Sample · Data card

Medical data-card field sample

The table is a field outline; below it is a fully fictional synthetic structure example (not stock, not real patients). Real samples are provided under agreed license only. Public inquiry forms do not accept patient data.

Group	Fields (examples)	Buyer use
Identity	dataset_id / version / freeze_date	Version freeze & change traceability
Task	task_id / modality / anatomy	Align training goal & sample scope
Spec	positive_criteria / windowing / guideline_ver	Checkable acceptance criteria
Scale	n_patient / n_study / n_series (separate)	Avoid mixing cases / frames / slides
De-ID	deid_method / residual_risk_note	Compliance & hospital review packs
License	license_scope / retrain / redistribute	Retrain & redistribution boundaries
QC	qc_threshold / disagreement_rate_den	Contract QC thresholds & disagreement denominator
Final review	reviewer_tier / audit_sample_rate	Auditable physician final review (controlled)
Engineering	pack_format / manifest_hash / reader_note	Ingest & read acceptance
Procurement	trial_scope / price_drivers / rework_rule	Trial delivery & rework boundaries

Request data card Gate report skeleton Medical Expert Data

Contracting entity: Beijing Langhui Technology Co., Ltd. (brand LumeSage). Not a medical-device promotion; no clinical efficacy claims.

What makes a deliverable verifiable

Agree the content, access level and timing of these materials before delivery.

Task and counting: Define modality, inclusion rules, deduplication keys and units. Cases, studies, series, slices and annotations are distinct.
Source and permissions: Check provenance, lawful processing basis, training use, retention and redistribution. De-identification does not replace authorization.
Samples and distributions: Review samples, fields, strata and missingness. Sample approval is not batch acceptance.
Expertise and quality: Specify expert fit, review, sample denominators, disagreement resolution, rework and acceptance records.
Files and versions: Agree paths, checksums, versions, format checks and image–text pairing. Settle against accepted, deduplicated units.
Evaluation separation: Freeze tasks and rubrics; agree access, contamination checks, retest conditions and role disclosure.