NIST shows cross-platform TBMNK standardisation across 42 cytometers and 21 sites

The NIST Flow Cytometry Standards Consortium coordinated a TBMNK interlaboratory study using common samples, control and assay reagents, and four SOPs across 42 cytometers from 7 manufacturers at 21 sites.

Results were reported as cells per microlitre for counts and Equivalent Reference Fluorophore units for marker expression, providing a common quantitative scale across platforms.

CFCM explainer — standardisation before AI

Common samples + SOPs → instrument calibration → quantitative units → harmonised analysis → reference dataset for AI/ML work.

Source/credit: CFCM explainer based on NIST and the open-access Frontiers in Immunology publication, published 28 September 2026.

The study supports the idea that AI transferability starts with measurement comparability rather than model architecture alone. It does not show that any particular clinical AI model is externally validated. The authors identify automated sample preparation as a next step for reducing remaining assay uncertainty.

AI Evidence Tracker: instrument harmonisation — strong; fluorescence calibration — strong, ERF-based; metadata completeness — substantial study/SOP metadata; external-site validation — strong for the assay standardisation framework, not for a diagnostic AI model; domain-shift evidence — directly addresses multi-platform variation; OOD/failure handling — not a focus of this study.