Metastatic Uveal Melanoma
Rare disease with only ~1,000 new patients annually in the US. ~50% (~500 patients) are eligible based on HLA-A*02:01 positivity. While Immunocore’s treatment demonstrates strong efficacy in this biomarker-defined population, oncologists are often more familiar with broader, pan-tumor immunotherapies.
Objective: Ensure engagement with the right physicians at the right time. Position Immunocore’s treatment as a targeted treatment option for eligible patients.
Transformer based Prediction
Existing alerts have a limited window for action: data delays and physician access delays lead to missed opportunities. Transformer based prediction replaces manual steps, utilizes more data, and delivers alerts earlier.
- Evaluate all UM patients, identify those likely to metastasize in the next 6-months
- Learns from the full patient journey
- Captures timing, sequence, context
- Clear clinical drivers behind each prediction enable trust and adoption
- Deploy in 1-month
Summary of Results
How predictive alerts translated into measurable commercial impact
- 2.8x more accurate than existing Machine Learning based model
- 3x more alert volume than Lab and EMR alert providers
- 1.9x more digital engagement, communicating to the doctor at the right time
- 2.1x more F2F rep visits after alert and engaging with digital material
- 3-months lead time gained before metastasis and key decision time-period
- +10% gain in market share amongst doctors who engaged with NPP and see rep F2F
Next Steps: Reinforce what works and provide field team with reasons to believe. Integrate alerts with HCP intelligence to scale impact.
Presenters: Eric Chung (Prospection) and Matthew Shindel (Immunocore). 2026 Annual Conference, Hyatt Regency, New Orleans, LA.
