Building biological world models for precision oncology.
We predict how a patient’s cancer will evolve — reading resistance from serial ctDNA months before it ever appears on a scan.
Read the trajectory, not the snapshot.
In blood, resistance rises along a curve. Span reads that curve as it forms — flagging the shift months before it crosses the imaging threshold.
AI learned language. Then vision. Then the rules of proteins. The next frontier is disease over time.
Not disease as static snapshots, but as living systems that evolve. Cancer is the proving ground — and drug resistance is where those trajectories are most learnable.
Care is decided from snapshots. Disease moves in trajectories.
A single biopsy, scan, or genomic report captures one moment in a moving disease. Yet more than 240 targeted therapies exist and nearly all eventually stop working — and resistance follows predictable biological pathways, often detectable in blood long before it shows on imaging.
Lead time by which serial ctDNA can detect disease progression before it becomes visible on standard imaging.
A longitudinal prediction layer on the tests hospitals already run.
Span turns isolated data points — serial ctDNA, EHR context, and standard-of-care imaging — into a patient trajectory. It doesn’t replace or compete with the assays you already use. It sits on top of all of them.
Find the right patient, earlier.
Identifies patients approaching specific resistance mechanisms and eligible for trials — weeks to months before standard screening would catch them.
See the next line before it’s needed.
Estimated time-to-resistance, the most likely resistance mechanism, and a ranked list of next-line options — including eligible trials.
From data points to a trajectory.
Serial signals over time
Serial ctDNA every 6–8 weeks, EHR context, and standard-of-care imaging.
A world model of disease
Learns how each patient’s disease evolves, not just where it stands today.
A decision, ahead of time
Time-to-resistance, the likely mechanism, and ranked next-line options and trials.
The window just opened.
Every patient makes the model better.
Span learns from the shape of disease over time — patient by patient, trajectory by trajectory. The more patients it helps, the sharper its predictions become.
Built by a small, focused team.
Leads Span’s vision and the partnerships that turn serial data into clinical decisions.
Builds the trajectory model and the systems that make each patient’s data compound.