SP
Loading FHIR patient…
Oncology context · source linked
Loading… Loading… Loading… Loading…
FHIR context loaded
Patient overview

Clinical context, evidence, and downstream use

Source-linked facts remain separate from derived outputs.

FHIR R4 · SMART + PKCE · mCODE · Genomics
Current clinical context

One structured oncology record supports multiple downstream workflows.

Source provenance and missingness remain visible while destination-specific rules govern trial, registry, and population outputs.
Source retained · uncertainty explicit · release rules applied
Retrieve SMART · FHIR REST · Bulk
Standardize mCODE · Genomics
Evaluate Evidence · Policy
Release Trials · Registry · Population

Destination routing

patient-level source class
Structured
Oncology Record
provenance retained
Candidate trials criterion evidence
READY
Cancer registry CCRR candidate
VALIDATE
Population analysis patient-level data blocked
BLOCKED
Audit record release traceability
TRACE
6
structured oncology facts
4 observed · 2 unknown
3
candidate trials identified
0 structured contradictions
100%
released facts with provenance
synthetic fixture
1
privacy route blocked
patient → population prohibited

Longitudinal oncology record

source-linked · longitudinal · missingness visible
Source
Diagnosis
Loading…
Loading…
Source
Staging
Loading…
Loading…
Source
Genomics
Loading…
Loading…
Current
Trial routing
candidate evidence
human review required
Current
Registry
candidate bundle
validation gate
Current
Population
aggregate only
privacy release

Evidence ledger

click source to verify
Dx
Primary cancer Loading…
St
Stage Loading…
Gx
Genomic finding Loading…
PS
ECOG performance status No source fact available
UNKNOWN

Provenance graph

FHIR source → structured fact
Condition Loading…
Observation Loading…
Observation Loading…
Patient Loading…
Structured record source links retained
Structured oncology record

Longitudinal record with source-level verification

Facts, missingness, and transformations remain inspectable throughout the workflow.

Current structured profile

synthetic record
Concept Canonical value State Source
Primary cancer Loading… LOADING
Stage Loading… LOADING
Genomics Loading… LOADING
ECOG — UNKNOWN no source
Prior PARP inhibitor — UNKNOWN insufficient medication history

Clinical state completeness

not a risk score
72% evaluable facts
Observed 4 facts

Directly represented in the synthetic FHIR source.

Unknown 2 facts

Missing evidence remains visible and cannot silently become a match.

Transformation ledger

source → standardization → destination

FHIR source graph

Patient, Condition, Observation and Genomics resources remain authoritative.

FHIR R4
→

Structured oncology record

Standardizes clinical semantics while retaining source identifiers and uncertainty.

PROVENANCE
→

Registry reporting

Applies destination-specific policy, terminology, validation, and release rules.

VERIFY
Clinical trial pre-screening

Candidate trials with criterion-level evidence

Structured contradictions are excluded; unresolved criteria remain visible for protocol review.

Candidate queue

3 frozen synthetic candidates
NCT-SYN-001 0.91

BRCA-associated ovarian cancer maintenance study

Recruiting 18 mi Phase III
NCT-SYN-002 0.78

Platinum-sensitive ovarian cancer precision study

Recruiting 42 mi Phase II
NCT-SYN-003 0.66

Advanced gynecologic malignancy basket study

Recruiting 63 mi Phase II

NCT-SYN-001 evidence matrix

candidate only · final eligibility requires protocol review
Diagnosis

Candidate criterion; compare with loading source…

REVIEW
Loading…
Age

Candidate criterion: 18–80 years

LOADING
Loading…
Administrative sex

Candidate criterion: female

LOADING
Loading…
Genomic criterion

Candidate criterion; compare with loading source…

REVIEW
Loading…
ECOG ≤2

Performance status required

UNKNOWN
No source fact
Prior PARP inhibitor

Protocol-specific exclusion review

REQUIRES REVIEW
Medication history incomplete
4/6
criteria directly evaluable
unknowns remain explicit
0
structured hard contradictions
age and sex filters passed
2
criteria requiring human review
missing criteria remain explicit
Registry reporting

Registry candidate with validation before release

The same source facts support registry reporting without a second oncology abstraction workflow.

Registry reporting pathway

candidate package · not yet declared CCRR conformant

Source FHIR

Patient + primary cancer + stage + genomics + treatment evidence.

SOURCE
→

Registry candidate

The registry candidate preserves references to the original source resources.

ASSEMBLED
→

Validation gate

Published CCRR profile validation is required before a conformance claim or release.

PENDING EXTERNAL TEST

Source-to-report traceability

field lineage
Registry concept Source State
Primary site / disease Loading… TRACEABLE
Stage Loading… TRACEABLE
Genomic finding Loading… TRACEABLE
ECOG Unavailable MISSING

Release guardrails

Profile validation REQUIRED

Critical profile failures block a conformance claim.

Provenance completeness 100%

Each current synthetic fact resolves to a source reference or explicit missingness.

Version pinning ON

Published baseline and implementation-guide versions are declared rather than silently tracking latest.

Population data release

Aggregate geography without exposing patient locations

Patient-level routing is blocked before the map layer; only policy-released aggregates are rendered.

Synthetic regional aggregates
illustrative aggregate geography · not patient points
k=11 release rule
higher released aggregate
intermediate aggregate
suppressed / generalized

Privacy release gate

Direct identifiers REMOVED

Population products do not receive direct identifiers.

Patient coordinates PROHIBITED

Raw geocoded points never enter the browser map product.

Small cells SUPPRESSED

Low-count cells are suppressed or generalized before release.

Environmental context AREA-LEVEL

Contextual measures are not represented as measured individual exposure.

1
patient-level route prevented
patient-level release blocked by policy
Standards & validation

Implementation and validation status

Implemented code, local tests, and external validation are deliberately separated.

FHIR R4
4.0.1
SMART App Launch
PKCE scaffold
mCODE
STU4 target
Genomics Reporting
STU3 target

Release evidence

local package validation
Gate Status Interpretation
Repository self-audit PASS package structure, claim boundaries, public ports
Python tests PASS unit/integration synthetic scenarios
C++ kernels PASS distance primitives
SMART App Launch test PASS recorded public-client sandbox run; receiving EHR acceptance still required
HL7 mCODE STU4 / Genomics STU3 PASS* frozen candidate: 0 fatal, 0 error; retained terminology warning
CCRR validation PASS* STU1 candidate: 0 fatal, 0 error, 1 retained histology warning

Audit log

prototype data
00:00:01 FHIR bundle received · tenant=setroja-prototype
00:00:01 primary cancer normalized · source=Condition/cancer-001
00:00:01 stage normalized · source=Observation/stage-001
00:00:01 genomics normalized · source=Observation/variant-001
00:00:01 ECOG unresolved · status=UNKNOWN
00:00:02 trial route permitted · human_review=true
00:00:02 registry route permitted · validator_required=true
00:00:02 population route BLOCKED_BY_POLICY · transform=aggregate_then_privacy_release
Care plan

Coordinated oncology care plan

Care tasks, unresolved data, ownership, and source status in one workspace.

4
active care-plan items
owners and source status visible
1
documentation gap
ECOG remains unresolved
4
care-team disciplines
oncology · nursing · navigation · research

Current workflow plan

prototype tasks
Item Owner State
Review performance-status documentation
No structured ECOG source 		Care team OPEN
Review BRCA-associated trial candidates Research team READY
Registry candidate validation Registrar PENDING
Navigation screen follow-up Navigator OPEN

Care team

Medical oncology treatment oversight
Oncology nursing symptom follow-up
Patient navigation access barriers and referrals
Clinical research trial prescreen and protocol review
Symptoms & ePRO

Patient-reported symptom monitoring

Severity and change are translated into workflow priority while clinical decisions remain with the care team.

Weekly symptom report

synthetic PRO data
Nausea CLINICAL REVIEW
Fatigue MONITOR
Neuropathy MONITOR
Pain MONITOR

Workflow state

Highest priority CLINICAL REVIEW

Nausea meets the configured prototype workflow threshold.

Patient contact NOT RECORDED

Follow-up remains a care-team action, not an automated treatment decision.

Instrument governance LOCAL CONFIG

Real-patient thresholds require validated instruments and institutional approval.

Tumor board

Multidisciplinary case review

Source facts, unresolved questions, and multidisciplinary recommendations are kept distinct.

Pre-meeting case summary

source-linked
DX
Loading… Loading…
OBSERVED
ST
Loading… Loading…
OBSERVED
GX
Loading… Loading…
OBSERVED
PS
Performance status No structured source
UNKNOWN

Multidisciplinary review

Medical oncology Surgical oncology Pathology Radiology Genetics Oncology nursing Pharmacy Research Navigation Palliative care
Board recommendation

Entered by the multidisciplinary care team. Recommendations are versioned and remain separate from source FHIR facts.

Navigation & access

Access barriers and care navigation

Transportation, financial, social, language, distance, and support needs can be surfaced for follow-up.

2
active barriers
transportation · care distance
1
high-priority need
transportation
3
support contacts
navigator · social work · financial counseling

Access screen

whole-person oncology
Transportation NEED IDENTIFIED
Financial concerns NO CURRENT NEED
Food security NO CURRENT NEED
Housing NO CURRENT NEED
Language support NO CURRENT NEED
Care distance BARRIER
Caregiver support AVAILABLE
Digital access AVAILABLE
Prior authorization NO ACTIVE DELAY
Insurance coverage DOCUMENTED
Survivorship & supportive care

Continuity across the cancer trajectory

Supportive care, goals of care, palliative-care status, and survivorship planning remain visible without replacing clinician judgment.

Supportive care

Symptom monitoring ACTIVE

Patient-reported symptoms flow to the clinical review workspace.

Palliative care AVAILABLE

Referral status and goals-of-care documentation can be represented independently of prognosis.

Goals & preferences

Goals of care

No structured goals-of-care document is present in the current synthetic record.

NOT DOCUMENTED

Survivorship

The prototype stores survivorship plan items, late-effect domains, follow-up ownership, and source references. Diagnosis- and treatment-specific recommendations require governed clinical content.

PLAN STRUCTURE READY
Cohorts & research

Research cohort discovery

Structured disease, stage, genomics, and workflow variables can be reused for research without a second abstraction pipeline.

Cohort builder

Current synthetic cohort

1 matching patient
Patient 001 MATCH

Loading… · Loading… · Loading…

Decision support

Consultation support with source-linked facts

Patient-friendly questions and structured facts support shared discussion without generating an autonomous treatment recommendation.

What is known

from the structured record
DX
Loading… Loading…
OBSERVED
ST
Loading… Loading…
OBSERVED
GX
Loading… Loading…
OBSERVED
PS
Performance status No structured source
UNKNOWN

Questions for the consultation

  1. What is the goal of the current treatment plan?
  2. Which benefits, risks, and uncertainties are most important for this decision?
  3. Are there clinical trials that should be reviewed?
  4. What symptoms or treatment effects should prompt contact with the care team?
  5. What practical barriers could affect treatment or follow-up?
Clinical boundary

Outcome estimates and treatment recommendations require validated evidence, licensed clinical content where applicable, and clinician review.