Live moduleA PathCopilot product

Variant interpretation
for hemato-oncology,
drafted in seconds.

HemeCopilot turns NGS panel results into a structured, guideline-grade interpretation — AMP/ASCO/CAP tiers, ELN 2022/2024 and IPSS-M risk, co-mutation context, therapy signals, and germline flags. Every claim is cited. You review, edit, and sign.

AMLMDSMPNALLCLL
See the engine
cases / HEM-2026-0142AML · 54F · BM
ELN 2022 RISK
Intermediate
ELN 2024 RECLASSIFIED
Adverse
confidence0.84
INPM1p.W288fs38%
IFLT3ITDmidostaurin42%
IIDNMT3Ap.R882H45%
IIITET2p.Q1261*CHIP?11%
2 open questions for pathologist · TET2 at 11% VAF — CHIP vs somatic driver? · Confirm FLT3-ITD reporting policy.
14 citations · sonnet-4-5

Illustrative case — AI-drafted, pathologist-reviewed

The workflow
New case → signed report in one pass
01 INGEST

Inbox & intake

VCF import or manual entry. Smart Queue prioritizes by risk and TAT.

02 ENRICH

KB lookup

Variants matched against OncoKB, ClinVar, CIViC before drafting.

03 DRAFT

AI interpretation

Tiers, risk, co-mutations, therapy, germline flags — with citations.

04 REVIEW

Edit inline

Adjust any field. Every change versioned against the AI draft.

05 SIGN OUT

Report & audit

Typed signature, PDF export, complete audit trail. Amendments tracked.

Interpretation engine

Guideline reasoning, not autocomplete.

The drafting prompt is versioned product code that encodes AMP/ASCO/CAP tiering and disease-specific frameworks. It classifies each variant independently, surfaces evidence conflicts, and refuses to invent therapy claims without tier-I support.

Per-variant AMP/ASCO/CAP tiering

Tier I–IV with explicit oncogenicity, never blurred across variants. Insufficient evidence becomes a documented VUS.

Dual-framework risk + reclassification alerts

ELN 2022 and 2024 scored side by side; flags when the update changes the outcome. IPSS-M for MDS.

Co-mutation context engine

Detects patterns that modify risk beyond single variants — NPM1+FLT3+DNMT3A triplet, biallelic TP53, spliceosome signatures.

Germline predisposition flags

Auto-detects DDX41, RUNX1, GATA2, CEBPA and other WHO-recognized genes; routes them to ACMG germline review, not somatic tiers.

co-mutation analysis3 patterns
biallelic TP53adverse

Two TP53 hits (R175H 62%, R248W 8%) → multi-hit WHO 2022 entity. Very poor prognosis.

NPM1 + FLT3-ITD + DNMT3Atriad

Composite triplet → midostaurin during induction (RATIFY). Monitor MRD.

low-VAF TET2CHIP candidate

11% VAF below clonal threshold — flagged for pathologist judgment, not auto-reported.

Knowledge suite

A reference library wired into the bench.

Curated hemato-onc annotations from OncoKB, ELN 2022, WHO 2022 and NCCN — plus the calculators, wizards and workbenches pathologists reach for mid-case. Versioned, with a 30-day changelog.

Risk Calculators

ELN 2022 · IPSS-M scoring with live inputs.

Germline Wizard

DDX41 · RUNX1 · GATA2 predisposition workup.

VUS Workbench

Weigh evidence for variants of uncertain significance.

AMP Criteria

Tiering rules from Li et al. 2017, searchable.

RAG Knowledge Base

Retrieval over guidelines and the lab's own notes.

Trial Matcher

FLT3 · IDH · TP53 trial eligibility surfacing.

PGx Reference

Pharmacogenomic dosing context at a glance.

Snippets & Templates

Reusable note blocks and report scaffolds.

Flashcard Trainer

Drill AMP tiering and ELN criteria.

patients / clonal evolution
VAF trend · 3 serial biopsiesdx → +6mo → +14mo
TP53 rising — emergent cloneNPM1 clearing

Lab intelligence

The whole lab, in view.

Beyond the single case: longitudinal patient tracking, cohort analytics, and surveillance that watches for knowledge-base changes affecting cases you've already signed out.

Clonal evolution & VAF trends

Serial biopsies per patient, with emergent and clearing clones surfaced automatically.

Gene landscape & co-occurrence

Mutation trends, co-occurrence heatmaps, and panel performance across your cohort.

CHIP & MRD surveillance

Dedicated trackers for clonal hematopoiesis and measurable residual disease over time.

Reclassification scanner

When OncoKB or a guideline changes, flags signed-out cases whose interpretation may now differ.

AI & variant concordance

Track where pathologists agree with the AI draft and where they override — by gene and over time.

Workflow & collaboration

Built for how labs actually sign out.

Smart Queue ranks the backlog by urgency. Batch operations, routing rules, tumor board and external second-opinion sharing move cases through the team.

Smart Queue & inbox triage

Priority-sorted by risk, STAT flags and TAT — urgent cases never sink.

Tumor board & second opinion

Assemble MTB packets; share a read-only case link with an external reviewer.

Batch reinterpret & sign-out

Re-run a cohort against an updated prompt; sign out a batch in one reviewed action.

Compliance & audit

Documentation that's already done.

Every AI call, edit, amendment and sign-out is logged with model, prompt version and tokens — built for CAP/CLIA review, not bolted on after.

CAP/CLIA metrics & Lab QC

Concordance, amendment rate and QC indicators tracked continuously.

Immutable audit log

Full request payload, raw response, parsed output and token counts per call.

Report versioning & amendments

Signed reports are frozen; amendments create a linked new version with a reason.

Disease coverage

Multiple hemato-oncology contexts, each with its own framework.

AML is the deepest, with ELN 2022/2024 risk and composite-mutation logic. Each context applies the right classifier and the right driver set.

AMLELN 2022 · 2024 riskNPM1 · FLT3 · CEBPA · TP53 · IDH1/2 · RUNX1 · ASXL1 · KIT
MDSIPSS-M riskSF3B1 · SRSF2 · TP53 · ASXL1 · del(5q)
MPNDriver classificationJAK2 · CALR · MPL · ASXL1 · EZH2
ALLPh+ / Ph-likeBCR-ABL1 · IKZF1 · fusion panel
CLLTP53 / del(17p) handlingTP53 · NOTCH1 · IGHV status

Validation

The prompt is product code.

HemeCopilot's interpretation prompt is versioned in git and changed through review. An expert-validated eval set runs on every change — tier accuracy, risk accuracy, therapy recall, germline flags and uncertainty quality. A regression over the threshold blocks the merge.

Every clinical claim carries a citation or a named guideline
Conflicting evidence is surfaced, never silently resolved
"I don't know" is a valid output — it becomes an uncertainty note
evals / heme_copilot_v1PASS · 0.86
Tier classification0.91
ELN / IPSS-M risk0.85
Therapy signal recall0.82
Germline flag accuracy0.95
20 cases · AML·MDS·MPN·CLL·ALLthreshold ≥0.80 ✓

HemeCopilot · design-partner program

Validate it on your lab's real cases.

We're partnering with molecular hemato-oncology pathologists to pressure-test the drafting engine against real sign-out workflows. Bring your panel and your edge cases.

← Back to platform

AI-drafted, pathologist-reviewed · Not a clinical product · Not for diagnostic use