Not a probability
The ratio is relative susceptibility on the engine's scale. It is not the chance that a patient has or will develop cancer, and no per-patient percentage is computed from it.
Before the first symptom
Cancer begins a decade before you feel it.
AEGIS helps you read a patient's risk inside the symptom-free window, while there is still time to act.
02The blind spot
Most cancers have no recommended screening test.
Routine screening covers four cancers. Most others are typically found when a patient already has symptoms.
4
Cancers with a broadly recommended screening test2
57%
Of diagnosed cancers have no recommended screen3
70%
Of cancer deaths come from those unscreened cancers3
Risk, not the calendar, should decide who you screen and how. That judgment stays yours. AEGIS informs it.
03The convergence
Four signals, read together.
HistoryGermline
genomicsImaging
reportsWearable
summaries
genomicsImaging
reportsWearable
summaries
Cancer Susceptibility Index (CSI): a per-cancer risk tier derived by a deterministic rules engine from published, peer-reviewed evidence, for review by a licensed healthcare professional. Not a diagnosis, and not a statement that a patient has or will develop cancer.
What feeds the index, and what stays separate
History, germline genomics, imaging reports and wearable-derived metrics feed the index. Imaging: Radiologist-reported findings, as reported. Wearables: Summary metrics, entered at discrete intervals. AEGIS does not acquire, process, or analyse medical images or continuous physiologic signal.
Genomic results are shown as the ordering laboratory reported them. Physician-ordered labs, MCED and other ordered results sit alongside the profile for the clinician to interpret, not folded into the index.
Biological age is estimated from routine blood results using a published method, on its own investigational tab. It is never part of the cancer score.
04The AEGIS loop
From information to action.
Then round again.
- 01New informationRecords and results.
- 02Cancer Susceptibility IndexRisk tier per cancer.
- 03Action itemsPhysician-reviewed plan.
- 04Follow-throughAct. Update. Repeat.
New information starts the next cycle.
The profile is re-scored at set intervals as new results are entered. Physician feedback informs governed updates to the rules; nothing retunes itself between releases.
05The index
One number per cancer, and the reasons behind it visible.
Each cancer is scored on its own. A patient's risk factors for that cancer combine into a single score, and that score sets one of five tiers (Low, Lowered, Average, Modestly elevated, Elevated). Nothing is weighted by hand, and the same inputs always give the same tier. The method itself is proprietary and is not described here.
Illustrative example, one cancer. Values are not engine values.
Factors that apply
Combined, illustrative
1.00
Start
Tier
Not self-tuning
The rules are versioned and change-controlled. Physician feedback informs governed updates to them. Nothing retunes itself.
Illustrative example. The factor names, values and tier above are made up to show the shape of the index; they are not engine values, and the tier scale shows names only. A relative-susceptibility index is not a cancer probability.
Trace any recommendation back to its evidence.
Follow it to the factor that triggered it, the strength of that evidence, and the literature it rests on.
This example is colorectal, from the 24-cancer rules engine. Tiers from the two published calculators (breast, lung) are traceable the same way, and the console names which method produced any given tier.
What the clinician can stand behind
Consider colonoscopy surveillance per guideline
With genetic counseling and a cascade-testing discussion, per NCCN
because of
Triggering factor
Lab-reported pathogenic MLH1 classification
Germline, Lynch syndrome, as reported by the ordering lab
graded by
Strength of evidence
Established causation
Hereditary syndrome, high penetrance
rooted in
Source
Peer-reviewed literature
Full citation attached in the live profile
You stay the physician. AEGIS surfaces the risk and shows its evidence. The diagnosis and the decision remain yours.
References
- Timing of driver mutations: Gerstung M et al. The evolutionary history of 2,658 cancers. Nature. 2020;578:122-128. doi:10.1038/s41586-019-1907-7. Yachida S et al. Distant metastasis occurs late during the genetic evolution of pancreatic cancer. Nature. 2010;467:1114-1117. doi:10.1038/nature09515. Back
- US Preventive Services Task Force. Cancer screening recommendations graded A or B: breast, cervical, colorectal and lung. Prostate screening is graded C for men aged 55 to 69, an individual decision, and D for men 70 and older. Checked 24 September 2026. Back
- NORC at the University of Chicago. New research highlights just one in seven diagnosed cancers found by a recommended screening test. 14 December 2022; estimates for 2017. The analysis was funded by GRAIL, which makes a multi-cancer early detection test. A peer-reviewed analysis with GRAIL-affiliated authors estimated that 31.4% of US cancer deaths in 2018 and 2019 came from cancer types with a recommended screening test: Ofman JJ et al. Cancer Biomarkers. 2025;42(1). doi:10.1177/18758592241308754. Back
General information, not medical advice. Figures are drawn from the public sources cited and are provided for context. The proportion-of-deaths figures depend on which cancers are counted as screened and over what period, so treat them as an order of magnitude rather than a precise share. AEGIS is clinical decision support, not a diagnostic device, and does not replace medical care.