BURExtract-Llama: An LLM for Clinical Concept Extraction in Breast Ultrasound Reports
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866917754808827904 |
|---|---|
| author | Chen, Yuxuan Yang, Haoyan Pan, Hengkai Siddiqui, Fardeen Verdone, Antonio Zhang, Qingyang Chopra, Sumit Zhao, Chen Shen, Yiqiu |
| author_facet | Chen, Yuxuan Yang, Haoyan Pan, Hengkai Siddiqui, Fardeen Verdone, Antonio Zhang, Qingyang Chopra, Sumit Zhao, Chen Shen, Yiqiu |
| contents | Breast ultrasound is essential for detecting and diagnosing abnormalities, with radiology reports summarizing key findings like lesion characteristics and malignancy assessments. Extracting this critical information is challenging due to the unstructured nature of these reports, with varied linguistic styles and inconsistent formatting. While proprietary LLMs like GPT-4 are effective, they are costly and raise privacy concerns when handling protected health information. This study presents a pipeline for developing an in-house LLM to extract clinical information from radiology reports. We first use GPT-4 to create a small labeled dataset, then fine-tune a Llama3-8B model on it. Evaluated on clinician-annotated reports, our model achieves an average F1 score of 84.6%, which is on par with GPT-4. Our findings demonstrate the feasibility of developing an in-house LLM that not only matches GPT-4's performance but also offers cost reductions and enhanced data privacy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_11334 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | BURExtract-Llama: An LLM for Clinical Concept Extraction in Breast Ultrasound Reports Chen, Yuxuan Yang, Haoyan Pan, Hengkai Siddiqui, Fardeen Verdone, Antonio Zhang, Qingyang Chopra, Sumit Zhao, Chen Shen, Yiqiu Computation and Language Artificial Intelligence Breast ultrasound is essential for detecting and diagnosing abnormalities, with radiology reports summarizing key findings like lesion characteristics and malignancy assessments. Extracting this critical information is challenging due to the unstructured nature of these reports, with varied linguistic styles and inconsistent formatting. While proprietary LLMs like GPT-4 are effective, they are costly and raise privacy concerns when handling protected health information. This study presents a pipeline for developing an in-house LLM to extract clinical information from radiology reports. We first use GPT-4 to create a small labeled dataset, then fine-tune a Llama3-8B model on it. Evaluated on clinician-annotated reports, our model achieves an average F1 score of 84.6%, which is on par with GPT-4. Our findings demonstrate the feasibility of developing an in-house LLM that not only matches GPT-4's performance but also offers cost reductions and enhanced data privacy. |
| title | BURExtract-Llama: An LLM for Clinical Concept Extraction in Breast Ultrasound Reports |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2408.11334 |