Platform documentation

heArtx

heArtx is a cloud platform for cardiac simulation. You bring a model of a heart, and the platform takes it from raw data to a finished, reproducible result. It prepares the model, fits it to your measurements, runs the simulation on cloud hardware, and streams the results back to your browser in real time.

Built for researchers, engineers, and technical teams who need dependable, shareable cardiac studies without managing the machinery behind them.

What the platform is for

Simulating how a heart moves and pumps is powerful but demanding work. It usually means stitching together specialized tools, then hand-tuning a model until it matches real measurements. After that, finding hardware to run it on, and tracking every input so the result can be trusted and repeated.

heArtx brings all of that into one place. You work entirely from your browser: upload a model, guide it through a short, well-marked set of steps, and let the platform handle the heavy computation in the cloud. Every run is recorded so you — or a colleague — can come back to it, understand exactly how it was produced, and reproduce it.

Core capabilities

Bring in a model

Start from a heart model derived from imaging, from a reconstruction, or even from an ultrasound video. The platform shows it in an interactive 3-D view as soon as it loads.

Prepare the model

Establish the model's natural resting state so the simulation starts from a sound, physically meaningful baseline.

Fit to measurements

Tune the model so its behavior matches your data, then lock in those settings as a reusable profile for the model.

Search for the best settings

Let the platform explore many options automatically to find settings that meet targets you choose, seeded by what worked before.

Run the simulation

Launch a full cardiac-cycle simulation on cloud hardware with a single click. Heavy computation runs remotely — nothing ties up your machine.

Watch it live

Follow progress as it happens — pressure–volume curves, strain, status, and a live log all update in real time while the run executes.

Review results

Open a results gallery with interactive plots and a single downloadable bundle containing everything the run produced.

Save your work

Group a model and its runs into a named workspace you can revisit, reload, and share — picking up exactly where you left off.

How a study flows

Every study moves through the same clear path. Most steps are optional — you can run a quick simulation from a ready model, or take the full route from data to a fitted, reproducible result.

  1. Ingest. Upload or reconstruct a heart model and inspect it in the 3-D view before going further.
  2. Prepare & calibrate. Optionally establish the resting state, fit the model to your measurements, or let an automated search find settings that hit your targets.
  3. Simulate. Submit a full cardiac-cycle run; it executes on cloud hardware while you keep working.
  4. Monitor. Follow live curves, status, and a streaming log as the run progresses.
  5. Results. Explore interactive plots and download one bundle with everything the run produced.
  6. Reproduce. Save the work as a named workspace so it can be reloaded, shared, and repeated later.

Collaboration & handoff

Work is organized into workspaces — named containers that keep a model and all of its runs together. A workspace is the unit you hand off: download it as a self-contained archive, or reopen a saved one to continue exactly where it was left.

Each run carries a complete record of the settings that produced it, so re-running from a saved snapshot gives the same study back. That makes it a reliable starting point for a colleague, a review, or your own follow-up work.

Cloud execution & scale

Runs happen in the cloud

The demanding computation runs on remote hardware, not your laptop. The interface stays responsive while a study is in flight.

Close the browser anytime

A run keeps going even if you disconnect. Come back later, reopen it, and pick up live monitoring where you left off.

Heavy work is queued

Submissions are scheduled and processed in order, so large jobs are handled dependably rather than competing for your machine.

Capacity grows with demand

Computation scales out across cloud resources, so the platform absorbs larger and more frequent studies without changing how you work.

Data & asset management

Everything you bring in and everything a run creates is tracked for you. Inputs, models, and generated outputs are stored in the cloud and made available on demand — you download directly, without round-trips through your machine.

Each completed run publishes a single output bundle: one archive containing all of that run's results, so there is never a scattered pile of files to chase. Your assets are private to you; you only ever see your own models, runs, and downloads.

Live monitoring

You are never left guessing about a run. While it executes, the platform streams its progress back to you continuously:

  • the pressure–volume loop, drawn live as cycles complete;
  • strain measures as they are computed;
  • a clear status message and a progress indicator with an estimated finish;
  • a raw, verbatim log for when you want the full detail.

Reproducibility & governance

Named workspaces

Save the full state of a study and restore it later with one click.

Complete history

Every run is kept with its status, timing, and outcome so nothing is lost.

Settings snapshots

Each run records the exact settings that produced it — the basis for repeating it.

Downloadable archives

Export a study or its results for your own records, sharing, or publication.

Private by default

You sign in to your own space and see only your own models, runs, and data.

Typical use cases

Patient-specific studies

Explore how an individual heart's shape and properties affect how it functions.

Calibration & sensitivity

Fit models to measurements and study how outcomes shift as settings change.

Imaging-derived models

Turn imaging — including ultrasound video — into a working model to simulate.

Reproducible benchmarks

Establish reference studies that others can reload and reproduce exactly.

Contact & citation

heArtx is developed by Vahid Ziaei-Rad. For questions, collaboration, or access, contact ziaeirad@msu.edu.

If you use this platform in your work, please cite the accompanying preprint: doi.org/10.21203/rs.3.rs-9204887/v1.