POAS
Six layers. Each with a defined analytical job.
POAS combines six complementary layers to organise the evidence in an existing molecular report: what is reported, how it relates to the cancer and the clinical question, what therapeutic or trial evidence may warrant review, what access questions remain and where uncertainty is unresolved.
Layer 01
CloneTrace
Variant distribution and tumour content
Reads the reported variants against the specimen and technical context, so each finding is interpreted with the quality and tumour content of the sample in view.
What it examines
- Reported SNVs and indels
- Variant allele fractions
- Copy-number findings
- Tumour fraction or purity estimates and specimen notes, where the report provides them
Questions it asks
Are observed variants broadly consistent with the reported specimen and tumour content? Could clonality or copy-number context affect how a finding is read? Are technical limitations or missing assay details relevant?
Potential output
A structured variant landscape with tumour-content context, copy-number notes, discordance flags and questions for the original laboratory or treating team.
How uncertainty is handled
Fields that are missing, unavailable or not applicable are labelled as such rather than inferred. Apparent discordance is flagged as a question, not resolved by assumption.
Evidence limitations
Interpretation can only be as complete as the supplied report. Tumour-content estimates vary between methods.
What CloneTrace does not establish
Variant allele fraction alone does not establish germline origin, clonality, pathogenicity or treatment actionability. POAS does not independently revalidate the original assay unless that is explicitly in scope.
Layer 02
PhenoMap
Tumour biology and molecular context
Places selected alterations in the context of tumour biology: the pathways, biomarker classes and disease-specific evidence relevant to this cancer.
What it examines
- Selected alterations mapped to tumour-associated pathways
- Biomarker classes and known molecular associations
- Disease-specific biological context
- The sources available for the case
Questions it asks
Is the alteration plausibly relevant to tumour biology in this cancer type? Is it a potential driver, a background finding or unresolved? Is the cited association disease-specific or extrapolated from another setting?
Potential output
A context-aware map of potential biological relevance, with associated biomarker classes, evidence labels and a note on competing interpretations.
How uncertainty is handled
Direct and extrapolated evidence are labelled separately. Competing interpretations are stated rather than collapsed into one.
Evidence limitations
Biological knowledge is uneven across genes and tumour types; rare alterations may have little evidence.
What PhenoMap does not establish
A pathway association or database annotation is not proof that a finding drives the individual tumour.
Layer 03
ImmunoLens
Immunotherapy-associated evidence
Structures immunotherapy-associated biomarkers with the assay and tumour context needed to interpret them responsibly.
What it examines
- Tumour mutational burden (TMB)
- MSI or MMR status
- PD-L1 and its scoring method
- HLA-related and other applicable biomarkers, where present
Questions it asks
Was the biomarker measured or inferred? Are the assay, scoring method and threshold appropriate to this tumour type and treatment context? Are required elements missing?
Potential output
An organised summary of available immunotherapy biomarkers, with assay-specific and tumour-specific caveats and follow-up questions.
How uncertainty is handled
Measured and inferred values are kept distinct. Thresholds are reported with the assay and tumour type they were validated for.
Evidence limitations
Biomarker values are not interchangeable across assays or tumour types.
What ImmunoLens does not establish
A biomarker does not by itself predict that a patient will benefit from immunotherapy.
Layer 04
SynerGX
Therapeutic evidence
Organises potentially relevant therapeutic associations by evidence strength, so options can be considered in order of how well they are supported in this cancer and setting.
What it examines
- Therapeutic associations with reported alterations or biomarkers
- Potential targeted options and combinations discussed in the literature
- Resistance or bypass considerations
- Strength and type of evidence for each association
Questions it asks
Is the association supported in this cancer and disease setting? Is the evidence guideline-supported, regulatory-approved, from a clinical study, preclinical or hypothesis-generating? Are there caveats or conflicting data?
Potential output
An evidence-aware hierarchy of possibilities for clinician consideration, with source and evidence-level distinctions rather than a flat list.
How uncertainty is handled
Each association carries its evidence level and setting. Conflicting data are shown alongside supporting data.
Evidence limitations
Evidence can change quickly; statements are dated. Availability and suitability are separate questions.
What SynerGX does not establish
SynerGX is not an automated prescription engine. It does not predict that a medicine will work or establish that it is available or suitable.
Layer 05
TrialGraph
Clinical-trial intelligence
Identifies potentially relevant clinical trials and organises their eligibility criteria against the information supplied, so the treating team knows exactly what to verify.
What it examines
- Relevant trial records and protocol details
- Tumour type, molecular requirements and disease stage
- Prior-treatment and measurable eligibility statements
- Location and recruiting status, where available
Questions it asks
Does the described cancer and biomarker profile appear to fit the published criteria? Which inclusion or exclusion items are unresolved? Are age, performance status, organ function, prior therapy or geography requirements missing?
Potential output
A shortlist of potentially relevant studies with identifiers, key eligibility points, known site information and an explicit checklist for the trial team to confirm.
How uncertainty is handled
Each criterion is marked as potentially satisfied, contradicted or unknown. The registry source and retrieval date are recorded.
Evidence limitations
Recruitment status, site activity, slots and protocol criteria change frequently.
What TrialGraph does not establish
A computationally identified match is not confirmation of eligibility, enrolment, access or clinical benefit.
Layer 06
AccessMatch
Treatment-access context
Adds practical context about whether a therapy under discussion is approved, available and affordable in India, and which facts still need confirmation.
What it examines
- Regulatory status for the proposed indication
- Treatment availability and supply routes in India
- Indicative costs, dated
- Potential access mechanisms, such as a named-patient pathway where legally and operationally applicable
Questions it asks
Is a therapy approved for the proposed indication or being discussed in another context? Is a reliable Indian supply route known? Are price estimates dated? What eligibility, paperwork, funding or logistics still need confirmation?
Potential output
Dated access notes that distinguish known information, estimates and facts that must be verified with the regulator, institution, pharmacy, supplier or access programme.
How uncertainty is handled
Every price and status note is dated and labelled as known, estimated or to be verified.
Evidence limitations
Costs, supply, approvals, import permissions and access-programme terms can change.
What AccessMatch does not establish
POAS does not guarantee affordability, supply or access to any medicine or trial.
Cross-layer synthesis and report writing
- The clinical question and the available diagnosis, histology, stage, prior treatments and response are integrated; a molecular result is never interpreted in isolation.
- Detected findings, evidence-supported associations, hypotheses, missing information and unresolved uncertainties are kept visibly distinct.
- The report summarises what it supports, what it does not establish, what further information may be useful and which questions should go to the treating clinician or MTB.
- The date and sources used for trial and access checks are stated. The report does not imply that every source or tool was run for every case unless the workflow actually did so.
Core safeguard
A detected alteration is not automatically actionable. A POAS report supports the treating clinician’s assessment; it does not prescribe treatment, independently determine the care plan or guarantee response, trial eligibility or access.
Short explainers, in production
- POAS in 60 secondsAbout 60–75 seconds