Home Modules Technology Validation Company

Module 01 / CARDIAC

HeartX

LIVE FLAGSHIP

Thirty cardiac metrics across eight regions, measured in place from one non-contrast CT.

01 · The single-patient journey

The single-patient journey.

Structures Viewer Body composition AHA-17 Similar patients 50 s · no audio

One patient, screen by screen: the 3D heart with all 28 structures shown and hidden by chamber, the CT viewer with its segmentation overlay, then the measurement panels — body composition, multi-compartment fat, the AHA 17-segment bullseye, myocardial texture, organ volumes, liver density — each value carrying its percentile within the cohort. It ends on the patients whose radiomic profile sits closest.

02 · The catalogue

Every cardiac metric, region by region.

The catalogue is the module’s spec sheet, and it says where each metric stands: eleven run in the product today, nineteen are specified and queued. This is the whole list, including the parts that are not built yet.

metrics

anatomical regions

live in product

on roadmap

HeartX metric catalogue — 30 metrics across 8 anatomical regions, each with its computation, clinical use, manual cost today, competitor coverage and status — Every metric with how it is computed, what it is used for, what it costs to do by hand today, and whether anything else on the market covers it. The catalogue’s own coverage column reads “25 have no FDA-cleared competitor” — the product’s count, shown as it appears.

03 · What it replaces

What these measurements cost by hand.

Straight from the catalogue’s “manual today?” column — what it takes to produce the same number without the module.

Measurement

By hand today

What that means

Myocardial fat volume

15–20 min

Per scan, by hand.

Myocardial fat fraction

Manual ROI

Or an MRI appointment instead.

AHA 6-region fat distribution

Not done

Not performed clinically at all today.

EAT volume

10–15 min

Semi-automatic, per scan.

04 · The analytics

Three ways in to the same values.

Unretouched captures of the running platform. HeartX carries its own accent inside the product — violet is CorpX, each module owns its hue.

Awaiting file

Fat fraction across the cohort, then whether a feature carries information beyond it — Pearson r on screen, redundant pairs listed and dropped.

Awaiting file

Associations across the cohort, patients compared side by side, then the clusters that fall out of 56 features.

CorpX steatosis analysis: fat-fraction dial at the 91st cohort percentile with a first-order feature radar and ten percentile bars — One patient against the cohort — percentiles, not verdicts.

Descriptive analytics — quantification and visualisation, not diagnosis.

05 · The endpoint package

The measurements your trial asks for.

A trial does not want thirty metrics; it wants the handful its protocol names. The report ships with endpoint presets — HFpEF, AF ablation, metabolic (SGLT2 / GLP-1), anti-inflammation, cardio-oncology, amyloidosis (ATTR), pulmonary hypertension and PE, AVC / aortic stenosis / TAVR — each measurement carrying its cohort percentile and the paper it comes from.

CorpX structured report — technique, findings, and a trial endpoint package with per-measurement literature citations and cohort percentiles — Measurements grouped into the endpoint package a given trial actually needs, each with the paper it comes from. Descriptive only — the report quantifies, it does not conclude.

06 · Validation

Proven before it ships.

Eleven metrics run in the product today, and each one traces to primary literature. Ground truth is generated synthetically — anatomy and tissue properties specified exactly — and the module must recover it within tolerance before release. What may be claimed was fixed at specification and has not drifted since.

The methodology, in full

07 · What you receive

Analysis first. The dataset underneath.

Every delivery is four things: access to the dashboards above, the browser viewer, the dataset as CSV or JSON, and the record that says how each number was produced.

The values

CSV and JSON, per region. Low-confidence flagged in the file.

The provenance

Pipeline version, weights and parameters, pinned per run.

The de-identification record

Proof the study was cleared before analysis.

08 · Who uses it

Cardiac trial endpoints

Documented, method-pinned measurements delivered as the trial’s endpoint package.

CT core labs

Consistent reads across sites and scanners, with the method recorded per run.

Cohort studies

Whole-heart phenotypes from scans already acquired.

Put HeartX on your cohort.

Request access