.50% | 72.90% |
| Rest MBF (mL/min/g) | 60.9% (43.1-78.7%) | 64.30% | 62.20% | 39.10% | 82.10% |
| Rest MBF index | 66.0% (46.8-85.3) | 42.90% | 91.90% | 66.70% | 81% |
| Stress MBF (mL/min/g) | 66.6% (49.6-83.7) | 100% | 36.80% | 36.80% | 100% |
| MPR | 75.1% (61-89.2) | 92.90% | 48.60% | 40.60% | 94.70% |
| MPR index | 78.8% (64.7-92.8) | 85.70% | 70.30% | 52.20% | 92.90% |
Quantitative stress perfusion CMR and donor-specific antibodies
We separately investigated the diagnostic accuracy of quantitative stress perfusion CMR in patients with a positive or negative DSA test result. These findings along with our interpretation of them are available in the Appendix.
Clinical utility
The net benefit purported by stress CMR as a gatekeeper to angiography was evaluated over a span of reasonable threshold probabilities of CAV (10-40%). In such analysis, a strategy of quantitative stress perfusion CMR guided by MPR index (angiography for MPR index < 2.1, no angiography otherwise) was compared with a strategy where guidance was provided by presence of inducible myocardial perfusion defects («visual» ischemia on stress CMR). Quantitative stress perfusion CMR conferred incremental benefit over visual ischemia across the prespecified spectrum of threshold probabilities. Use of MPR index to gatekeep patients to angiography would lead to as much as 42% net reduction in unnecessary angiographies performed, independently of patient-physician preferences (Figure 11 and Table 6).
Clinical utility and decision curve analysis
Analysis of decision curves is an increasingly supported method to evaluate the utility of newly-introduced tests over a reference standard (1). This fits in the larger debate around the importance and real-world usefulness of metrics such as test calibration, sensitivity, specificity, or area under the curve, which are considered «statistical abstractions» (1) not inherently informative of a test’s clinical value, and poorly understood by clinicians (2). The elemental building block of a decision curve is the net benefit, which captures the trade-off between a test’s true positive rate and false positive rate:
\[ \text { Net benefit }=\frac{\text { True positives }}{N}-\frac{\text { False positives }}{N} \]
As clinicians and/or patients may be comfortable with different levels of risk before deciding to proceed with a test (i.e., pre-test risk thresholds for testing are subjective), there is not such a thing as “one-size-fits-all” risk threshold. To address this, a decision curve is created by assessing the overall advantage or benefit of a test at a range of thresholds that are considered reasonable in a clinical setting. Each point on the decision curve reflects the benefit of the test for a specific probability threshold, taking into account the diverse preferences and perspectives of clinicians and patients. Decision curves are easily interpreted in that the test that confers the greater benefit over the spectrum of threshold probabilities offers greater decisional performance.
A test’s clinical utility can also be expressed as the net reduction in the number of unnecessary downstream tests. In the present study, we measure the net reduction in angiographies deemed unnecessary (i.e. those angiographies conducted in patients in which no CAV was found, ~ 39%) obtained by referring the patient to stress CMR.
Andrew JV, Ben Van C, Ewout WS. Net benefit approaches to the evaluation of prediction models, molecular markers, and diagnostic tests. BMJ 2016;352:i6.
Whiting PF, Davenport C, Jameson C et al. How well do health professionals interpret diagnostic information? A systematic review. BMJ Open 2015;5:e008155.
Stress CMR is compared with the standard of angiography, and against a strategy of no testing. A strategy of QP-CMR, guided exclusively by a finding of MPR index < 2.1, conferred incremental benefit over visual assessment of inducible ischemia (Table 6). QP-CMR, quantitative stress perfusion CMR.
Tabulated results at 5% increments for probability thresholds (lighter is better). A) at a pre-test risk threshold of 10%, neither QP-CMR or visual ischemia conferred additional benefit over a referral to angiography. For risk thresholds 15-40%, a strategy of QP-CMR guidance to further angiographic testing would confer incremental benefit over visual assessment of inducible ischemia. B) for patients at intermediate (> 15%) risk of CAV, a QP-CMR would reduce unnecessary angiographies by 20-42%.
Correction for overfit
Ten-fold cross-validation with 25 repeats was used to evaluate model performance after accounting for overfit. The cross-validated quantitative stress perfusion CMR model yielded similar added benefit to the original model for threshold probabilities from 15% to 40%, while benefit for patients willing to undergo angiography for risk < 15% was not maintained (Supplemental Figure 1).
Prognostic yield
During follow-up, 9 patients underwent unplanned cardiac hospitalization and 2 died, for a total of 11 CV events recorded (18.3%) over a median (IQR) follow-up of 1.8 years (0.9 to 2.7) years.
Baseline characteristics
Traditional CV risk predictors such as age and presence of diabetes and dyslipidemia at baseline were not prognostic. Instead, presence of hypertension and a higher resting heart rate were related to higher risk of future MACE.
Stress CMR parameters
Among stress CMR parameters, stress MBFi and MPRi, but not qualitative assessment of ischemia, were prognostic (Table 5). Multivariable adjustment was not attempted due to the low sample size. Exploratory plotting of hazard ratios on the spectrum of stress MPR index values and stratification by MPR index for Kaplan Meier curves demonstrated that patients with an MPR index lower than 2.0 were at significantly increased risk of future MACE (Figure 12 and Supplemental Figure 2).
At a median follow-up of 1.8 years, eleven patients incurred a MACE. MPR index was found to be a significant predictor of future MACE. A MPR index ≤ 2 was associated with increased CV risk (log-rank p = 0.022).
Other CMR parameters
The prognostic value of non-stress perfusion CMR parameters was also investigated. Among these, peak global longitudinal strain was the only significant predictor of future CV events (Supplemental Table 2); other functional parameters, such as left and right ventricular ejection fraction and markers of longitudinal function, were not prognostic in this patient population, consisting mainly of outpatients with well-functioning cardiac grafts. Presence and extent of LGE were not prognostic.
Discussion
A 2010 consensus statement issued by the ISHLT defined ICA as a «screening tool to grossly detect the presence of CAV»10. This statement may be seen as controversial given earlier reports of limited utility of ICA in the setting of cardiac transplantation15, the radiological risk associated with repeated, typically yearly, testing, and the significant periprocedural complication rate, which is to be balanced against limited, where not not null, interventional benefit16. IVUS has been shown to yield increased sensitivity and might be used to select patients in need of closer mid-term surveillance23, but it does not address the issue of risk and requires additional expertise. It is also subject to variable local availability.
To date, dobutamine stress echocardiography had low diagnostic yield for CAV31. SPECT may not be considered as a diagnostic gatekeeper to ICA due to its unacceptable accuracy36, and while novel advances in MBF and MPR quantification were recently validated38, these are far from widespread clinical implementation. Potential noninvasive gatekeeper tests include CTCA, PET and stress CMR39. CTCA currently holds a central role in CAV as it allows to image the coronary arteries with great detail84. However, radiological risk is nonnegligible with CTCA as with ICA85, being especially high in young patients86 and cumulated with repeated scans, which would be recommended at a frequency of every two years in heart transplant recipients44. Despite the growing interest in tissue characterisation and CT perfusion, these remain purely investigational and evidence is lacking that CTCA might offer a “one-stop shop” for CAV surveillance and rejection. Bearing in mind that patients with a previous cardiac transplantation face significant risks for different complications, including graft rejection and CAV, it seems highly unlikely that CTCA might allow to reduce rates of utilization of endomyocardial biopsy and invasive coronary angiography; the same goes for PET, which might have high yield in CAV, but has no established value in the context of graft rejection, confers radiation exposure, and has limited worldwide availability and expertise87.
We now have high-level evidence from a randomized controlled trial that a CMR strategy is non-inferior to an invasive strategy of endomyocardial biopsy in the setting of graft rejection surveillance62. This raises the bar in terms of patient needs following cardiac transplantation; as we are able to prevent more and more patients from undergoing unnecessary biopsies, it appears just as reasonable to pursue diagnostic strategies that reduce unnecessary angiographies for CAV detection and surveillance. Noninvasive strategies allowing for quantification of myocardial blood flow for CAV assessment include PET32,33 and CMR69,73,88,89, with preliminary findings for SPECT90 and CT49.
In our study, we observed adenosine stress perfusion CMR to be safe and feasible in a substantial cohort of heart transplant recipients, with 25% stressed at < 3 years from transplantation, and able to elicit adequate cardiovascular and symptomatic responses.
We find assessment of quantitative stress perfusion cardiac MRI as a viable, noninvasive, radiation-free alternative to angiography, which provides acceptable rule-in and excellent rule-out ability, has incremental prognostic value, and high decisional performance in a real-world outpatient population of heart transplant recipients. In particular, we found in the MPR index, as measured in-line during the scan, a single parameter which would confer operational accuracy in CAV detection, with limited added value from other parameters such as stress MBF and rest MBF index. Importantly, we confirm results from a previous, non-comparative study in which qualitative assessment of inducible myocardial ischemia had poor diagnostic value in CAV64. We build on that finding by showing not only that quantitative stress perfusion CMR has incremental value over qualitative assessment of ischemia, but also that it might lead to a major reduction in the use of ICA and CTCA, with a net benefit in terms of newly-diagnosed patients which largely outweighs any missed diagnoses and fits well within diverse decision-making frameworks. These results on the clinical utility of quantitative stress perfusion CMR do not consider the additional benefit conferred by reduction in radiation exposure, for both CTCA and ICA, and relative harm and costs of ICA. The single parameter with the greatest accuracy (~79%) was MPR index, while all other parameters at rest and stress had smaller AUCs at ~60-70%.
We also evaluated the prognostic value of stress CMR long-term after heart transplantation. Our results on adenosine stress perfusion CMR expand on previous work on risk prediction by qualitative regadenoson stress CMR88,91 and semi-quantitative MPR index measurement with adenosine92. In their study, Kazmirczak, Shenoy and colleagues88 showed that an abnormal stress CMR was prognostic for a composite endpoint of composite endpoint of myocardial infarction, percutaneous revascularization, cardiac hospitalization, retransplantation or death. However, they reported on prognostic yield of broad CMR assessment, comprehensive of LV function, scarring and inducible myocardial ischemia upon visual inspection. We instead report, for the first time, on the specific prognostic value of fully quantitative perfusion parameters in patients with suspected CAV. These results can be weighed against the corresponding PET literature33,34. With the present work, CMR and PET may be nominally equiparable in terms of their diagnostic and prognostic yield, but the current lack of comparative studies justifies the need for future prospective head-to-head evaluation.
We additionally report, for the first time to our knowledge, on a correlation between the rate-pressure product and MBF during stress, which is not observed in the general population or in the first years after cardiac transplantation76 (see Appendix). In this context, we demonstrate that such relationship is independent of the presence of CAV and might indicate the need for adjustment of the stress MBF for the stress RPP for this parameter to achieve greater diagnostic and prognostic yield. We further evaluated the influence of positivity to donor-specific antibodies on diagnostic accuracy metrics, specifically for stress rate-pressure product-adjusted stress MBF, which is found to be significantly reduced among patients with no significant CAV who are DSA-positive, hinting at possible microvascular dysfunction in relation with subclinical rejection (see Appendix).
In this study, quantitative perfusion was superior to qualitative stress perfusion CMR for CAV detection and patient prognostication. At present, stress CMR is infrequently utilized for CAV assessment87. Results of this study demonstrate the value of quantitative stress perfusion CMR for CAV surveillance in heart transplant recipients and warrant future validation in a prospective trial.
Limitations
This study has the following limitations. The observational, single-centre, retrospective study design allowed to demonstrate the safety, feasibility and clinical potential, but not the real-world effectiveness of a strategy of stress CMR as gatekeeper to angiography, which should be the objective of future research effort; time between CMR and angiography varied significantly among patients, possibly leading to imperfect comparison between the respective findings; in this study, both ICA and CTCA findings were utilized in a composite angiographic endpoint (and accepted as reported), whereas the latest ISHLT recommendations, despite being more than 12 years old, state that CTCA cannot be seen as an alternative to ICA for its lack of accuracy in assessing branch vessels; additionally, we found evidence of selection bias in referral to ICA versus CTCA, presumably owing to the real-world physician preference of CTCA in low-risk patients; we cannot exclude a biased representation of cardiac transplantation outpatients in this cohort, as patients not undergoing CMR, the stress component of CMR, or angiography for any reason were systematically excluded; additionally, we have not addressed the issue of donor-transmitted coronary artery disease versus new-onset CAV of the graft heart; we have not fully evaluated the additional value of quantitative stress perfusion CMR parameters compared with non-stress CMR parameters; we did not assess the applicability of parametric mapping to CAV assessment, for which there is some initial evidence74; although we confirmed the prognostic value of global longitudinal strain in this cohort as previously demonstrated93, we did not formally compare the diagnostic and prognostic yield of strain and perfusion modules, specifically because this was a study of patients undergoing stress CMR with a limited number of events. The study sample was adequate for detection of differences in quantitative perfusion parameters between patients with and without significant CAV, but may be considered modest for the prognostic analysis. This is reflective of practice of the outpatient surveillance service for cardiac transplantation recipients which does not routinely or exclusively refer patients to CMR.
Future perspectives
Our study results pave the way for a future prospective, multi-centre, randomized controlled trial to test the hypothesis that CMR-guided catheterisation is non-inferior to current angiography referral practice in terms of safety and efficacy, and leads to reduced downstream costs owing to a reduction in use of unnecessary ICA, in long-term heart transplant recipients. As different competing tests have been proposed as credible, non-invasive gatekeepers to angiography (mainly CTCA, PET, and CMR), prospective comparisons of these tests has been recommended against a diagnostic and prognostic endpoint39. Such a study design may be challenged by the slowly-progressing nature of CAV, and it is outside of the scope of the present work to indulge in power calculations or endpoint definitions. However, where a randomized controlled trial would be deemed unfeasible, quasi-randomized designs of interrupted time-series analysis and difference-in-difference studies might instead prove feasible, less costly, and possibly as informative94. Finally, methodological and implementational aspects of this proof-of-concept study will need future investigation. There is growing awareness of the need to make cardiovascular imaging increasingly sustainable from ethical, economic and environmental standpoints95. In this light, our study findings appear incomplete and should be weighed against the comparative yield and cost-effectiveness of stress CMR and other non-invasive imaging modalities in the cardiac transplant outpatient setting, which should be the object of future investigation.
Conclusion
In this retrospective observational study, stress CMR with quantitative perfusion showed excellent safety profile and feasibility in HT recipients. Quantitative stress perfusion CMR, but not qualitative stress CMR, was diagnostic for cardiac allograft vasculopathy and yielded superior decisional and prognostic performance, findings which substantiate its utility as a gatekeeper to invasive testing. These findings warrant prospective and comparative validation.
References
1.
2.
Avery, R. K. Cardiac-allograft vasculopathy. New England Journal of Medicine 349, 829–830 (2003).
3.
Ellenbogen, K. A., Thames, M. D., DiMarco, J. P., Sheehan, H. & Lerman, B. B. Electrophysiological effects of adenosine in the transplanted human heart. Evidence of supersensitivity. Circulation 81, 821–8 (1990).
4.
Sai Kiran, B., Stephen, P. & Jayan, P. Cardiac transplantation: Indications, eligibility and current outcomes. Heart 105, 252 (2019).
5.
Giblin, G. et al. Survival trends post cardiac transplantation: A comparative analysis of irish and international data (1985-2019). The Journal of Heart and Lung Transplantation 39, S290 (2020).
6.
Yamani, M. H. & Starling, R. C. Long-term medical complications of heart transplantation: Information for the primary care physician. Cleve Clin J Med 67, 673–80 (2000).
7.
Lund, L. H. et al. The registry of the international society for heart and lung transplantation: Thirty-fourth adult heart transplantation report—2017; focus theme: Allograft ischemic time. The Journal of Heart and Lung Transplantation 36, 1037–1046 (2017).
8.
Stewart, S. et al. Revision of the 1990 working formulation for the standardization of nomenclature in the diagnosis of heart rejection. The Journal of Heart and Lung Transplantation 24, 1710–1720 (2005).
9.
Avery, R. K. Infectious disease and transplantation: Messages for the generalist. Cleve Clin J Med 65, 305–14 (1998).
10.
Mehra, M. R. et al. International society for heart and lung transplantation working formulation of a standardized nomenclature for cardiac allograft vasculopathy-2010. J Heart Lung Transplant 29, 717–27 (2010).
11.
Gao, S.-Z., Alderman, E. L., Schroeder, J. S., Silverman, J. F. & Hunt, S. A. Accelerated coronary vascular disease in the heart transplant patient: Coronary arteriographic findings. Journal of the American College of Cardiology 12, 334–340 (1988).
12.
Taylor, D. O. et al. Registry of the international society for heart and lung transplantation: Twenty-second official adult heart transplant report–2005. J Heart Lung Transplant 24, 945–55 (2005).
13.
Hiemann, N. E. et al. Prognostic impact of microvasculopathy on survival after heart transplantation: Evidence from 9713 endomyocardial biopsies. Circulation 116, 1274–82 (2007).
14.
Costanzo, M. R. et al. Heart transplant coronary artery disease detected by coronary angiography: A multiinstitutional study of preoperative donor and recipient risk factors. Cardiac transplant research database. J Heart Lung Transplant 17, 744–53 (1998).
15.
Clague, J. R., Cox, I. D., Murday, A. J., Charokopos, N. & Madden, B. P. Low clinical utility of routine angiographic surveillance in the detection and management of cardiac allograft vasculopathy in transplant recipients. Clin Cardiol 24, 459–62 (2001).
16.
Schmauss, D. & Weis, M. Cardiac allograft vasculopathy. Circulation 117, 2131–2141 (2008).
17.
Kobashigawa, J. A. et al. Effect of pravastatin on outcomes after cardiac transplantation. New England Journal of Medicine 333, 621–627 (1995).
18.
Keogh, A. et al. Sirolimus in de novo heart transplant recipients reduces acute rejection and prevents coronary artery disease at 2 years. Circulation 110, 2694–2700 (2004).
19.
Grant, S. C., Brooks, N. H. & Levy, R. D. Routine coronary angiography after heart transplantation. Heart 78, 101 (1997).
20.
Störk, S. et al. Assessment of cardiac allograft vasculopathy late after heart transplantation: When is coronary angiography necessary? J Heart Lung Transplant 25, 1103–8 (2006).
21.
Torres, H. J. et al. Prevalence of cardiac allograft vasculopathy assessed with coronary angiography versus coronary vascular ultrasound and virtual histology. Transplant Proc 43, 2318–21 (2011).
22.
Kobashigawa, J. A. et al. Multicenter intravascular ultrasound validation study among heart transplant recipients: Outcomes after five years. Journal of the American College of Cardiology 45, 1532–1537 (2005).
23.
Nelson, L. M. et al. Intravascular ultrasound-guided selection for early noninvasive cardiac allograft vasculopathy screening in heart transplant recipients. Clin Transplant 34, e14124 (2020).
24.
Bacal, F. et al. Dobutamine stress echocardiography predicts cardiac events or death in asymptomatic patients long-term after heart transplantation: 4-year prospective evaluation. J Heart Lung Transplant 23, 1238–44 (2004).
25.
Dipchand, A. I. et al. A prospective study of dobutamine stress echocardiography for the assessment of cardiac allograft vasculopathy in pediatric heart transplant recipients. Pediatr Transplant 12, 570–6 (2008).
26.
Eroglu, E. et al. Quantitative dobutamine stress echocardiography for the early detection of cardiac allograft vasculopathy in heart transplant recipients. Heart 94, e3 (2008).
27.
Mastrobuoni, S. et al. Dual-source CT coronary angiogram in heart transplant recipients in comparison with dobutamine stress echocardiography for detection of cardiac allograft vasculopathy. Transplantation 87, 587–90 (2009).
28.
Minardi, G. et al. The pressure/volume relationship during dobutamine stress echocardiography in transplanted heart: Comparison with quality of life and coronary anatomy. Cardiovasc Ultrasound 10, 44 (2012).
29.
Sade, L. E. et al. Follow-up of heart transplant recipients with serial echocardiographic coronary flow reserve and dobutamine stress echocardiography to detect cardiac allograft vasculopathy. J Am Soc Echocardiogr 27, 531–9 (2014).
30.
Chirakarnjanakorn, S., Starling, R. C., Popović, Z. B., Griffin, B. P. & Desai, M. Y. Dobutamine stress echocardiography during follow-up surveillance in heart transplant patients: Diagnostic accuracy and predictors of outcomes. J Heart Lung Transplant 34, 710–7 (2015).
31.
Elkaryoni, A. et al. Diagnostic accuracy of dobutamine stress echocardiography in the detection of cardiac allograft vasculopathy in heart transplant recipients: A systematic review and meta-analysis study. Echocardiography 36, 528–536 (2019).
32.
Chih, S. et al. PET assessment of epicardial intimal disease and microvascular dysfunction in cardiac allograft vasculopathy. J Am Coll Cardiol 71, 1444–1456 (2018).
33.
Bravo, P. E. et al. Diagnostic and prognostic value of myocardial blood flow quantification as non-invasive indicator of cardiac allograft vasculopathy. Eur Heart J 39, 316–323 (2018).
34.
Robert, J. H. M. et al. Comparative prognostic and diagnostic value of myocardial blood flow and myocardial flow reserve after cardiac transplantation. Journal of Nuclear Medicine 61, 249 (2020).
35.
Wu, Y. W. et al. Diagnostic and prognostic value of dobutamine thallium-201 single-photon emission computed tomography after heart transplantation. J Heart Lung Transplant 24, 544–50 (2005).
36.
Thompson, D., Koster, M. J., Wagner, R. H., Heroux, A. & Barron, J. T. Single photon emission computed tomography myocardial perfusion imaging to detect cardiac allograft vasculopathy. Eur Heart J Cardiovasc Imaging 13, 271–5 (2012).
37.
Aguilar, J. et al. Clinical utility of SPECT in the heart transplant population: Analysis from a single large-volume center. Transplantation 106, 623–632 (2022).
38.
Ko, K. Y. et al. Myocardial flow assessment after heart transplantation using dynamic cadmium-zinc-telluride single-photon emission computed tomography with (201)tl and (99m)tc tracers and validated by (13)n-NH(3) positron emission tomography. Circ Cardiovasc Imaging 16, e015034 (2023).
39.
Ruddy, T. D. & Wells, R. G. Measurement of myocardial blood flow with positron emission tomography and single-photon emission computed tomography for diagnosis of cardiac allograft vasculopathy. Circ Cardiovasc Imaging 16, e015593 (2023).
40.
Nunoda, S. et al. Evaluation of cardiac allograft vasculopathy by multidetector computed tomography and whole-heart magnetic resonance coronary angiography. Circ J 74, 946–53 (2010).
41.
Bastarrika, G. et al. Systolic prospectively ECG-triggered dual-source CT angiography for evaluation of the coronary arteries in heart transplant recipients. Eur Radiol 21, 1887–94 (2011).
42.
Barthélémy, O. et al. Multislice computed tomography to rule out coronary allograft vasculopathy in heart transplant patients. J Heart Lung Transplant 31, 1262–8 (2012).
43.
Beitzke, D. et al. Dual-source cardiac computed tomography angiography (CCTA) in the follow-up of cardiac transplant: Comparison of image quality and radiation dose using three different imaging protocols. Eur Radiol 25, 2310–7 (2015).
44.
Rohnean, A. et al. Heart transplant patient outcomes: 5-year mean follow-up by coronary computed tomography angiography. Transplantation 91, 583–8 (2011).
45.
Ferencik, M., Brady, T. J. & Hoffmann, U. Computed tomography imaging of cardiac allograft vasculopathy. J Cardiovasc Comput Tomogr 6, 223–31 (2012).
46.
Khan, R. & Jang, I. K. Evaluation of coronary allograft vasculopathy using multi-detector row computed tomography: A systematic review. Eur J Cardiothorac Surg 41, 415–22 (2012).
47.
Wever-Pinzon, O. et al. Coronary computed tomography angiography for the detection of cardiac allograft vasculopathy: A meta-analysis of prospective trials. J Am Coll Cardiol 63, 1992–2004 (2014).
48.
Nous, F. M. A. et al. Clinical implementation of coronary computed tomography angiography for routine detection of cardiac allograft vasculopathy in heart transplant patients. Transpl Int 34, 1886–1894 (2021).
49.
Ahn, Y. et al. CT coronary angiography and dynamic CT myocardial perfusion for detection of cardiac allograft vasculopathy. JACC Cardiovasc Imaging 16, 934–947 (2023).
50.
Almenar, L. et al. Utility of cardiac magnetic resonance imaging for the diagnosis of heart transplant rejection. Transplant Proc 35, 1962–4 (2003).
51.
Taylor, A. J. et al. Diagnostic performance of multisequential cardiac magnetic resonance imaging in acute cardiac allograft rejection. Eur J Heart Fail 12, 45–51 (2010).
52.
Usman, A. A. et al. Cardiac magnetic resonance T2 mapping in the monitoring and follow-up of acute cardiac transplant rejection: A pilot study. Circ Cardiovasc Imaging 5, 782–90 (2012).
53.
Krieghoff, C. et al. Assessment of sub-clinical acute cellular rejection after heart transplantation: Comparison of cardiac magnetic resonance imaging and endomyocardial biopsy. Eur Radiol 24, 2360–71 (2014).
54.
Miller, C. A. et al. Multiparametric cardiovascular magnetic resonance surveillance of acute cardiac allograft rejection and characterisation of transplantation-associated myocardial injury: A pilot study. J Cardiovasc Magn Reson 16, 52 (2014).
55.
Butler, C. R. et al. Correlation of cardiovascular magnetic resonance imaging findings and endomyocardial biopsy results in patients undergoing screening for heart transplant rejection. J Heart Lung Transplant 34, 643–50 (2015).
56.
Lu, W. et al. Diagnostic performance of cardiac magnetic resonance for the detection of acute cardiac allograft rejection: A systematic review and meta-analysis. J Thorac Dis 7, 252–63 (2015).
57.
Vermes, E. et al. Cardiovascular magnetic resonance in heart transplant patients: Diagnostic value of quantitative tissue markers: T2 mapping and extracellular volume fraction, for acute rejection diagnosis. J Cardiovasc Magn Reson 20, 59 (2018).
58.
Dolan, R. S. et al. Multiparametric cardiac magnetic resonance imaging can detect acute cardiac allograft rejection after heart transplantation. JACC Cardiovasc Imaging 12, 1632–1641 (2019).
59.
Miller, R. J. H. et al. Quantitative myocardial tissue characterization by cardiac magnetic resonance in heart transplant patients with suspected cardiac rejection. Clin Transplant 33, e13704 (2019).
60.
Sethi, N. et al. Quantitative cardiac magnetic resonance T2 imaging offers ability to non-invasively predict acute allograft rejection in children. Cardiol Young 30, 852–859 (2020).
61.
Han, D. et al. Diagnostic accuracy of cardiovascular magnetic resonance for cardiac transplant rejection: A meta-analysis. JACC Cardiovasc Imaging 14, 2337–2349 (2021).
62.
Anthony, C. et al. Cardiovascular magnetic resonance for rejection surveillance after cardiac transplantation. Circulation 145, 1811–1824 (2022).
63.
Korosoglou, G. et al. Strain-encoded cardiac magnetic resonance for the evaluation of chronic allograft vasculopathy in transplant recipients. Am J Transplant 9, 2587–96 (2009).
64.
Colvin-Adams, M. et al. Qualitative perfusion cardiac magnetic resonance imaging lacks sensitivity in detecting cardiac allograft vasculopathy. Cardiol Res 2, 282–287 (2011).
65.
Machida, H. et al. Magnetic resonance assessment of left ventricular diastolic dysfunction for detecting cardiac allograft vasculopathy in recipients of heart transplants. Int J Cardiovasc Imaging 28, 555–62 (2012).
66.
Hussain, T. et al. Detection and grading of coronary allograft vasculopathy in children with contrast-enhanced magnetic resonance imaging of the coronary vessel wall. Circ Cardiovasc Imaging 6, 91–8 (2013).
67.
Machida, H. et al. Usefulness of left ventricular diastolic function assessed by magnetic resonance imaging over invasive coronary flow reserve measurement for detecting cardiac allograft vasculopathy in heart transplant recipients. Int J Cardiovasc Imaging 29, 151–7 (2013).
68.
Braggion-Santos, M. F. et al. Late gadolinium enhancement assessed by cardiac magnetic resonance imaging in heart transplant recipients with different stages of cardiac allograft vasculopathy. Eur Heart J Cardiovasc Imaging 15, 1125–32 (2014).
69.
Miller, C. A. et al. Multiparametric cardiovascular magnetic resonance assessment of cardiac allograft vasculopathy. J Am Coll Cardiol 63, 799–808 (2014).
70.
Mirelis, J. G. et al. Magnetic resonance for noninvasive detection of microcirculatory disease associated with allograft vasculopathy: Intracoronary measurement validation. Rev Esp Cardiol (Engl Ed) 68, 571–8 (2015).
71.
Chih, S. et al. Perfusion cardiac magnetic resonance imaging as a rule-out test for cardiac allograft vasculopathy. Am J Transplant 16, 3007–3015 (2016).
72.
Erbel, C. et al. Myocardial perfusion reserve and strain-encoded CMR for evaluation of cardiac allograft microvasculopathy. JACC Cardiovasc Imaging 9, 255–66 (2016).
73.
DeSa, T. B. et al. Semi-quantitative myocardial perfusion MRI in heart transplant recipients at rest: Repeatability in healthy controls and assessment of cardiac allograft vasculopathy. Clin Imaging 61, 62–68 (2020).
74.
Abbasi, M. A. et al. Multiparametric cardiac magnetic resonance imaging detects altered myocardial tissue and function in heart transplantation recipients monitored for cardiac allograft vasculopathy. J Cardiovasc Imaging 30, 263–275 (2022).
75.
Muehling, O. M. et al. Reduced myocardial perfusion reserve and transmural perfusion gradient in heart transplant arteriopathy assessed by magnetic resonance imaging. J Am Coll Cardiol 42, 1054–60 (2003).
76.
Narang, A. et al. Myocardial perfusion reserve and global longitudinal strain as potential markers of coronary allograft vasculopathy in late-stage orthotopic heart transplantation. Int J Cardiovasc Imaging 34, 1607–1617 (2018).
77.
Heeswijk, R. B. van et al. Quantification of myocardial interstitial fibrosis and extracellular volume for the detection of cardiac allograft vasculopathy. Int J Cardiovasc Imaging 36, 533–542 (2020).
78.
Xue, H., Brown, L. A. E., Nielles-Vallespin, S., Plein, S. & Kellman, P. Automatic in-line quantitative myocardial perfusion mapping: Processing algorithm and implementation. Magn Reson Med 83, 712–730 (2020).
79.
Gobel, F. L., Norstrom, L. A., Nelson, R. R., Jorgensen, C. R. & Wang, Y. The rate-pressure product as an index of myocardial oxygen consumption during exercise in patients with angina pectoris. Circulation 57, 549–56 (1978).
80.
Dietz, M. et al. Comparison of the prognostic value of impaired stress myocardial blood flow, myocardial flow reserve, and myocardial flow capacity on low-dose rubidium-82 SiPM PET/CT. J Nucl Cardiol 30, 1385–1395 (2023).
81.
Shrestha, U. M. et al. Assessment of late-term progression of cardiac allograft vasculopathy in patients with orthotopic heart transplantation using quantitative cardiac 82Rb PET. The International Journal of Cardiovascular Imaging 37, 1461–1472 (2021).
82.
Fitzgerald, M., Saville, B. R. & Lewis, R. J. Decision curve analysis. Jama 313, 409–10 (2015).
83.
Andrew, J. V., Ben Van, C. & Ewout, W. S. Net benefit approaches to the evaluation of prediction models, molecular markers, and diagnostic tests. BMJ 352, i6 (2016).
84.
Ojha, V. et al. Detection of cardiac allograft vasculopathy on dual source computed tomography in heart transplant recipients: Comparison with invasive coronary angiography. Br J Radiol 95, 20211237 (2022).
85.
Einstein, A. J. et al. Radiation dose and cancer risk estimates in 16-slice computed tomography coronary angiography. J Nucl Cardiol 15, 232–40 (2008).
86.
Einstein, A. J., Henzlova, M. J. & Rajagopalan, S. Estimating risk of cancer associated with radiation exposure from 64-slice computed tomography coronary angiography. JAMA 298, 317–323 (2007).
87.
Olymbios, M., Kwiecinski, J., Berman, D. S. & Kobashigawa, J. A. Imaging in heart transplant patients. JACC Cardiovasc Imaging 11, 1514–1530 (2018).
88.
Kazmirczak, F. et al. Safety and prognostic value of regadenoson stress cardiovascular magnetic resonance imaging in heart transplant recipients. J Cardiovasc Magn Reson 21, 9 (2019).
89.
Patel, S. G., Husain, N., Rigsby, C. K. & Robinson, J. D. Safety and efficacy of regadenoson for pediatric stress perfusion cardiac MRI with quantification of myocardial blood flow. Children (Basel) 9, (2022).
90.
Ko, K.-Y. et al. Myocardial flow assessment after heart transplantation using dynamic cadmium-zinc-telluride single-photon emission computed tomography with 201Tl and 99mTc tracers and validated by 13N-NH3 positron emission tomography. Circulation: Cardiovascular Imaging 16, e015034 (2023).
91.
Jiménez-Jaso, J. M. et al. Cardiac magnetic resonance imaging myocardial perfusion reserve index in heart transplant patients. Radiologia (Engl Ed) 62, 493–501 (2020).
92.
Hofmann, N. P. et al. Comprehensive bio-imaging using myocardial perfusion reserve index during cardiac magnetic resonance imaging and high-sensitive troponin t for the prediction of outcomes in heart transplant recipients. Am J Transplant 14, 2607–16 (2014).
93.
Shenoy, C. et al. Cardiac magnetic resonance feature tracking global longitudinal strain and prognosis after heart transplantation. JACC Cardiovasc Imaging 13, 1934–1942 (2020).
94.
Kontopantelis, E., Doran, T., Springate, D. A., Buchan, I. & Reeves, D. Regression based quasi-experimental approach when randomisation is not an option: Interrupted time series analysis. Bmj 350, h2750 (2015).
95.
Picano, E. Economic, ethical, and environmental sustainability of cardiac imaging. Eur Heart J 44, 4748–4751 (2023).
96.
Czernin, J. et al. Influence of age and hemodynamics on myocardial blood flow and flow reserve. Circulation 88, 62–9 (1993).
97.
Grupper, A., Gewirtz, H. & Kushwaha, S. Reinnervation post-heart transplantation. Eur Heart J 39, 1799–1806 (2018).
Appendix
Statistical software and packages
The thesis was compiled with R and Quarto in VS Code. R packages utilized included ggplot2, ggstatsplot, pROC, dca, dplyr, tidyr, knitr, survival, and gtsummary.
Supplemental results
Relationship between MBF and the rate-pressure product
Adjustment of CMR-derived MBF and MPR to the rate-pressure product (RPP) is a matter of ongoing debate. The rate-pressure product is the double product of heart rate and systolic blood pressure, with both parameters typically measured at rest. Since its first introduction in 1993, the RPP was utilized for adjustment of the rest MBF96 to account for differences in baseline haemodynamic conditions across patients; it was in fact observed that the significant correlation between RPP and MBF would lose strength and significance when both parameters were measured during vasodilator-induced stress. For this reason, stress MBF was typically not normalized by the RPP. This was also the case in heart transplant recipients75. We tested for the existence of a correlation between RPP and MBF at both rest and stress in the present cohort. As exemplified by Supplemental Figure 3, Supplemental Figure 4, a stronger, significant correlation was identified between stress RPP and stress MBF, than it was found with both parameters at rest. One reason why the rationale of adjusting only the rest MBF might not entirely apply to heart transplant recipients is that these patients, besides presenting with different baseline haemodynamic conditions, also show diverse haemodynamic response to pharmacological vasodilation, possibly in keeping with the varying degree of graft reinnervation and thus, autonomic control over the implanted heart97. In their study of heart transplant recipients undergoing stress perfusion CMR published 20 years ago75, Muehling and colleagues found significant correlation between rest MBF and rest RPP, but not between these parameters at stress (personal correspondence, courtesy of Peter Kellman and Michael Jerosch-Herold). They also reported that, in contrast with healthy controls, the RPP in heart transplant recipients was reduced upon vasodilation with adenosine, rather than increased. This was not the case in this cohort. In fact, we show that the RPP modestly increased between rest and stress (10,764±2,413 to 11,309±2,627, p = 0.06 for difference; Supplemental Figure 5). Importantly, our cohort differed significantly from the one investigated by Muehling et al., since they enrolled patients between 2 and 14 years from transplantation, while in our cohort, we reported on patients with graft age ranging from < 1 to > 30 years, with the large majority of patients at > 10 years from graft implantation. We investigated the observed heterogeneity in RPP change between rest and stress. Age and time from transplantation were not significantly associated with change in RPP (p = 0.45 and 0.92, respectively); this plausibly excludes the hypothesis that the haemodynamic response to adenosine after transplantation is explained simply by reinnervation, which remains limited and incomplete long-term after the procedure97. Instead, male sex and positivity to DSA were significant predictors of a positive change in RPP (Supplemental Figure 6, Supplemental Figure 7). Finally, stress MBF index, but not MPR index or rest MBF index, was inversely correlated with graft age, defined as the time between graft implantation and CMR (Supplemental Figure 8). The pathophysiological foundations and possible prognostic implications of such «pseudonormalization» of the RPP response on the lifetime trajectory of cardiac grafts are unknown.
Stress MBF index
Given our finding of a significant correlation between RPP and MBF at stress, we conducted exploratory analyses on the diagnostic yield of the stress MBF adjusted to the stress rate-pressure product. We found comparable diagnostic and prognostic yield to the more «conventional» MPR index. Stress MBF index was found to significantly discriminate patients with moderate to severe CAV from those without significant coronary disease (p = 0.014). At a diagnostic threshold of 2.2, stress MBF index had a sensitivity of 92.9% and a specificity of 55.3%, and was also prognostic (HR 0.11, 95%CI 0.02-0.54, p = 0.007).
Stress MBF index and positivity to donor-specific antibodies
The low specificity observed with stress MBF index indicated a high rate of false positive tests, or patients with reduced stress MBF index who had no significant CAV on the index angiography. Patients with a false positive result were similar in baseline characteristics to patients with other test results, but they had a significantly higher prevalence of donor-specific antibodies (DSA) (p = 0.029) on blood samples tested within 36 (11-99) days from the CMR study. We separately evaluated the diagnostic accuracy in patients with a negative DSA test, and found no difference in sensitivity (90.0% vs 92.9%) and a nominal increase in specificity (72.7% vs 55.3%), with an overall improvement in AUC from 0.69 to 0.80.
Supplemental tables
Supplemental figures
License
The content of this thesis is licensed under a Creative Commons License, Attribution - Noncommercial - NoDerivative Works 4.0 International: see www.creativecommons.org. The text may be reproduced for non-commercial purposes, provided that credit is given to the original author. I hereby declare that the contents and organisation of this dissertation are the product of my own original work and do not compromise in any way the rights of third parties, including those relating to the security of personal data.