What question did this study set out to answer?

To assess differences in volumetry, image quality, and acquisition time between conventional sensitivity encoding (SENSE) and deep learning-based super-resolution reconstruction using compressed sensitivity encoding (CS-SR).

February 11, 2026The International Journal of Cardiovascular ImagingOpen Access

Cardiac MR function analysis with DL-based super resolution reconstruction: application in the clinical setting

Q: What are the key findings of this study?

CS-SR cardiac MRI significantly reduced acquisition time to 165.6s compared to 411.1s for standard SENSE (p<0.001), without impairing left ventricular volumetric analysis.

Q: What is the clinical evidence from this study?

Study design: Observational. Population: Ischemic and non-ischemic cardiomyopathies (n=31). Intervention: Compressed sensing with deep learning-based super-resolution reconstruction (CS-SR) vs. Standardized sensitivity encoding (SENSE). Primary outcome: End-diastolic volume (EDV) (Mean difference 0.04 ml, 95% CI -11.19 to 11.26, p=0.970).

Q: What does this research mean for the field?

Deep learning-based super-resolution reconstruction (CS-SR) significantly shortens cardiac MRI acquisition times without impairing left ventricular volumetric analysis or overall image quality. Novelty: ClaimNovelty.INCREMENTAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

View Full Paper

Ask AI

Bookmark

Key result

CS-SR cardiac MRI cuts acquisition time ~60% vs standard SENSE without impairing LV volumetry.

Mean difference −0.04 ml
95% CI -11.19 to 11.26
P=0.970
n=31

Why the study?

Differences in volumetry, image quality, and acquisition time between standardized SENSE cine sequences and deep learning-based super-resolution reconstruction using compressed sensitivity encoding remained to be assessed.

Does deep learning-based super-resolution reconstruction with C-SENSE reduce acquisition time without impairing left ventricular volumetry and image quality compared to standard SENSE in patients undergoing cardiac MRI?

Key Points

To assess differences in volumetry, image quality, and acquisition time between conventional sensitivity encoding (SENSE) and deep learning-based super-resolution reconstruction using compressed sensitivity encoding (CS-SR).
Retrospectively evaluated 31 consecutive patients (mean age 61.2 ± 13.1 years, 26% female [8/31], 74% male [23/31]) undergoing 1.5T cardiac MRI for ischemic and non-ischemic cardiomyopathies.
Acquired balanced steady-state free precession cine sequences using parallel imaging SENSE (R = 2) and compressed sensing C-SENSE (R = 4), reconstructing C-SENSE datasets using deep learning denoising and super-resolution.
Conducted manual left ventricular segmentation by two blinded readers and subjective image quality evaluations by three independent readers using five-point Likert scales.
Left ventricular volumetry showed high agreement between SENSE and CS-SR (r = 0.98–1.00), with no significant differences for end-diastolic volume (mean difference 0.04 ml, LoA -11.19 to 11.26 ml; p = 0.970) or end-systolic volume (mean difference 1.60 ml, LoA − 7.48 to 10.68 ml; p = 0.064).
Total acquisition time was significantly shorter for C-SENSE compared to SENSE (165.6 ± 21.5 s vs. 411.1 ± 47.7 s; p < 0.001), while per-breath-hold duration was longer (14.2 ± 1.4 s vs. 7.6 ± 1.2 s; p < 0.001).
Overall image quality was comparable (p = 0.061), with CS-SR achieving reader scores ≥ 4 in sharpness more often than SENSE (52% [16/31] vs. 13% [4/31]; p < 0.001), but scoring ≥ 4 for artifact freedom less often (19% [6/31] vs. 39% [12/31]; p < 0.001).

Study Design

Type

Observational (n=31)

Blinding

Blinded readers

Multicenter

Structured PICO

Population

31 consecutive patients (mean age 61.2 ± 13.1 years, 74% male) undergoing cardiac MRI for ischemic and non-ischemic cardiomyopathies. Exclusions: cardiac rhythm abnormalities (atrial fibrillation, ventricular extrasystoles), age below 18 years, incomplete SAx cine imaging datasets, and declined or missing general consent.

Intervention

Deep learning-based super-resolution reconstruction based on high-resolution images acquired with compressed sensitivity encoding (C-SENSE, undersampling factor R=4) on a 1.5T MRI scanner.

Comparator

Standardized sensitivity encoding (SENSE, undersampling factor R=2) approach.

Outcome

Differences in left ventricular volumetry (end-diastolic volume and end-systolic volume), image quality, and acquisition time.surrogate

Main Result

Effect estimate: Mean difference 0.04 ml (95% CI -11.19 to 11.26)

Absolute Event Rate: 167.5% vs 167.5%

p-value: p=0.970

Deep learning-based super-resolution reconstruction of cardiac MRI significantly reduces acquisition time while maintaining accurate left ventricular volumetric analysis and comparable overall image quality.

Limitations

Small patient cohort
Single-center nature of the trial
Use of a vendor-specific DL-based reconstruction algorithm
Potential effect on regional wall motion abnormalities and detectability of small anatomic structures was not considered
Potential effect of the reconstruction algorithm on regional wall motion abnormalities and the detectability of small anatomic structures was not considered

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

May shorten cardiac MRI acquisition without impairing LV volumetry; leaves open prospective validation of artifact effects.

Cite This Study

Adomat et al. (2026) conducted an observational in Ischemic and non-ischemic cardiomyopathies (n=31). Compressed sensing with deep learning-based super-resolution reconstruction (CS-SR) vs. Standardized sensitivity encoding (SENSE) was evaluated on End-diastolic volume (EDV) (Mean difference 0.04 ml, 95% CI -11.19 to 11.26, p=0.970). CS-SR cardiac MRI significantly reduced acquisition time to 165.6s compared to 411.1s for standard SENSE (p<0.001), without impairing left ventricular volumetric analysis.

synapsesocial.com/papers/698c1c33267fb587c655e748 https://doi.org/10.1007/s10554-026-03642-8

View Full Paper

Ask AI

Bookmark

Cardiac MR function analysis with DL-based super resolution reconstruction: application in the clinical setting

Key Points

Study Design

Structured PICO

Main Result

Limitations

Discussion

Implication

Cite This Study

Cardiac MR function analysis with DL-based super resolution reconstruction: application in the clinical setting

Key Points

Study Design

Structured PICO

Main Result

Limitations

Discussion

Implication

Cite This Study

Also Consider

Also Consider