Genomate Health

The computational model is the product. The platform calculates a cumulative score for each match between a drug and a genomic profile, ranks therapies targeting cancer driver alterations, and returns a ranked list of drugs most likely to be effective for that individual. Without the scoring model there is no output, only a variant list the oncologist already has.

Among the better designed oversight postures in this index, on two counts. Structurally, the output feeds molecular tumor board recommendations rather than reaching the prescribing decision directly, and clinics in Central and Eastern Europe incorporated it into that workflow. Architecturally, the company describes its tool as open box AI in contrast to black box models, meaning the reasoning behind a ranking is inspectable. For a system telling an oncologist which drug is most likely to work, inspectability is the difference between decision support and an oracle.

The method is described concretely rather than in marketing terms: a knowledge model performing automated data compilation, evaluation, and synthesis, scoring each drug to genomic profile match and producing a cumulative rank. The company's own open box framing is a transparency claim it invites scrutiny against. Peer reviewed and conference publications describe the model's performance. Falls short of an A because no technical specification, feature set, or model documentation is published for independent audit.

Evidence is validated against trial data rather than asserted, which is rare in precision oncology decision support. The company reports that in the SHIVA01 clinical trial, patients who received therapy recommended by its model had a 4X higher positive response rate than those whose therapy was selected on the presence of a single biomarker but not recommended by the model. A published performance analysis in that trial exists, alongside a real world performance analysis in pediatric tumors and an ASCO abstract evaluating the model's clinical performance in immune checkpoint therapy response prediction for lung cancer. Selection into the Mayo Clinic Platform_Accelerate program adds an independent validation signal. Caveat for buyers: the SHIVA01 analysis is retrospective application to trial data rather than a prospective trial of the tool itself.

No published data protection or PHI stewardship framework was located. The platform processes tumor genomic profiles, which are identifiable and among the most sensitive data classes, across both European and United States operations.

No HIPAA or BAA commitment located. The company entered the United States market from a European base, so a US institution should establish both HIPAA standing and GDPR interaction terms directly.

No SOC 2, ISO 27001, or equivalent attestation was located in public materials.

No FDA clearance is claimed and no regulatory pathway is disclosed. This is the central open question for the product: software that ranks therapies by predicted response sits close to the Software as a Medical Device boundary, and the tool's positioning as decision support informing a tumor board rather than directing treatment is what keeps it on the safer side. European regulatory status, including any CE marking under IVDR or MDR, was not located either. A buyer should establish the regulatory basis for their jurisdiction before clinical deployment.

No formal governance program is published, but the open box design is itself a partial answer to the governance question, since an inspectable model can be audited by the clinicians using it. Participation in the INSTAND-NGS4P initiative with clinical testing across three European sites indicates engagement with external standardization. No bias evaluation across populations or tumor types was located, which matters given the model is trained on published research evidence that carries its own representational skew.

Integration is with the genomics pipeline rather than the chart. The company works with next generation sequencing reporting partners on a joint solution for precision oncology reporting, tested across three European clinical sites, which is meaningful interoperability in the diagnostic workflow. No EHR integration, HL7, or FHIR support was located.

No hosting, tenancy, or residency disclosure was located. The company operates from Cambridge Massachusetts with a satellite in Budapest and serves clinics in both Europe and the United States, so cross border data flow terms are a live question a buyer must raise.

No published pricing. The company offers both a clinician facing decision support platform and a genomics driven cancer second opinion service for patients, and neither carries public pricing.

Single specialty, which is oncology, and within it the scope is targeted therapy and immune checkpoint therapy selection based on molecular profile. Published evidence spans solid tumors, pediatric tumors, and lung cancer immune checkpoint response. Geographic footprint covers Central and Eastern European oncology clinics with a stated United States market entry, so US deployment references are thinner than European ones.