A Dataset for N-ary Relation Extraction of Drug Combinations

Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 3190 - 3203 July 10-15, 2022 ©2022 Association for Computational Linguistics A Dataset for N-ary Relation Extraction of Drug Combinations Aryeh Tiktinsky∗,1 Vijay Viswanathan∗,2 Danna Niezni3 Dana Meron Azagury3 Yosi Shamay3 Hillel Taub-Tabib1 Tom Hope1,4 Yoav Goldberg1,5 1Allen Institute for AI 2Language Technologies Institute, Carnegie Mellon University 3Faculty of Biomedical Engineering, Technion – Israel Institute of Technology 4School of Computer Science & Engineering, The Hebrew University of Jerusalem 5Computer Science Department, Bar Ilan University aryeht@allenai.org vijayv@andrew.cmu.edu Abstract Combination therapies have become the stan- dard of care for diseases such as cancer, tu- berculosis, malaria and HIV. However, the combinatorial set of available multi-drug treat- ments creates a challenge in identifying effec- tive combination therapies available in a situ- ation. To assist medical professionals in iden- tifying beneficial drug-combinations, we con- struct an expert-annotated dataset for extract- ing information about the efficacy of drug com- binations from the scientific literature. Beyond its practical utility, the dataset also presents a unique NLP challenge, as the first relation ex- traction dataset consisting of variable-length relations. Furthermore, the relations in this dataset predominantly require language under- standing beyond the sentence level, adding to the challenge of this task. We provide a promising baseline model and identify clear ar- eas for further improvement. We release our dataset,1 code,2 and baseline models3 publicly to encourage the NLP community to partici- pate in this task. 1 Introduction “So far, many monotherapies have been tested, but have been shown to have limited efficacy against COVID-19. By contrast, combinational therapies are emerging as a useful tool to treat SARS-CoV-2 infection.” (Ianevski et al., 2021). Indeed, combining two or more drugs together has proven to be useful for treatments of various medical conditions, including cancer (DeVita et al., 1975; Carew et al., 2008; Shuhendler et al., 2010), AIDS (Bartlett et al., 2006), malaria (Eastman and * Equal contribution. 1https://huggingface.co/datasets/ allenai/drug-combo-extraction 2https://github.com/allenai/ drug-combo-extraction 3https://huggingface.co/allenai/ drug-combo-classifier-pubmedbert-dapt Fidock, 2009), tuberculosis (Bhusal et al., 2005), hypertension (Rochlani et al., 2017) and COVID- 19 (Ianevski et al., 2020). In this work, we examine the clinically signifi- cant and challenging NLP task of extracting known drug combinations from the scientific literature. We present an expert-annotated dataset and base- line models for this new task. Our dataset contains 1600 manually annotated abstracts, each mention- ing between 2 and 15 drugs. 840 of these abstracts describe one or more positive drug combinations, varying in size from 2 to 11 drugs. The remaining 760 abstracts either contain mentions of drugs not used in combination, or discuss combinations of drugs that do not give a combined positive effect. For the clinical setting, solving the drug com- bination identification task can help researchers suggest and validate complex treatment plans. For example, when searching for effective treatments for cancer, knowing which drugs interact synergisti- cally with a first line treatment allows researchers to suggest new treatment plans that can subsequently be validated …

Content truncated to 500 words