Skip to content

Empowering breakthroughs in cancer care through faster and smarter response assessment

Plan a pilot Review the evidence
Illustrative radiology monitor displaying chest CT views with a linked enlarged detail

Know whether the drug is working by week 6.

An earlier answer means more time to act.

First dose

With Oncometra AI

Week 6

Without Oncometra

Months later

Waiting for mature outcomes

A single CT view with a magnifying glass enlarging a detail of the same scan

The same week-6 scan.Oncometra AI

Advance

Back promising
treatments.

Focus investment on the next phase of development.

Adapt

Refine the
trial.

Revisit the cohort or study design while you can act.

Stop

Protect time
and budget.

Stop unpromising arms and redirect resources.

Put resources behind the treatments that show promise.

Giving you the data to make decisions, faster.

For Clinical Investigation Centers

Explore treatment response in greater detail.

Pair a continuous score with the standard-of-care read to study response patterns and refine trial analysis.

  • Audit trail

    Every output paired with its inputs and confidence intervals.

  • Routine follow-up imaging

    No new scans or protocol amendments needed

For Oncology Care Teams

Bring more context to the treatment discussion.

Give oncologists and radiologists an additional perspective on treatment response when discussing whether to continue the current treatment or consider a change.

  • Signal at week 6

    after therapy initiation

Using
the score.

Response, redefined as a spectrum.

A continuous, calibrated number with a confidence interval on outcome.

  • Patient-level predicted benefit

    A 0–1 signal from baseline and early follow-up imaging.

  • Cohort-level trial outcome

    An early view of where the arm is heading.

  • A minimum confidence threshold

    If confidence falls below the threshold, no score is displayed.

Explore the use cases

Stable Disease

Tumor volume, RECIST reads, and the AI benefit score across six timepoints.

BASELINE TUMOR VOLUME · % OF BASELINE 200 100 0 t0 t1 t2 t3 t4 t5 RECIST AI BENEFIT BASELINE 0.50 SD 0.72 SD 0.80 PR 0.88 PR 0.93 PR 0.97
200 100 0 t0 t1 t2 t3 t4 t5

RECIST READ · AI BENEFIT

  1. t0 BASELINE 0.50
  2. t1 SD 0.72
  3. t2 SD 0.80
  4. t3 PR 0.88
  5. t4 PR 0.93
  6. t5 PR 0.97

Pseudoprogression

An early rise in tumor volume, followed by a decline.

BASELINE TUMOR VOLUME · % OF BASELINE 200 100 0 t0 t1 t2 t3 t4 t5 RECIST AI BENEFIT BASELINE 0.50 PD 0.68 PD 0.78 PD 0.88 PD 0.95 PD 1.00
200 100 0 t0 t1 t2 t3 t4 t5

RECIST READ · AI BENEFIT

  1. t0 BASELINE 0.50
  2. t1 PD 0.68
  3. t2 PD 0.78
  4. t3 PD 0.88
  5. t4 PD 0.95
  6. t5 PD 1.00

Early Cohort Forecasting

Drug A and Drug B, viewed from week 6 through month 18.

SURVIVAL · % OF ARM 100 50 0 MO 0 MO 6 MO 12 MO 18 Prediction generated at week 6 NOT YET OBSERVED Drug A Drug B Synthetic SOC WEEK-6 PREDICTION Drug A 86% Likely to outperform SOC Drug B 12% Unlikely to outperform SOC
100 50 0 MO 0 MO 9 MO 18 Week 6 A B SOC

WEEK-6 PREDICTION

  1. Drug A 86% Likely to outperform SOC
  2. Drug B 12% Unlikely to outperform SOC

The therapies we study.

Four treatment classes, each with its own
response signature on imaging.

An intravenous medication bag hanging above its drip chamber.

Chemotherapy

Volumetric and textural change.

A blister pack of plain tablets with one tablet beside it.

Targeted therapy

Lesion-level heterogeneity.

A glass medication vial with a blue cap and a capped syringe beside it.

Immunotherapy

Pseudoprogression and delayed response.

Two distinct medication vials standing together on a clinical tray.

Combinations

Interacting treatment mechanisms.

The images you collect.
The response you study.

Oncometra brings baseline and follow-up CT images together to assess treatment response.

Illustrative radiology workstation displaying chest CT images

  1. Start with routine CT scans.

    Baseline and follow-up images provide the starting point.

  2. Measure what changes.

    Compare the same lesions over time to assess treatment response.

  3. Review the findings.

    Explore patient and cohort readouts alongside the images behind them.

What the model sees

Five families of imaging features. From the tumor to the whole body.

Imaging biomarker / Lesion

Tumor volume

The three-dimensional extent of tumor tissue.

Imaging biomarker / Lesion

Tumor heterogeneity

Patterns and variation within tumor tissue.

Imaging biomarker / Lesion

Tumor shape

The geometry and contours of a tumor.

Imaging biomarker / Organs

Organ involvement

The distribution of lesions across organs.

Imaging biomarker / Whole body

Body composition

The composition of muscle and fat compartments.

The evidence.

Peer-reviewed research. We let the work speak for itself.

Our team’s research has appeared in
  • The Lancet Oncology
  • Annals of Oncology
  • JAMA Oncology
  • nature communications
  • Journal for ImmunoTherapy of Cancer
  • Clinical Cancer Research
  • JNCI Journal of the National Cancer Institute
  • The Journal of Nuclear Medicine
  • European Journal of Nuclear Medicine and Molecular Imaging
  • European Journal of Cancer
  • European Radiology

Journal names and mastheads are trademarks of their respective publishers. They indicate journals in which our research has been published, and do not imply endorsement of, or affiliation with, Oncometra.

The IMPACT Consortium.

International Multicenter Partnership for Assessing Cancer Treatment Response.

Leading cancer centers working together to assemble one large multicenter imaging dataset that spans multiple cancer types, a pan-cancer resource for understanding how tumors respond to treatment.

  1. Assemble

    A shared imaging dataset from participating cancer centers.

  2. Develop

    Robust imaging biomarkers and AI tools for treatment response.

  3. Validate

    Externally, across centers, scanners and cancer types.

Join the consortium
Multicenter network·Multiple cancer types·One shared dataset

A healthier world.

Findings go to peer review, not only to a data room.

Every method that reaches a claim is published, with its limits stated. Progress in reading response early belongs to everyone treating this disease.

Plan a pilot
A healthier world

About us.

Imaging scientists and oncologists who spent a decade reading response in trials.

The signal may already have been in the imaging. The work was bringing it forward in time, and making it auditable enough to act on.

We are passionate about cancer care and helping patients. Our team comes from leading cancer centers in Europe and the United States.

Contact us

Book a call.

Tell us what you are working on and what you need.

We will answer your questions and discuss the next steps with you.

Find a time to talk.

Book a call

You can also email contact@oncometra.ai.

Privacy Notice