CardiacNexus documentation

CardiacNexus is a cardiovascular magnetic resonance imaging (CMR) phenotype extraction pipeline for UK Biobank. This documentation is organized for clinicians, imaging researchers, and paper readers who need to understand what each phenotype means, how it was derived, and how it should be interpreted.

Understand the pipeline

Follow image preparation, segmentation, feature extraction, QC, and output generation.

Trace a modality

Start from the CMR sequence, UKB field provenance, and downstream phenotype families.

Find a phenotype

Browse chamber, vascular, flow, tissue, and strain phenotypes with units and caveats.

Interpret values

Review reference ranges, disease associations, QC limitations, and citation guidance.

Read by disease

Open clinician-facing primers that connect feature families to common cardiac disease contexts.

Reading model

Use modality pages when you want to know which UK Biobank CMR images are used. Use phenotype pages when you want definitions, units, formulas, output columns, reference ranges, and disease associations. Use disease primers when you want to read CardiacNexus outputs as a clinical pattern. Use methods pages when you need to understand how CardiacNexus turns images into tabular and time-series outputs.

Clinical reading principle

CardiacNexus phenotypes are quantitative imaging descriptors. Disease associations in this site are literature and context links, not standalone diagnostic rules.

Evidence labels

Reference ranges and feature summaries use evidence labels to separate values that are source-verified from values that remain method-dependent, literature-context only, not current pipeline outputs, or still in need of source-level verification.

Current maturity

The site is being converted from manuscript-style notes and pipeline development pages into a web-native reference. Pages marked as draft provide the intended structure and known provenance while detailed reference ranges and validation notes are filled in.