AI-assisted Knowledge Discovery in Biomedical Literature to Support Decision-making in Precision Oncology

TL;DR

This study evaluates NLP models, including BERT variants and PubTator, for extracting relevant entities and relations from biomedical literature to aid precision oncology decision-making.

Contribution

It compares the performance of multiple NLP models on NER and RE tasks, identifying BioBERT as the most effective for relation extraction in biomedical literature.

Findings

PubTator 3.0 and BioBERT excelled in NER tasks.

BioBERT achieved the highest F1-score in RE tasks.

Models effectively recognized entities and relations in biomedical texts.

Abstract

The delivery of appropriate targeted therapies to cancer patients requires the complete analysis of the molecular profiling of tumors and the patient's clinical characteristics in the context of existing knowledge and recent findings described in biomedical literature and several other sources. We evaluated the potential contributions of specific natural language processing solutions to support knowledge discovery from biomedical literature. Two models from the Bidirectional Encoder Representations from Transformers (BERT) family, two Large Language Models, and PubTator 3.0 were tested for their ability to support the named entity recognition (NER) and the relation extraction (RE) tasks. PubTator 3.0 and the BioBERT model performed best in the NER task (best F1-score equal to 0.93 and 0.89, respectively), while BioBERT outperformed all other solutions in the RE task (best F1-score 0.79) and a specific use case it was applied to by recognizing nearly all entity mentions and most of the relations.