Target initial classification benchmark
6 hours → 3.5 minutes
Target initial cell classification across a 1,000-cell review packet.
The app pre-classifies. The pathologist reviews, confirms, and finalizes.
AI-assisted hematology review • morphology + Flow
From raw slides and FCS files to review-ready evidence in minutes.
NanoCure AI pre-classifies cells, maps every call to its source slide, surfaces candidate Flow populations, and packages the evidence for pathologist confirmation—without requiring new lab hardware.
MGG human results
Sysmex human results
Raw FCS list-mode review
Windows + Linux desktop bridge
Every cell mapped to its source
Candidate populations stay linked across plots
Static views from the live product • no concept interface
Performance evidence
Speed
100×
Target faster initial classification.
356.5 minutes returned per 1,000-cell review packet at the target operating bar.
Pathologist comparison
≤10%
Blast-% difference versus pathologist review.
Agreement claim—not generalized diagnostic accuracy.
Human smear workflows
MGG + Sysmex
Strong human results across two staining workflows.
Cell-level source traceability remains intact across both.
Internal workflow target uses an approximately 2,500-cell review packet versus manual initial classification. Source slides/images remain linked, but throughput is measured by reviewed cells. Pathologist confirmation time is separate; validation and scaling are ongoing.
Human workflow evidence
Built to perform across real stain variation—not only a single idealized image set.
NanoCure AI has demonstrated strong human smear workflow results with MGG and Sysmex staining while preserving the same case totals, per-cell evidence, full-slide mapping, and pathologist review controls.
MGG
Human bone marrow and peripheral blood smear workflow evaluation with cell-level source mapping.Sysmex
Human Wright / Wright-Giemsa stain workflow evaluation through the same reviewer-controlled product path.Case-levelBlast, healthy, accepted-cell, and blast-percentage totals across all submitted images.Designed to move the reviewer from raw images to a prioritized case summary.
Cell-levelCrop, predicted call, probability, source file, cell ID, and review status remain visible together.Every output remains inspectable rather than becoming a detached score.
Source-linkedReviewers move bidirectionally between an identified cell and its original slide coordinates.Full-slide mapping preserves where the evidence came from.
Reviewer-controlledAccept, reject wrong call, reject unusable image, correct, annotate, and export the review record.The pathologist remains the final reviewer.
Morphology workspace
Start with a prioritized case—not a blank slide.
The morphology workflow turns a multi-image case into case totals, per-cell calls, source-slide maps, and preserved reviewer decisions before the pathologist begins final interpretation.
Live morphology interface
Real product viewFull-slide mapping • source image list • category filters • selected-cell evidence
Multi-image case ingestion
Submit multiple stained smear images under a de-identified case and preserve image-level and case-level context.
Case totals and blast percentage
See accepted cells, blast-like cells, healthy cells, multiple-cell calls, and the resulting case-level blast percentage.
Per-cell evidence
Inspect the crop, classification, probability, source image, cell ID, and review status in one row.
Full-slide cell mapping
Every mapped cell remains linked to its original coordinates with blast, healthy, multiple-cell, reviewed, and low-confidence filters.
Pathologist adjudication
Accept results, reject an incorrect call, reject an unusable image, clear a decision, and preserve the reviewer record.
Export the evidence—not only the prediction.Generate CSV, visual/PDF, and full-artifact outputs while retaining source links and reviewer actions.
Flow Cytometry FCS workspace
Interrogate candidate populations across every plot—without losing event context.
The Flow workflow goes well beyond static plots. It combines raw FCS ingestion, linked event selection, population scaling, candidate clusters, marker adequacy, QC, and evidence export in one reviewer-controlled workspace.
Raw list-mode FCS ingestion
Review raw FCS 2.0, 3.0, and 3.1 files with case, specimen, disease context, requested panel, profile, strategy, reference, and compensation context.
Linked brushing across all plots
Select events with rectangle, circle, or lasso tools and keep the same events highlighted across every synchronized diagram.
Population scaling and focus
Switch between whole-file and selected-event views, then use fit, 2×, 4×, or 8× scaling without breaking event linkage.
Candidate clusters and event masks
Inspect color-coded clusters, candidate burden, event counts, selected-population notes, and exportable event masks.
Marker adequacy and QC
Surface present and missing markers, panel readiness, missing-marker impact, acquisition/time anomalies, compensation context, and sample-quality flags.
Review packet and evidence output
Preserve confirm, deny, or correct decisions alongside dot plots, histograms, boxplots, multiparameter maps, evidence reports, and JSON.
Live Flow Cytometry interface
Real product viewLinked brushing • rectangle / circle / lasso • synchronized event context
Interactive reviewSelections remain linked across FSC/SSC, CD45/SSC, and disease-oriented marker-pair views.
Assay readinessMarker adequacy and missing-channel impact are presented before a reviewer over-interprets the file.
Traceable outputEvidence report, JSON, event mask, reviewer notes, and decision state remain connected to the case.
One controlled review path
AI does the first pass. The pathologist makes the final call.
Case first. Modality second. Reviewer always in control.
01 / UPLOAD
Upload
Submit smear images or raw FCS files under a de-identified Case ID through the desktop bridge or browser workflow.
02 / ANALYZE
Analyze
Run morphology inference or Flow analysis while retaining input, assay, and case context.
03 / ORGANIZE
Organize
Generate case totals, per-cell rows, full-slide maps, linked plots, candidate populations, QC, and marker context.
04 / CONFIRM
Confirm
Accept, reject, correct, annotate, and inspect the linked source evidence before final interpretation.
05 / EXPORT
Export
Create CSV, JSON, visual/PDF, Flow reports, event masks, and full evidence artifacts for QA or handoff.
Technical readiness
Built for the laboratory workflow you already have.
NanoCure AI adds a digital review layer around existing image and Flow workflows instead of forcing a new instrument purchase.
Desktop + browser workflow
Windows and Linux desktop ingestion connects to the same browser-based morphology and Flow review experience.
Existing lab inputs
Use current microscope/camera images, scanner files, and raw list-mode FCS output without replacing core laboratory hardware.
Structured evidence outputs
CSV, JSON, PDF/visual evidence, full artifacts, Flow reports, and event masks support QA, collaboration, and handoff.
Reviewer-controlled record
De-identified case IDs, source links, reviewer actions, notes, and preserved decisions keep the workflow inspectable.
Pathologists
More time for interpretation.
Start with a prioritized case, focus on discordant or difficult cells, move directly between crops and source images, and preserve decisions for QA.
Expert attention goes to confirmation and difficult cases.Hospital + lab leaders
More capacity from the same specialist team.
Shorten repetitive first-pass work, reduce queue pressure, standardize evidence, and measure cases per expert hour and turnaround.
A scalable operating model without a hardware swap.Hematology operations
Better coverage across shifts and sites.
Route de-identified digital cases to available reviewers, escalate suspicious findings with source evidence, and maintain a consistent review record.
A shared digital review layer across the network.60-day pilot program
Prove the operational value in your own workflow.
Run NanoCure AI alongside your current process using de-identified cases. Measure the first-pass time returned, pathologist touch time, agreement, turnaround, reviewer acceptance, and evidence-output completeness before making a broader deployment decision.
Processing timeMeasure app first-pass time by case and batch.
Pathologist touch timeSeparate confirmation time from pipeline processing.
AgreementCompare blast percentage and reviewer decisions.
TurnaroundMeasure time from input readiness to review-ready evidence.
Cases per expert hourQuantify capacity returned to the specialist team.
Reviewer acceptanceTrack usability and retained decisions.
Export completenessVerify evidence packets, reports, and source links.
Deployment fitDefine integration and next-step requirements.