Cardiac Cine with Arrhythmia: Trade-offs in Segmented Acquisition
A patient with frequent premature ventricular contractions (PVCs) undergoes a cardiac MRI. The ordered sequence is a retrospective ECG-gated segmented cine (segments of 12 views per cardiac phase, 30 phases, 1.5 T). The reconstructed cine shows 'stair-step' discontinuities and blurring in phases near the PVCs, and the effective temporal resolution is worse than the prescribed 40 ms. The patient's heart rate is irregular, ranging from 55 to 90 bpm.
1. Explain why the segmented cine degrades with PVCs and how the 'effective' temporal resolution differs from the prescribed value.
Read the worked answer
Segmented cine assumes a regular R-R interval: views acquired over multiple heartbeats are binned into the same cardiac phase based on the assumption that each beat has the same duration and phase timing. With PVCs, R-R intervals vary, so views acquired in different beats are misregistered in cardiac phase, producing stair-step discontinuities and blurring. The reconstruction may also reject or accept beats inconsistently, causing phase-to-phase intensity jumps.
The prescribed temporal resolution (e.g., 40 ms) is the duration of one segment (views per segment x TR). The effective temporal resolution is worse because:
- The reconstruction may combine data across beats with different R-R intervals, effectively averaging over a longer window.
- Arrhythmia rejection may discard beats, reducing the number of views per phase and requiring more beats to fill the segment, which lengthens the acquisition and increases the chance of motion between contributing beats.
- If the scanner uses retrospective gating with a fixed number of phases, it may interpolate across irregular intervals, blurring the phase.
So the nominal 40 ms is not achieved in practice; the true temporal footprint is longer and inconsistent.
2. What acquisition changes would you make to improve robustness, and what are the costs?
Read the worked answer
Options:
- Real-time cine (non-gated, single-shot per frame): robust to arrhythmia because each frame is acquired independently. Cost: lower SNR and spatial resolution, and temporal resolution limited by readout (typically 40-80 ms per frame).
- Arrhythmia rejection with prospective gating: accept only beats within a tolerance window (e.g., +/-10% of the running R-R). Cost: longer scan, possible incomplete coverage if many beats are rejected.
- Reduce views per segment to shorten the segment duration, improving temporal resolution at the cost of SNR and longer scan time.
- Use a self-gated or motion-corrected reconstruction (e.g., MOCO, compressed sensing real-time) if available, which tolerates irregular rhythm.
- Increase averages to recover SNR lost by shorter segments, at the cost of scan time.
A practical plan: switch to real-time cine for functional assessment, keep a prospectively gated segmented cine with arrhythmia rejection for high-resolution anatomy, and document the rhythm irregularity.
3. The technologist proposes simply increasing the number of phases from 30 to 50 to 'capture the PVCs.' Evaluate this.
Read the worked answer
Increasing phases does not fix the underlying problem. The artifact comes from phase misregistration across beats, not from insufficient sampling of the cardiac cycle. More phases means each phase contains fewer views (shorter segment), which:
- Lowers SNR per phase.
- Increases the number of beats needed to fill all phases, giving more opportunity for irregular beats to corrupt the data.
- May exceed the achievable temporal resolution if the segment becomes shorter than TR x views, forcing interpolation.
It could marginally help if the goal is to resolve rapid events, but with irregular R-R intervals the misregistration persists. Better to address rhythm robustness directly (real-time, arrhythmia rejection, self-gating) rather than oversampling phases. If more phases are used, pair them with arrhythmia rejection and verify that the segment duration remains physically achievable.