What is CardiacNexus
CardiacNexus is a UK Biobank cardiovascular magnetic resonance phenotype extraction pipeline and documentation site. It turns multimodal CMR acquisitions into structured cardiac phenotypes, time-resolved curves, QC artifacts, and source-audited interpretation notes.
- Modality
- Multimodal CMR
- UKB source
- UK Biobank cardiac MRI fields 20207-20214
- Pipeline step
- Preparation, segmentation, feature extraction, aggregation, and documentation validation
- Outputs
- CSV phenotypes, NPZ time series, QC images, segmentation-derived visualizations, phenotype dictionary entries
- Maturity
- Source-audited overview page
What the project produces
CardiacNexus documents quantitative CMR phenotypes across ventricles, atria, myocardium, aorta, valves, phase-contrast flow, native T1 tissue characterization, and cross-chamber coupling. The current documentation separates three layers:
| Layer | Purpose | Examples |
|---|---|---|
| Modality pages | Explain source acquisitions and implementation boundaries | short-axis cine, long-axis cine, phase-contrast flow, native T1 |
| Phenotype pages | Define exact measurements, units, outputs, reference context, disease interpretation, and QC caveats | ventricular structure, atrial function, aortic distensibility, cross-chamber phenotypes |
| Reference/method pages | Compact data contracts and shared methodology | CSV schema, NPZ keys, quality control, feature formulas |
Why phenotype-first documentation
Embeddings and learned representations can be useful, but they are hard to audit without a traceable anatomical or physiological meaning. CardiacNexus keeps a phenotype-first layer: every promoted phenotype page should name the source modality, implementation source, exact output column, unit, missingness behavior, and interpretation boundary.
Phenotype
A CardiacNexus phenotype is a quantitative descriptor of cardiac structure, function, tissue, vascular geometry, blood flow, or motion derived from CMR images and written into a documented output contract.
Current documentation standard
Promoted phenotype pages are expected to be source-audited against the current implementation and supporting registries. For draft rollout, a page may defer public figure reuse permission if the figure provenance is complete. Public release readiness is stricter and requires reuse wording to be resolved.
The site avoids presenting disease badges as diagnostic classifiers. Disease context is used for navigation and interpretation, while the output rows remain quantitative research phenotypes.
Reader path
Start with UK Biobank CMR protocol to map source fields to modalities. Use Pipeline overview for the extraction flow. Then read modality and phenotype pages for feature-specific details. Use Outputs and data model when working with CSV, NPZ, and QC artifacts.
Source audit
- Scope and route organization were checked against the current website source under
website/src/app/docs/**. - Output-contract expectations were checked against
docs/data/output_column_inventory.ymlanddocs/data/phenotype_dictionary.yml. - Validation expectations were checked against the current docs validators in
website/scripts/**. - Textbook context boundary: broad clinical textbook context is not surfaced here because this overview page is project/provenance context rather than disease interpretation.
References
- Bai W, Sinclair M, Tarroni G, Oktay O, Rajchl M, Vaillant G, Lee AM, Aung N, Lukaschuk E, Sanghvi MM, Zemrak F, Fung K, Paiva JM, Carapella V, Kim YJ, Suzuki H, Kainz B, Matthews PM, Petersen SE, Piechnik SK, Neubauer S, Glocker B, Rueckert D. Automated Cardiovascular Magnetic Resonance Image Analysis with Fully Convolutional Networks. Journal of Cardiovascular Magnetic Resonance. 2018;20(1):65.
- Bai W, Suzuki H, Qin C, Tarroni G, Oktay O, Matthews PM, Rueckert D. Recurrent Neural Networks for Aortic Image Sequence Segmentation with Sparse Annotations. In: Medical Image Computing and Computer Assisted Intervention - MICCAI 2018. Lecture Notes in Computer Science, vol. 11073. Springer International Publishing; 2018:586-594.
- Puyol-Anton E, Ruijsink B, Baumgartner CF, Masci PG, Sinclair M, Konukoglu E, Razavi R, King AP. Automated Quantification of Myocardial Tissue Characteristics from Native T1 Mapping Using Neural Networks with Uncertainty-Based Quality-Control. Journal of Cardiovascular Magnetic Resonance. 2020;22(1):60.
- Petersen SE, Matthews PM, Francis JM, Robson MD, Zemrak F, Boubertakh R, Young AA, Hudson S, Weale P, Garratt S, Collins R, Piechnik S, Neubauer S. UK Biobank's Cardiovascular Magnetic Resonance Protocol. Journal of Cardiovascular Magnetic Resonance. 2016;18(1):8.