Precision medicine tailors prevention and treatment to each patient's own genomic and molecular profile — the right drug at the right dose, the therapy matched to a tumour's biology, and inherited risk caught early enough to prevent. Ardia is building the decision-support that helps clinicians personalise that care, and the coverage layer that gets the enabling tests paid so patients aren't blocked from it.
Two people with the same diagnosis can need very different care. Precision medicine reads each patient's genomic and molecular signals to choose the therapy most likely to work, avoid the drugs that won't, and flag inherited risk early — better outcomes, fewer adverse events, and far less trial-and-error.
The barrier is access. These molecular tests are complex and costly, and independent molecular labs see a 23–31% average denial rate (industry benchmark · XiFin 2024) — with 65% of denials never appealed (MGMA 2023). Every denied test can be a patient who never gets the personalised answer their care depends on. Ardia exists to remove that barrier — and to build the decision-support on top of it.
Reads the patient's own pharmacogenomics — CYP2C19, CYP2D6, DPYD, TPMT, HLA-B — to flag drug–gene risks and guide a safer choice and dose (CPIC guidelines), so they avoid adverse reactions and medications that won't work for them. Decision-support for the prescriber; non-diagnostic.
Interprets a tumour's molecular profile — EGFR, ALK, BRAF, KRAS, HER2, MSI-H/TMB — and surfaces guideline-matched targeted therapies, immunotherapy and open clinical trials (NCCN, companion diagnostics) to support the oncologist's personalised plan.
Flags pathogenic hereditary variants — BRCA1/2, Lynch syndrome — with ACMG/AMP classification, then points to personalised screening, risk-reducing prevention and family cascade testing. Preventing disease, not only treating it.
Brings genomic, molecular, clinical and lifestyle signals into one evolving, personalised risk-and-prevention view — including liquid-biopsy and ctDNA monitoring for earlier detection and response tracking.
None of this reaches a patient if the test isn't covered. Ardia's reimbursement layer — MolDX & Z-code dossiers, medical-necessity documentation, NGS coding (81445/81455) and cited appeals — exists to remove that barrier so the personalised answer actually gets paid for. It's the plumbing, not the point: MolecuIQ is the model we're building for this (modelled target, non-diagnostic, human-in-the-loop), and the clinical decision-support above is on our roadmap.
What we're building first is the access & coverage engine — MolecuIQ plus deterministic code/dossier checks that get the enabling test paid. The clinical decision-support that personalises care — pharmacogenomic prescribing and molecular therapy-matching — is the roadmap: non-diagnostic, human-in-the-loop, grounded in CPIC, NCCN and ACMG.
Drafts variant, coding and medical-necessity rationale; non-diagnostic documentation and reimbursement support
Encodes CPIC PGx tables, ACMG/AMP tiers and CPT/PLA-to-MolDX code mappings as auditable deterministic rules
Checks MolDX/DEX Z-Code dossier completeness and NCCI plus CARC/RARC edits before claim submission
Build deterministic code-mapping and CPIC/ACMG rule tables, ground the MolecuIQ persona in NCCN/CPIC/MolDX retrieval with guardrails, and stand up dossier and PGx-alert checklists on public references only.
Pilot with a molecular lab under a signed BAA on a limited data set (LDS/DUA), human-in-the-loop, to validate dossier completeness and PGx-alert coverage — no autonomous billing or diagnosis.
Extend across all six therapeutic areas, add payer pre-submission automation, and integrate via HL7 FHIR into LIS/EHR workflows once pilots validate the approach.
Any predictive or device-like scoring is FDA-gated (510(k)) and deferred until validated — kept out of the shipped decision-support product.
In each area a person's molecular profile points to a different, better-targeted plan — and Ardia helps get the test that reveals it covered.
Interpreting a patient's molecular profile against NCCN and CPIC to support a personalised plan — and getting the enabling test covered. Modelled on synthetic data.
