Our mission

Genomic answers, within reach of every patient.

Targeted cancer therapy depends on knowing a tumor's mutations, yet most patients never get that information. Imagenomix exists to change that, using AI to predict genetics from the pathology slides every patient already has.

The gap we exist to close

The gap

A diagnosis is not the same as an answer.

Knowing which mutations drive a tumor is what makes targeted therapy possible. For most patients, that knowledge never arrives, blocked by cost, tissue, or turnaround.

~0% in the United States

never receive molecular testing.

Tested (~20) Left behind (~80)
~0% worldwide

never receive molecular testing.

Tested (~2) Left behind (~98)
~0%

of samples are insufficient for sequencing — even patients who are tested can be turned away.

Imagenomix closes the gap

0%

IGX Predict runs on every confirmed cancer case — reading actionable genetics from the standard H&E slide already on the bench.

These are genomically underserved patients. Imagenomix exists to bridge that gap.

Why it matters: when a tumor's drivers are known, patients can be matched to targeted therapies that meaningfully improve outcomes. Testing is the door to that treatment, and today it stays closed for most.

Access to molecular testing, United States and worldwide.

A first in cancer AI

The first to predict cancer mutations from a single image.

Imagenomix pioneered predicting a tumor's genetics from its histology, then proved it, patented it, and brought it to the clinic.

  1. 2017
    DiscoveredHistology-based mutation prediction shown at NYU Langone.
  2. 2018
    First to publishCoudray et al., Nature Medicine.
  3. 2018 to 2022
    ProvenValidated across multiple cancer types.
  4. 2022
    PatentedUS Patent 11,367,180.
  5. 2023
    CommercializedIGX Predict goes live.

How it works

The answer is already in the tissue.

Every test begins with an H&E stained slide. Imagenomix partitions that whole slide image into thousands of tiles, scores each one for the molecular signatures of driver mutations, then maps the result back onto the slide.

No DNA extraction. No additional tissue. The genetics, predicted directly from the image.

Standard H&E stained whole-slide image before analysis
Standard H&E slide
AI-predicted mutation probability heatmap for the same slide
Predicted mutation map

Why IGX Predict

Fast. Affordable. Accessible.

Minutes
Fast

Predictions in minutes, not the weeks sequencing can take. Treatment decisions cannot wait.

Fraction of NGS
Affordable

Computational, not wet-lab, bringing molecular insight within reach of every confirmed cancer case.

Any H&E slide
Accessible

Runs on the standard H&E slide every patient already has, using the scanners and image systems labs already own.

For research use. Performance varies by product, see each product page for validation detail.

The platform

IGX Predict™
One platform. Four products. Four cancer types.

IGX Predict is the platform. Its engines prepare each case, then a tumor-type product predicts the driver mutations from the standard H&E slide.

IGX Clear
Cleans the input, normalizing staining, scanner, and resolution before analysis.
IGX Seek
High sensitivity to catch every positive, built for screening and trial enrichment.
IGX Focus
High specificity for the most reliable call, built for trials and companion diagnostics.

Four products across four cancer types. Pulmogene is commercial. BLISS, NeuroInsight, and DermPredict are in development.

Imagenomix cancer detection visualization

Access and equity

No patient left untested.

Health outcomes should not depend on where a patient lives or where their sample is sent. Our models are trained across diverse patient populations so they perform accurately across racial and ethnic backgrounds, not only in well-resourced cohorts.

Imagenomix was selected for the AWS Health Equity Initiative to scale IGX Predict toward genomically underserved patients in every country, the foundation for reaching patients who would otherwise have no path to molecular testing at all.

Where we began

Spun out of NYU Langone.

Imagenomix grew out of research at NYU Langone Health showing that a tumor's genetics leave readable traces in its histology. We licensed that technology, assembled a team of AI scientists, pathologists, and oncologists, and set out to bring it to patients.

Patent-protected scienceTechnology discovered in 2017 at NYU Langone Health and validated across peer-reviewed studies. US Patent 11,367,180.
Built to run on any H&E slideNo special preparation and no additional tissue, using the scanners labs already own.
Image management system agnosticIntegrating with the leading digital pathology and image management systems used globally.

Work with us

Two ways to bring genetics to every image.

Pathology team reviewing digital slide images

For pathology practices

Molecular insight in minutes

Run IGX Predict on confirmed cancer cases and return driver mutation predictions with spatial heatmaps, without sending tissue out or waiting weeks.

See the platform
Archival cancer tissue samples used for biopharma screening

For biopharma partners

Screen archives, enrich trials

Prescreen archival slides for molecular drivers to find eligible cases faster and spatially map heterogeneity before more expensive sequencing.

Explore partnerships

Help us reach the patients no one is testing.

Whether you are a clinician, a sponsor, or a future colleague, there is a place for you in this. Reach the team to run a case, pilot the platform, or partner on a trial.

Imagenomix Corp · New York, NY · Contact