| Title |
Transformer-based Pathology Report Analysis Algorithm for Automated Clinical Data Information Extraction |
| Authors |
한주원(Ju-Won Han) ; 최지우(Jee-Woo Choi) ; 박승(Seung Park) |
| DOI |
https://doi.org/10.5573/ieie.2022.59.3.44 |
| Keywords |
Artificial intelligence; medical data analysis; electronic medical record; natural language processing; biomedical information extraction |
| Abstract |
Artificial intelligence(AI) technology in the medical field has led to efficient allocation of medical resources and cost reduction. Especially, AI techniques for prognosis prediction of cancer patients may improve the quality of life of patients by providing early prediction the cancer recurrence. However, since the training data used for cancer recurrence and prognosis artificial intelligence are extracted from free-form texts, which are difficult to extract high-quality data due to the high cost and time-consuming nature of the data labeling process. To alleviate this problem, this paper presents an automated clinical data extraction algorithm to build medical AI datasets. Experimental results performed on the thyroid cancer recurrence prediction dataset show that the proposed algorithm extracts the clinical factors with high accuracy. |