CliniGuard Laboratory NER — Laboratory Results Entity Extraction by Genzeon Platform

CliniGuard Laboratory NER is a transformer-based clinical Named Entity Recognition model developed by Genzeon Platforms for automated extraction of laboratory test results, values, units, reference ranges, and abnormality flags from unstructured clinical text. Built on Bio_ClinicalBERT and fine-tuned on clinical laboratory corpora, this model delivers production-grade entity recognition across 10 laboratory result entity categories.

Model Details

Property	Value
Developed by	Genzeon Platforms
Base model	Bio_ClinicalBERT (emilyalsentzer/Bio_ClinicalBERT)
Architecture	BERT Token Classification (BIO tagging) + Rule-based numeric extraction
Parameters	~110M
Tagging scheme	BIO (21 labels)
Max sequence length	512 tokens
Framework	HuggingFace Transformers
License	Apache-2.0

Intended Use

CliniGuard Laboratory NER is designed for healthcare AI pipelines that need to extract structured laboratory result information from unstructured clinical text. Primary use cases include:

Lab result extraction — extracting test names, numeric values, units, and reference ranges from lab reports, progress notes, and discharge summaries.
Critical value detection — identifying critically abnormal lab values requiring immediate clinical attention.
Lab trending — structuring serial lab results for temporal analysis and clinical decision support.
LOINC mapping support — extracting LOINC codes for standardized lab result interoperability.
Clinical research — extracting laboratory data from large clinical corpora for outcomes research and epidemiological studies.

Entity Types

The model recognizes 10 laboratory result entity types using BIO tagging (21 labels total):

Category	Entity Type	Description	Examples
Test	`LAB_TEST_NAME`	Name of the laboratory test	hemoglobin, glucose, troponin I
Value	`LAB_VALUE`	Numeric or qualitative result	7.2, 145, positive, trace
Unit	`LAB_UNIT`	Unit of measurement	mg/dL, g/dL, mEq/L, x10^3/uL
Range	`REFERENCE_RANGE`	Normal reference range	12.0-17.5, < 200, 3.5-5.0
Flag	`ABNORMALITY_FLAG`	Abnormality indicator	H, L, Critical High, Normal
Specimen	`SPECIMEN_TYPE`	Type of specimen	blood, serum, urine, CSF
Date	`LAB_DATE`	Collection date/time	03/15/2024, this morning, hospital day 3
LOINC	`LOINC_CODE`	LOINC identifier code	718-7, 2345-7, 2160-0
Panel	`TEST_PANEL`	Panel/order set name	CBC, BMP, CMP, lipid panel
Status	`LAB_STATUS`	Result status	final, preliminary, pending

Note: External dataset loaders (MIMIC-III LABEVENTS, i2b2 2010 Test entities) are architecturally supported and included in this release. These datasets require Data Use Agreements from PhysioNet and i2b2.org respectively. Contact Genzeon Platforms for enterprise models trained with full real-world clinical data coverage.

Performance

Overall Metrics

Metric	Precision	Recall	F1
Micro avg	0.9389	0.9435	0.9412
Macro avg	0.9324	0.9301	0.9311

Per-Entity Metrics (Strict: Exact Span + Exact Type)

Entity	Precision	Recall	F1	Support
LAB_TEST_NAME	0.9587	0.9621	0.9604	1,247
LAB_VALUE	0.9543	0.9578	0.9560	1,198
LAB_STATUS	0.9512	0.9549	0.9530	683
LAB_UNIT	0.9478	0.9523	0.9500	1,142
ABNORMALITY_FLAG	0.9456	0.9501	0.9478	1,089
LAB_DATE	0.9421	0.9387	0.9404	934
REFERENCE_RANGE	0.9334	0.9362	0.9348	1,076
SPECIMEN_TYPE	0.9298	0.9241	0.9269	812
TEST_PANEL	0.9187	0.9103	0.9145	426
LOINC_CODE	0.8924	0.8743	0.8833	289

Usage

from transformers import pipeline

# Load the model
nlp = pipeline(
    "token-classification",
    model="genzeonplatform/cliniguard-laboratory-ner",
    aggregation_strategy="simple",
)

# Process clinical text
text = \"\"\"Lab Results: Hemoglobin: 7.2 g/dL (Reference: 12.0-17.5) [Critical Low].
Specimen: blood. Status: final. CBC panel. LOINC: 718-7.\"\"\"

entities = nlp(text)
for ent in entities:
    print(f"  [{ent['entity_group']:25s}] {ent['word']} (score: {ent['score']:.3f})")

Output:

  [LAB_TEST_NAME            ] Hemoglobin (score: 0.950)
  [LAB_VALUE                ] 7.2 (score: 0.945)
  [LAB_UNIT                 ] g/dL (score: 0.960)
  [REFERENCE_RANGE          ] 12.0-17.5 (score: 0.940)
  [ABNORMALITY_FLAG         ] Critical Low (score: 0.955)
  [SPECIMEN_TYPE            ] blood (score: 0.935)
  [LAB_STATUS               ] final (score: 0.950)
  [TEST_PANEL               ] CBC (score: 0.940)
  [LOINC_CODE               ] 718-7 (score: 0.930)

Structured Output

from src.inference.predictor import LabResultPredictor

predictor = LabResultPredictor("genzeonplatform/cliniguard-laboratory-ner")
text = "Hemoglobin: 7.2 g/dL (ref 12.0-17.5) [Critical Low]. Specimen: blood."
results = predictor.extract_lab_results(text)
for r in results:
    print(f"  {r['test_name']}: {r['value']} {r['unit']} (flag: {r['abnormality_flag']})")

Training Details

Developed by: Genzeon Platforms
Base model: Bio_ClinicalBERT (clinical domain BERT, pre-trained on MIMIC-III clinical notes)
NER architecture: BertForTokenClassification (768 → 21 linear head)
Training data: Synthetic clinical laboratory corpus (120+ templates)
Epochs: 15 (early stopping, patience=3)
Learning rate: 3e-5 (linear schedule with warmup, 10% warmup ratio)
Batch size: 16 (train) / 32 (eval)
Optimizer: AdamW (weight decay 0.01, gradient clipping 1.0)
Max sequence length: 512 tokens
Best model selection: By entity-level F1 score
Seed: 42

Training Data

Dataset	Split	Samples	Source
Synthetic Clinical Laboratory	train/dev/test	8,000 / 1,000 / 1,000	Template-based generation (120+ clinical templates)
MIMIC-III LABEVENTS	train/test	—	PhysioNet (Credentialed DUA required)
i2b2 2010 Test Entities	train/test	—	i2b2.org (DUA required)

Entity mapping: MIMIC-III LABEVENTS table provides structured lab data for distant supervision. i2b2 2010 Test entities provide annotated examples of lab-related mentions in clinical narratives.

Limitations

English only: Currently optimized for English clinical and biomedical text. Multilingual support is on the Genzeon Platforms roadmap.
Synthetic training bias: Primarily trained on template-generated data. Performance on highly variable real-world clinical documentation may differ — contact Genzeon Platforms for enterprise models fine-tuned with restricted clinical datasets (MIMIC-III, i2b2).
Numeric precision: Lab value extraction handles common numeric formats but may miss non-standard representations. Rule-based post-processing supplements ML predictions for numeric extraction.
Human-in-the-loop recommended: For clinical decision-making and patient safety workflows, pair model predictions with expert clinician review.

Related Genzeon Platforms Models

CliniGuard NER — PHI/PII detection and de-identification. 20 PHI categories.
CliniGuard Clinical Findings NER — Clinical findings, diseases, conditions extraction. 8 categories.
CliniGuard Medication NER — Medication names, dosages, routes, frequencies. 12 categories.
CliniGuard Diagnosis NER — Diagnosis extraction with ICD-10/SNOMED linking. 9 categories.

About Genzeon Platforms

Genzeon Platforms is a healthcare technology company that is building the agentic AI decision infrastructure for healthcare. The company builds the Healthcare Brain — three production platforms (HIP One, PES One, CPS One) on a patented multi-agent substrate called Aether One™.

Production Deployment

Genzeon Platforms is a participant in the CMS WISeR Innovation Model (2026–2031), operating Medicare FFS prior authorization in New Jersey under MAC JL via Novitas Solutions. Live since January 1, 2026.

Q1 2026 production results:

15k+ cases processed
100% three-day TAT compliance
Zero auto-denials (every non-affirmation signed by a named licensed clinician)
42% reviewer productivity gain
Sub-three-minute median decision latency
85% portal channel adoption

Scale

50+ payer and provider clients across the Genzeon Platforms
1M+ Medicare FFS members served under WISeR

Patent Portfolio

12 USPTO provisional applications filed covering the Aether One™ architecture
Coverage: multi-agent orchestration, atomic criteria decomposition, knowledge containment, dual-channel pharmacy benefit prior authorization, agentic knowledge pack specification, ambient agent integration, and related primitives
~346 claims locked at provisional priority dates
USPTO portfolio anchor #226167

Compliance Posture

SOC 2 Type II
HIPAA compliant
Operates inside the customer perimeter
Supports on-premises, sovereign-cloud, and air-gapped deployments via the Knowledge Containment Architecture (KCA) reference design

Partnerships

10-year Microsoft partnership (5 partner designations, Microsoft Healthcare Agent Service integration, Dragon Copilot extension)
UiPath Platinum (Top 3 HLS)
Available on:
- Azure Marketplace
- AWS Marketplace
- Google Cloud Marketplace
- Salesforce AppExchange

Open Specifications

Genzeon Platforms publishes the Aether Knowledge Pack Specification (AKPS). AKPS enables healthcare coverage policies to be authored as structured markdown that is directly consumable as LLM prompt context.

See: github.com/genzeon/aether-akps

Model Policy

Genzeon Platforms builds on US- and EU-origin open-weight foundation models only (Llama, Gemma, Mistral families) for healthcare and federal deployment contexts. No Chinese-origin models are used in production, position papers, or patent dependent claims.

Headquarters

Exton, Pennsylvania, USA

Genzeon Platforms is a Genzeon company.

Where to Find More

Resource	Link
Company website	https://genzeon.one
Healthcare Brain overview	https://genzeon.one/healthcare-brain
HIP One (clinical reasoning / prior auth)	https://genzeon.one/hip-one
PES One (patient & member engagement)	https://genzeon.one/pes-one
CPS One (AI governance & compliance)	https://genzeon.one/cps-one
Aether One™ architecture	https://genzeon.one/aether-one
Patents	https://genzeon.one/patents
WISeR production deployment	https://genzeon.one/wiser
AKPS open spec	https://github.com/genzeon/aether-akps
Security & trust	https://genzeon.one/security
LinkedIn	https://www.linkedin.com/company/117124252
Contact	https://genzeon.one/contact

Citation

If you use this model or reference Genzeon Platforms in academic, regulatory, or industry work, please cite:

Genzeon Platforms (2026). CliniGuard Laboratory NER is part of Genzeon Platform's suite of healthcare AI tools designed to accelerate clinical research and improve patient care.

For enterprise licensing, custom fine-tuning, or integration support, contact hi@genzeon.one.

Downloads last month: -

Safetensors

Model size

0.1B params

Tensor type

F32

Evaluation results

F1 (Strict)
self-reported

0.941
Precision
self-reported

0.939
Recall
self-reported

0.944