Comparison of Evaluation Metrics for Landmark Detection in CMR Images

Koehler, Sven; Sharan, Lalith; Kuhm, Julian; Arman, Ghanaat,; Gordejeva, Jelizaveta; K., Simon, Nike; M., Grell, Niko; André, Florian; Engelhardt, Sandy

doi:10.1007/978-3-658-36932-3_43

Cited by 4 publications

(2 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Section 5.3 ) according to the 16 myocardial AHA segments ( Vec2Strain ) [ 26 ]. This model has a modified U-Net architecture [ 45 ] and was reused from earlier works [ 36 , 44 , 46 ], still achieving state-of-the-art mean Dice scores (cf. Section 2.4 ) without any specific modifications.…”

Section: Methodsmentioning

confidence: 99%

Artificial Intelligence to derive aligned strain in cine CMR to detect patients with myocardial fibrosis: an open and scrutinizable approach

Koehler,

Kuhm,

Huffaker

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Cine Cardiac Magnetic Resonance (CMR) is the gold standard for cardiac function evaluation, incorporating ejection fraction (EF) and strain as vital indicators of abnormal deformation. Rare pathologies like Duchenne muscular dystrophies (DMD) are monitored with repeated late gadolinium-enhanced (LGE) CMR for identification of myocardial fibrosis. However, it is judicious to reduce repeated gadolinium exposure and rather employ strain analysis from cine CMR. This solution is limited so far since full strain curves are not comparable between individual cardiac cycles and current practice mainly neglects diastolic deformation patterns. Our novel Deep Learning-based approach derives strain values aligned by key frames throughout the cardiac cycle. In a reproducibility scenario (57+82 patients), our results reveal five times more significant differences (22 vs. 4) between patients with scar and without, enhancing scar detection by +30%, improving detection of patients with preserved EF by +61%, with an overall sensitivity/specificity of 82/81%.

show abstract

Section: Methodsmentioning

confidence: 99%

Artificial Intelligence to derive aligned strain in cine CMR to detect patients with myocardial fibrosis: an open and scrutinizable approach

Koehler,

Kuhm,

Huffaker

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…5) per dataset and cardiac keypoint in the original temporal resolution in Table 1. Additionally, we investigate the sensitivity of different C n on the pF D and compare the LV blood-pool center of mass C lv , the mean septum landmark (center between the average anterior and inferior right ventricular insertion points (RVIP) [11]) C sept , the CMR-volume center C vol and the center of mass for a quantile-threshold mean squared error mask averaged along the temporal axis C mse . Finally, we qualitatively evaluate the general pattern of α t and | v| t on both datasets (cf.…”

Section: Methodsmentioning

confidence: 99%

Self-supervised motion descriptor for cardiac phase detection in 4D CMR based on discrete vector field estimations

Koehler¹,

Hussain²,

Hussain³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Cardiac magnetic resonance (CMR) sequences visualise the cardiac function voxel-wise over time. Simultaneously, deep learningbased deformable image registration is able to estimate discrete vector fields which warp one time step of a CMR sequence to the following in a self-supervised manner. However, despite the rich source of information included in these 3D+t vector fields, a standardised interpretation is challenging and the clinical applications remain limited so far. In this work, we show how to efficiently use a deformable vector field to describe the underlying dynamic process of a cardiac cycle in form of a derived 1D motion descriptor. Additionally, based on the expected cardiovascular physiological properties of a contracting or relaxing ventricle, we define a set of rules that enables the identification of five cardiovascular phases including the end-systole (ES) and end-diastole (ED) without usage of labels. We evaluate the plausibility of the motion descriptor on two challenging multi-disease, -center, -scanner short-axis CMR datasets. First, by reporting quantitative measures such as the periodic frame difference for the extracted phases. Second, by comparing qualitatively the general pattern when we temporally resample and align the motion descriptors of all instances across both datasets. The average periodic frame difference for the ED, ES key phases of our approach is 0.80 ± 0.85, 0.69 ± 0.79 which is slightly better than the inter-observer variability (1.07 ± 0.86, 0.91 ± 1.6) and the supervised baseline method (1.18 ± 1.91, 1.21 ± 1.78). Code and labels will be made available on our GitHub repository. https://github.com/Cardio-AI/cmr-phase-detection

show abstract

Self-supervised Motion Descriptor for Cardiac Phase Detection in 4D CMR Based on Discrete Vector Field Estimations

Koehler

Hussain

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Comparison of Evaluation Metrics for Landmark Detection in CMR Images

Cited by 4 publications

References 10 publications

Artificial Intelligence to derive aligned strain in cine CMR to detect patients with myocardial fibrosis: an open and scrutinizable approach

Artificial Intelligence to derive aligned strain in cine CMR to detect patients with myocardial fibrosis: an open and scrutinizable approach

Self-supervised motion descriptor for cardiac phase detection in 4D CMR based on discrete vector field estimations

Self-supervised Motion Descriptor for Cardiac Phase Detection in 4D CMR Based on Discrete Vector Field Estimations

Contact Info

Product

Resources

About