AI-Guided Additive Scoring Model for Differential Diagnosis of Primary Liver Cancer
Blood vessel fragments combined with routine blood tests show promise for distinguishing two types of liver cancer, though larger studies are needed before clinical use.
This pilot proof-of-concept study combined small extracellular vesicle (EV) phenotyping with routine serum markers to differentiate HCC from iCCA in LR-M lesions, achieving internal AUROCs up to 0.96 via ML-guided modeling and a simplified 5-point paper-and-pencil PRISM score. Results are internally validated only on a small cohort (n=50) and require multicenter external validation before clinical use.
What the study was
- Study design
- Pilot proof-of-concept (small cohort, internal validation, 80:20 train-test split)
- Population
- LR-M liver lesion patients (n=50; 25 HCC, 25 iCCA)
- Sample size
- 50
- Category
- Diagnostics
- Maturity
- Exploratory
- Journal
- JHEP Reports
Why it surfaced
Novel EV-based liquid biopsy approach to resolve diagnostically ambiguous LR-M liver lesions; translational PRISM score design is clever; however, n=50 cohort with internal-only validation caps evidence maturity at Exploratory.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.