A locally deployable agentic framework for clinical data deidentification
Year: 2026
Type: preprint
Abstract: Multimodal clinical data contain identifiers across diverse formats. We developed the Multimodal Anonymizer, a locally deployable framework combining multimodal language models, specialist networks, deterministic transformations and iterative verification. We compared 16 orchestrator configurations using automated evaluation on 250 constructed MIMIC-IV-derived patient bundles with synthetic identifiers. A selected local configuration underwent manual evaluation on these same bundles supplemented with head CT scans, face images, handwritten notes, audio recordings and synthetic German clinical notes from unrelated sources. Separate evaluations used the German GRASCCO corpus and 250 Charité partograms. On this expanded benchmark, patient-level deidentification sensitivity was 98.80% (95% CI, 97.20–100), per-identifier sensitivity 99.82% (95% CI, 99.76–99.88), patient-level critical clinical preservation 99.60% (95% CI, 98.80–100) and the clinical preservation score 99.61% (95% CI, 99.51–99.71). All partograms met the study-defined deidentification criterion. These findings support feasibility, while constructed bundles, synthetic identifiers, limited natural-data validation and absent component ablations constrain generalizability.
Source: medRxiv
Institutions: HTW Berlin - University of Applied Sciences, Humboldt-Universität zu Berlin, Berlin Institute for the Foundations of Learning and Data, Technische Universität Berlin, Charité – Universitätsmedizin Berlin, Berlin Institute of Health at Charité - Universitätsmedizin Berlin, ETH Zurich, Institute for Machine Learning, German Centre for Cardiovascular Research, Deutsches Herzzentrum der Charité
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7163660880