01 / Magnifications
4×–100×
4× · 10× · 20× · 40× · 100×
Five supported magnifications for stained leukemia patient images.
Leukemia morphology + breast histology
AI review across
bone marrow, blood, and tissue.
Two focused models for leukemia morphology and breast histology. Bring the images your laboratory already captures. Inspect source-linked evidence. Keep the specialist in control.
Leukemia / Image flexibility
4×·10×·20×·40×·100×
MGG Sysmex Leishman
Microscope-camera images. Even cell-phone captures.
Real-world image quality. Not just ideal captures.
Explore image support and qualityReview existing digital images in the browser. Research and workflow evaluation only.
Leukemia morphologyCELL REVIEW
Stained smearsSource → cell → review
Breast histologyREGION REVIEW
H&E tissueSlide → region → review
Leukemia / Built for real-world capture
Your microscope.
Your staining workflow.
Not one magnification. Not one stain. Not one camera. Bring leukemia images from the setup you already use—without a dedicated NanoCure scanner.
Explore the leukemia model02 / Stain workflows
MGG. Sysmex.
Leishman.
Work with these staining protocols across peripheral blood and bone marrow smears.
03 / Capture devices
Even a
cell phone.
Use a microscope camera or capture the microscope field with a cell phone. No proprietary NanoCure capture hardware required.
04 / Image quality
Real-world
image quality.
Not limited to pristine captures. Submit leukemia images with variation in focus, lighting, and resolution.
Broad input support. Expert judgment. These are leukemia input capabilities, not a guarantee of equal performance at every magnification or image quality. Specialists review the visible evidence and can exclude unusable images.
Two models / One review platform
Different evidence.
Built for the right expert.
Leukemia morphology evaluates individual cells in stained smear images. Breast histology ranks areas of H&E tissue for pathologist attention. Each workflow has its own evidence and limits.
01 / HEMATOLOGYAML lead human evaluation
Leukemia morphology
Identify and review suspected malignant cells in stained bone marrow and peripheral blood images. Inspect each proposed cell call in its source image and record the expert decision.
Explore leukemia review02 / PATHOLOGYLimited current validation
Breast histology
H&E breast tissue images and digital slides. Move from a ranked worklist to review-priority heatmaps, tissue regions, tile evidence, and structured pathologist feedback.
Explore histology review01 / Leukemia morphology model
A leukemia foundation.
Built for real laboratories.
Our leukemia program supports the identification and review of suspected malignant cells, with AML → ALL → CLL as its disease-specific development sequence. Public training data and authorized expert-labeled samples support continued evaluation.
Developed in partnership with
Chandrasekharan LabsUniversity of UtahHuntsman Cancer Institute
01
Learn from the cell and its context.
The AML model learns from stained cell images to propose blast / non-blast calls. The review workspace keeps each crop, probability, and source location together so a specialist can inspect the result.
02
Build depth through partner data.
Our data strategy combines public morphology datasets with proprietary, expert-labeled samples shared under research agreements. Regional cohorts and stain variation guide further training and evaluation.
03
Earn each disease expansion.
AML is the lead human-evaluated application. ALL and CLL require their own labeled cohorts, review criteria, and performance evaluation as we extend the model.
Model developmentAML-led foundation
Public morphology dataCell diversity and baseline training
Authorized partner dataExpert labels and human stain variation
NanoCure AI / leukemia program
Leukemia morphology model
Cell-level learning with an AML starting point and disease-specific expansion toward ALL and CLL.
Evaluation against expert reference callsBlast sensitivity · non-blast specificity · case agreement · review time
Development framework. Each new disease requires separate evidence before broader use.
Routine inference is separate from model development. Any use of partner data for training requires a separate agreement and authorization.
01 / Cell-level review workflow
From the case overview
to the cell that needs you.
Bring the image, proposed call, and review decision into one workspace. Move between individual cells and their original image context as you work through the case.
Open an organized case.
Submit stained smear images—including cell-phone microscope captures—under a de-identified case ID. Review detected cells and proposed classifications alongside their source images.
Inspect the evidence.
See the cell crop, predicted call, probability, and original image location. Use category and review-status filters to focus your attention.
Apply your judgment.
Accept a proposed call, reject an incorrect classification, or mark an image unusable. Keep the reviewer’s decision attached to the cell.
Carry the record forward.
Export the reviewer workbook, CSV, visual/PDF evidence, and source-linked artifacts for case discussion, quality review, and handoff.
ACTUAL PRODUCT
Every cell retains its source.
Move between the cell crop and original image coordinates without losing case context.
The reviewer’s decision stays attached.
Review the proposed classification, flag unusable images, and preserve the record.
REVIEW ACTIONSAccept, reject, or exclude.
Distinguish a wrong classification from an image that cannot support interpretation.
EVIDENCE EXPORTTake the case evidence with you.
CSV, visual/PDF output, and full artifacts retain the context behind the review.
A proposed cell label is review evidence. A non-blast call does not establish that a cell is benign, and morphology alone does not establish a leukemia diagnosis or subtype.
02 / Breast histology model
From the whole slide
to regions worth reviewing.
Upload de-identified H&E breast tissue images or digital slides. The model organizes ranked slide and region evidence; a qualified pathologist investigates the tissue and records the interpretation.
Breast Histology Review ItineraryACTUAL PRODUCT VIEW
Start with the tissue.
Create a de-identified case and upload H&E tissue images or WSI files, including SVS, TIFF, NDPI, MRXS, PNG, or JPEG.
Find regions to inspect.
Use the ranked slide worklist, review-priority heatmap, spatial region overlay, and tile gallery to focus the review.
Record the pathologist's decision.
Inspect the source tissue and ranked areas, then save a decision, reason, or corrected region boundary as appropriate.
Carry the review forward.
Download the offline Reviewer Workbook and export review records for case discussion, quality checks, and research workflow evaluation.
Review priority is not a diagnosis. A highlighted area does not independently establish malignancy, subtype, grade, stage, or treatment. MIL attention appears only when trained weights are deployed; unavailable evidence is shown as N/A.
01 / Leukemia development roadmap
AML first.
Then ALL. Then CLL.
A focused path for malignant-cell identification and expert review in AML, ALL, and CLL patient images. Each expansion adds disease-specific data, review criteria, and evaluation.
01 / CURRENT FOCUSHuman evaluation
AML
Acute myeloid leukemia
Our lead morphology workflow for blast-focused review of bone marrow and peripheral blood smear images, with source mapping and cell-level adjudication.
Near-term priorityExpand human evaluation and improve non-blast discrimination while protecting blast sensitivity.
02 / NEXT EXPANSIONIn development
ALL
Acute lymphoblastic leukemia
Extend the model to lymphoblast morphology using ALL-specific training data and human evaluation. Establish performance independently from AML.
Evidence neededExpert-labeled ALL cohorts and disease-specific evaluation across specimen and stain variation.
03 / FOLLOWING EXPANSIONPlanned expansion
CLL
Chronic lymphocytic leukemia
Develop support for mature lymphoid morphology and abnormal lymphocyte review, with a disease-specific approach suited to CLL.
Evidence neededCLL-specific cell labels and comparison cohorts that distinguish abnormal from reactive lymphoid morphology.
The sequence describes our development priorities. ALL and CLL are expansion programs; AML results do not establish performance in either disease.
03 / In your laboratory
Your images.
Your existing equipment.
Start with the capture tools already in your laboratory. Leukemia review accepts multiple magnifications, stain workflows, and cell-phone captures; breast histology has its own H&E tissue workflow.
No dedicated NanoCure hardware.
For leukemia, use images from your microscope camera or cell phone. Breast histology uses its separate H&E image and digital-slide workflow.
Existing capture equipment · Browser-based reviewGive each specialist the right view.
Hematopathologists review cell calls in their source image; breast pathologists review ranked tissue regions and spatial evidence. Both workflows run in the browser.
Distinct cell and tissue review workspacesMeasure the workflow on your cases.
Compare leukemia cell calls with disease-specific expert labels; for AML, assess blast and non-blast performance. Evaluate tissue-region prioritization and feedback separately for breast histology.
Separate evaluation criteria for each model04 / Start with expert evaluation
Bring your everyday images.
Let experts decide.
Request access for leukemia morphology, breast histology, or both. For leukemia, bring de-identified images with the magnifications, stains, capture devices, and quality variation your team actually sees.
Choose your model.
Tell us whether you want to evaluate stained leukemia smears, breast H&E tissue images, or both. Agree on case mix and data-use terms.
Review the evidence.
Compare cell calls with expert labels for leukemia; assess ranked tissue regions and pathologist decisions for breast histology.
Check the fit.
Confirm image compatibility, reviewer experience, exports, and practical integration before defining pilot acceptance criteria.
Request free expert access
Useful starting details: specialty, model of interest, magnifications, staining workflows, capture method, file format, and approximate case volume.
Questions from the laboratory
Before you begin.
Practical details for hematopathologists, breast pathologists, and laboratory teams considering an evaluation.
What do the two models currently support?
Leukemia morphology supports expert review of proposed cell calls from stained bone marrow and peripheral blood smear images; AML is the lead human-evaluated workflow. The separate breast histology model prioritizes regions on H&E tissue images for qualified pathologist review and has limited current validation. ALL and CLL remain disease-specific development programs.
What do you mean by a leukemia foundation model?
We are building a shared morphology model that can be extended across leukemia applications using expert-labeled data. AML is the starting point. The term describes our development direction; it does not imply that one model is already validated for every leukemia type or that it establishes a diagnosis.
Which leukemia magnifications and stain workflows are supported?
The leukemia workflow accepts 4×, 10×, 20×, 40×, and 100× images from MGG, Sysmex, and Leishman staining workflows. This input scope applies to stained peripheral blood and bone marrow images. Breast histology is a separate H&E workflow.
Can we use leukemia images taken with a cell phone?
Yes. Leukemia images captured through a microscope using a cell phone can be submitted for review, alongside images from microscope cameras. A proprietary NanoCure microscope or scanner is not required. The source image and proposed cell calls remain available for expert inspection.
Do leukemia images have to be pristine?
No. The leukemia workflow accepts varied-quality images, not only ideal captures. Acceptance does not establish interpretability or equivalent accuracy across magnifications and image qualities. The specialist reviews the visible evidence and can exclude unusable images. Evaluate performance on your own capture conditions and case mix.
Do we need a new microscope or scanner?
No proprietary NanoCure microscope or scanner is required. For leukemia, use existing microscope-camera images or cell-phone captures through the microscope. For breast histology, use H&E tissue images or supported digital-slide files. Confirm file-format compatibility and intended workflow during evaluation.
What does the specialist control?
In leukemia review, the hematopathologist can inspect each cell and source image, accept or reject a proposed call, and identify unusable images. In breast histology, the qualified pathologist inspects ranked tissue regions and records a decision and reason. Specialists remain responsible for interpretation in the full clinical context.
Does a highlighted breast region mean cancer?
No. The breast model marks review priority only. It does not independently establish malignancy, histologic subtype, lineage, grade, stage, or treatment. Unavailable model evidence appears as N/A, not as a negative result.
Are evaluation images used to train the model?
Data use is agreed before an evaluation begins. Routine inference is separate from research and model development. Any contribution to training requires a separate agreement and explicit authorization.
What can we export from the review?
Leukemia morphology supports cell review workbooks, CSV, visual/PDF output, and source-linked artifacts. Breast histology supports an offline Reviewer Workbook with structured review records and export options. We assess case handoff and integration requirements with each laboratory.