Saved Results

No results saved yet.

Enter a patient name and hit Save on a result.

Clinical Reference
July 24, 2026
OncoToolkit Team

PREDICT Breast v3.1 Official Tool Guide

How to use and interpret the official PREDICT Breast v3.1 population estimates.

Evidence-Based Guide

PREDICT Breast v3.1 Official Tool Guide

Generate estimates at the official PREDICT Breast v3.1 website. OncoToolkit does not reproduce the PREDICT algorithm or provide an alternative result recorder.

What the official tool reports

PREDICT estimates survival after surgery for early invasive breast cancer and displays the potential survival benefit associated with selected adjuvant treatments. Its output includes:

Component effects are combination-dependent. Do not add the displayed component benefits independently; use the official combined-treatment output.

Supported scope

The documented source population is women aged 25-85 with early invasive breast cancer after surgery, before adjuvant treatment, and without distant metastases. DCIS and LCIS alone are outside scope.

Post-neoadjuvant and bilateral cases require particular caution because model inputs and transportability may not match an individual patient’s clinical situation. Check the current official instructions and disclaimers before use.

Interpretation boundary

PREDICT reports population estimates based on outcomes in similar women. The values are not patient-specific certainty and do not determine treatment. OncoToolkit does not:

The official PREDICT site and clinical team remain the source for generating and interpreting the result.

Sources

  1. Official PREDICT Breast v3.1 tool.
  2. PREDICT Breast overview.
  3. Official disclaimer.
  4. Official algorithm information.
  5. Official technical information.
  6. Winton Centre PREDICT v3 R package and MIT license.
  7. PREDICT v3 development and validation publication, PMCID PMC10789872.
  8. Comparative validation of PREDICT v3.1 and v2.2, PMID 41991871.

This page is a clinical reference to the official tool. It is not an independent implementation or validation of PREDICT.