Many medicines work well for some patients and poorly for others. Giving treatments to people who will not benefit costs time, money and, most importantly, a patient's chance to try something that would work. We design every programme around the patients most likely to respond.
Patients are grouped by shared biological features — a gene mutation, a protein level, a tumour subtype — so a treatment is tested and used where it is most likely to help.
The same principle taken further: decisions on whether and how to treat that reflect an individual's own biology, history and response.
Each candidate is linked to a measurable biomarker before it advances.
Studies enrol the patients the biology predicts will respond, so smaller trials can give clearer answers.
We plan how patients will be identified in practice, not just in trials.
Response data feed back to refine who benefits, moving from groups towards individuals.
| Programme | Disease | Stratification marker |
|---|---|---|
| KLS-101 | DIPG / diffuse midline glioma | H3 K27-altered tumour status |
| KLS-201 | Frontotemporal dementia (FTD-GRN) | GRN mutation carriers; progranulin levels |
| KLS-301 | Chondrosarcoma (advanced / metastatic) | IDH1 / IDH2 mutation status and tumour subtype |
Some compounds have been set aside because their benefit was diluted across a broad, mixed population. When the patients who do respond share a measurable feature, a medicine that looked marginal can become meaningful.
Combining repurposing with stratification is especially powerful in rare disease, where every patient counts and trials must be small, precise and well designed.
Our rare disease initiative →