Study population
This prospective study was approved by the Institutional Ethical Committee of Beijing Chaoyang Hospital (2020-ke-225) and was performed in agreement with the Declaration of Helsinki. From January to February 2021, we recruited 14 consecutive first-time STEMI patients (11 men, mean age: 62 ± 11 years) successfully treated by primary percutaneous coronary intervention (PPCI) < 12 hours from the onset of symptoms at Beijing Chaoyang Hospital. Diagnosis and treatment of STEMI were as per current guidelines [15, 16]. Exclusion criteria were < 18 years of age, previous cardiac events (myocardial infarction, cardiac surgery, PCI), valvular diseases, cardiomyopathies, abnormal renal function, or contraindications to the use of gadolinium-based contrast agents or CMR. All patients underwent baseline myocardial 18F-AIF-NOTA-FAPI PET/CT imaging (6 ± 2 days post MI) and CMR (8 ± 2 days post MI). Ten patients had convalescent CMR scans (84 ± 4 days post MI). Patients or their guardians signed the informed consent before enrollment in the study.
Blood analysis
Serial blood samples were tested in all patients at admission and in the cardiac care unit before transferring the patients to ordinary wards, including complete blood count, metabolic panel (troponin I, creatine kinase-MB [CK-MB], creatine kinase [CK], B-type natriuretic peptide [BNP], lactate dehydrogenase [LDH], cholesterol and triglyceride levels, liver and renal function tests), and arterial blood gas tests.
FAPI images acquisition and interpretation
18F-AIF-NOTA-FAPI was radio-labelled as previously described [17]. PET/CT images were acquired 60 minutes after injection (2.5-3.0 MBq/kg) using a 16-slice PET/CT scanner (Discovery STE, GE, USA). CT parameters were 140 kV, 120 mA, pitch 1.375, 16 ×0.625 mm collimation, and section width of 5 mm. Two beds of PET images (5 min/bed, 3D mode) were acquired, and the heart was set in the center of the view. Attenuation-corrected PET images (voxel size, 3.9 × 3.9 × 3.3 mm) were reconstructed from the CT data using a 3D ordered-subset expectation maximization algorithm (14 subsets, 2 iterations). Integrated PET and CT images were obtained automatically on AW VolumeShare 2 (GE Healthcare).
FAPI images were independently evaluated by two nuclear physicians (XYX and MFY) blinded to the CMR and clinical information. Disagreements were resolved by consensus. Myocardial FAPI uptake was analyzed in the global and segmental manner, respectively. For global analysis, a volume of interest (VOI) was manually drawn around the myocardium with FAPI uptake, and a region grow algorithm with a threshold of 50% of the maximum uptake was set to determine the FAPI volume. The FAPI volume was further expressed as a percentage of left ventricular (LV) volume derived from MRI–this percentage was defined as the FAPI extent (FAPI%). Within this VOI, the maximum standardized uptake values (SUVmax) were automatically derived, and a target-to-background ratio (TBRmax) was calculated by dividing myocardial SUVmax by blood pool SUVmean to indicate the FAPI intensity. Finally, the total FAPI was calculated by the product of FAPI% and FAPI TBRmax.
For segmental analysis, the American Heart Association 17-segment model was applied using QPS software (version 3.1, Cedars-Sinai Medical Center, Los Angeles, CA, USA). Segmental FAPI uptake was automatically calculated by the software and expressed as a percentage of the LV SUVmax, and the product of this percentage and LV TBRmax was further calculated to represent the segmental TBRmax.
CMR image acquisition and interpretation
CMR was performed on a 3.0-T scanner (Prisma, Erlangen, Siemens Healthcare, Germany) using the following protocol: balanced steady-state free precession (bSSFP) acquisitions for long- and short-axis cine; short time inversion recovery (STIR) prepared T2WI of three short axis slices (basal, mid-ventricular and apical); standard LGE imaging matching long- and short-axis bSSFP cine slices for the quantification of scar burden; and modified Look-Locker inversion recovery (MOLLI) sequence for pre- and post-contrast T1 mapping of 3 short-axis slices. LGE and post-contrast T1 mapping images were collected 10-15 min after the administration of 0.1 mmol/kg Gd-DOTA (DotaremTM, Guerbet S.A., Paris, France). Typical acquisition parameters for bSSFP were TE/TR = 1.4/44 ms, flip angle 44°, voxel size 0.8 × 0.8 × 8.0 mm, and matrix 192 × 256. Parameters for STIR-prepared T2WI were TE/TR = 50/2100 ms, voxel size 1.4 × 1.4 × 8.0 mm, and matrix 135 × 256. The MOLLI acquisition followed the 5(3)3 protocol for pre-contrast and the 4(1)3(1)2 protocol for post-contrast T1 mapping with typical parameters of TE/TR = 1.12/283 ms, voxel size 1.4 × 1.4 × 8.0 mm, and matrix 148 × 256. The LGE parameters were TE/TR = 1.96/450 ms, flip angle 20°, and voxel size 1.4 × 1.4 × 8.0 mm.
CMR images were analyzed on a commercial software (Cvi42, Version 5.12.2, Circle Cardiovascular Imaging Inc., Calgary, Canada) by two experienced CMR operators (JXW and SHZ) who were blinded to both clinical and FAPI data; disagreements were resolved by consensus. Volumetric and functional parameters were calculated from the cine images. The infarcted myocardium was identified on LGE images using 5 SD above the remote myocardium [14]. Myocardial edema was defined by a high signal intensity on T2WI two standard deviations (SD) above the signal intensity of the remote myocardium. The extent of T2WI and LGE was expressed as the percentage of LV mass (T2WI% and LGE%). Microvascular obstruction (MVO) was identified in LGE images as hypointense recesses within the hyperenhanced myocardium. ECV were quantified according to the established formula [18]: ECV = (1 – hematocrit) × (1/T1myopost – 1/T1myopre) / (1/T1bloodpost – 1/T1bloodnative) and expressed as ECV%. The apex segment was excluded in the segment-to-segment comparative analysis between FAPI and CMR images.
Statistical analysis
SPSS Statistics (Version 24; IBM) was adopted. The normality of distribution was assessed by Kolmogorov–Smirnov test. Continuous variables were described as mean (SD) or medians (ranges), depending on the normality of distribution. Categorical variables were expressed as absolute numbers and percentages. A paired t test was used to compare the different imaging parameters in the same individual. ANOVA and non-parametric tests were used to compare the differences of parameters among segments. Pearson/Spearman’s correlation analysis was conducted to explore the relationship. Boxplot was used to determine the outliers. A p-value < 0.05 was considered to be statistically significant.