Skip to main content
2025 SCMR Conference Main banner
Final Program


Schedule at a Glance
Back
On-Demand Video

Early Career

Early Career Award Translation

Diagnosing Intramyocardial Hemorrhage Using Free-Breathing, Cardiac Quantitative Susceptibility Mapping with Full LV Coverage – From Foundational Canine Studies to Initial Clinical Experience

Thursday, January 30, 2025
5:00 PM – 5:10 PM East Coast USA Time
Location: Blue Room
Background: Intramyocardial hemorrhage (IMH) following reperfused myocardial infarction (MI) can lead to focal iron deposition and contribute towards significant adverse cardiovascular outcomes. The recent statement from the Canadian Cardiovascular Society has identified IMH as the highest risk factor for acute myocardial infarction (AMI), emphasizing the need for accurate characterization of IMH in AMI management (1). While quantitative susceptibility mapping (QSM) has been proposed to measure blood oxygenation in the heart, its application in IMH assessment has been met with significant technical challenges (2). To overcome these challenges and existing limitations of T2* CMR (the current standard for imaging IMH), we developed a free-running, whole-heart high-dynamic-range QSM (HDR-QSM) technique. We validated HDR-QSM using animal models and tested it in patients reperfused for acute myocardial infarction (AMI).

Methods: All studies were performed on a 3T MRI scanner (MAGNETOM Prisma, Siemens Healthineers, Forchheim, Germany). A 3D, free-breathing, non-ECG gated multi-gradient echo (mGRE) research sequence was developed and reconstructed with a low-rank tensor (LRT) framework (TE1/ΔTE: 1.77/1.59 ms, number of echos: 8, voxel size: 1.6 × 1.6 × 6 mm³) (3). QSM images with reduced artifacts were constructed using an HDR-QSM algorithm (4). We tested and validated HDR-QSM in canine models of reperfused MI (n=14; 8 were positive for IMH), which were studied against ex-vivo R2* (1/T2*) scans and histology. Subsequently, we tested the clinical translatability of the approach in healthy subjects (n=5) and acute MI patients (n=40; 15 positive for IMH).

Results: HDR-QSM significantly improved the image quality and reduced imaging artifacts in canines, showing close alignment with ex-vivo histology (Fig. 1A) and comparable iron quantification accuracy to the ex-vivo ground truth (Fig. 1B). Results from human subjects are shown in Fig. 2, which demonstrate that HDR-QSM can significantly reduce off-resonance artifacts (indicated by the red arrows) in both patients and healthy volunteers when compared to R2* maps. This is captured in Fig. 3, where both the area under the curve (AUC) and accuracy of HDR-QSM are greater than 3D R2* Maps: AUC of 0.79 (R2*) vs 0.91 (HDR-QSM) and accuracy of 0.75 (R2*) vs. 0.87 (HDR-QSM).

Conclusion: The proposed free-running, whole-heart HDR-QSM can be used to detect and quantify IMH with greater accuracy than R2* mapping and overcomes the known limitations of previous QSM approaches. The utility of the approach remains to be investigated in larger clinical studies.

Fig. 1. Proof-of-Concept in Canines Demonstrating Improved Imaging of IMH with HDR-QSM with Ex-vivo Validation. In panel A, red arrows highlight the off-resonance at the lateral wall in the R2* map, which is successfully mitigated by the HDR-QSM. White arrows in histology image point to the IMH region that is positive for iron in HDR-QSM. Panel B demonstrates that HDR-QSM offers comparable quantification of iron concentration to ex-vivo standards.

Fig.2. Representative T2* vs HDR-QSM from a healthy volunteer (top row), non-hemorrhagic MI patient (middle row), and hemorrhagic MI patient (bottom row). Proposed HD-QSM approaches overcomes the image artifacts at the heart-lung interface observed with standard (breath-held and ECG-gated 2D) and 3D T2* images (volunteer and patient, red arrows). In addition, orange arrows indicate the motion-artifacts observable in standard T2* images; Blue arrows point to the non-hMI region. White arrows point to IMH regions in HDR-QSM image from a MI patient.

Fig.3. Diagnostic Performance of R2* and HDR-QSM in Reperfused MI Patients. ROC from AUC analysis (A) indicates that HDR-QSM significantly enhances accuracy in detecting IMH compared to 3D R2* (0.79 (3D R2* Maps) vs. 0.91 (HDR-QSM)). The increase in AUC is likely from the reduction in off-resonance imaging artifacts seen in R2* maps but markedly reduced in HDR-QSM. This is further supported by the accuracy metrics in panel B, where HDR-QSM shows significantly higher sensitivity, specificity, and overall accuracy compared to the 3D R2* approach.