Skip to main content
2025 SCMR Conference Main banner
Final Program


Schedule at a Glance
Back

Rapid Fire Abstracts

Rapid Fire 09- Interventional MRI & Miscellaneous

Evaluation of beat detection during MRI with an MR-compatible 12-lead ECG device. Comparison to Philips Expression MR400 (RF_SA_453)

Saturday, February 1, 2025
9:30 AM – 9:40 AM East Coast USA Time
Location: Blue Room PreFunction
Background: Synchronization of cardiac magnetic resonance (CMR) imaging with the heartbeat is vital for accurate, motion-free heart images. Timing the acquisition to specific cardiac phases, especially diastole, minimizes motion artifacts and improves quality. Poor synchronization impacts diagnostic accuracy and prolongs acquisition time since missed beats require rescans [1-3].


This study compares the MiRTLE system to the Philips Expression MR400 system by analyzing their beat detection (BD) performance during MRI.


Methods: All MRI scans (10 typical iCMR sequences) were conducted using a 1.5T Siemens Aera scanner on 21 volunteers (average age: 41 ± 16 years, BMI: 26 ± 5) for the dual instrumentation data collection (DID). The MR400 system was given priority for electrode placement. The heartbeats detected by both systems were synchronized using a DATAQ DI-1100 acquisition system: MiRTLE BD was connected to Channel 1, MR400 BD to Channel 2, and MR400 ECG to Channel 3.



To quantify the results, we manually identified the true beat positions, comparing them to those detected by each system. Beats were classified as True Positive (TP) if detected within a ± 25 ms window around the true beat; False Negative (FN) if no beat was detected within the window; and False Positive (FP) if a beat was detected outside the window or more than one beat was detected. Metrics like Positive Predictive Value (PPV = TP / [TP + FP]), True Positive Rate (TPR = TP / [TP + FN]), and F1-Score (F1 = 2TP / [2TP + FP + FN]) were calculated to assess performance.


“Advanced Filtering” was used to filter the input signals on MR400 [4]; for MiRTLE we used 4 different filtering options [5], each activated during 3 or 4 unique pulse sequences. The choice of the filters was based on the recommendation list (from MiRTLE).


Results: Table 1 outlines the subjects' details, while Figure 1 showcases ECG waveform examples from both systems. In total, 30,607 heartbeats were analyzed. The mean PPV was 0.94 ± 0.04 for MiRTLE and 0.48 ± 0.22 for MR400. The TPR was 0.92 ± 0.05 for MiRTLE and 0.66 ± 0.19 for MR400, while the F1-scores were 0.93 ± 0.04 and 0.55 ± 0.20, respectively.


Conclusion: This study shows with optimized filtering MiRTLE’s BD performance was significantly better compared to MR400 during the studied interventional cardiac imaging sequences. Since MiRTLE's filters were recommended for each pulse sequence, there was a possibility of a bias in the results. To balance this, the "Advanced Filtering" setting on MR400 was used as a recommended option by Philips during scanning.