A three-pass LLM framework (extractor, detector, false-positive verifier) more than doubled PPV and halved estimated review costs for radiology report error detection, while the absolute number of confirmed errors stayed stable.
Large-scale validation of the feasibility of GPT-4 as a proofreading tool for head CT reports
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
A Multi-Pass Large Language Model Framework for Precise and Efficient Radiology Report Error Detection
A three-pass LLM framework (extractor, detector, false-positive verifier) more than doubled PPV and halved estimated review costs for radiology report error detection, while the absolute number of confirmed errors stayed stable.