Noninvasive early detection and grading of pneumoconiosis via plasma proteomics and machine learning: PRSS3 as a potential biomarker
Blood protein patterns may detect early coal worker lung disease before symptoms appear, though findings require confirmation in independent populations.
Plasma proteomic profiling (1,239 proteins) with ML identified PRSS3 as a potential single-biomarker with AUC=1.00 in training for detecting early-stage coal workers' pneumoconiosis versus dust-exposed controls; dysregulated lipid metabolism and inflammatory pathways implicate a metabolic-immune axis. AUC=1.00 in training requires independent external validation before clinical inference.
What the study was
- Study design
- Plasma proteomic profiling with ML model development and validation
- Population
- 158 participants: 28 healthy controls, 30 dust-exposed workers, 100 CWP patients across 3 stages
- Sample size
- 158
- Category
- Diagnostics
- Maturity
- Exploratory
- Journal
- Clin Proteomics
Why it surfaced
Novel plasma proteomics + ML approach for noninvasive occupational lung disease grading. PRSS3 as single biomarker is intriguing but AUC=1.00 in training set raises overfitting concern; small N=158 limits generalizability. Relevant as methodology for AI-assisted blood-based diagnostics.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.