NSSC: a neuro-symbolic AI system for enhancing accuracy of named entity recognition and linking from oncologic clinical notes

Research output: Contribution to journalArticleResearchpeer review

Authors

  • Álvaro García-Barragán
  • Ahmad Sakor
  • Maria Esther Vidal
  • Ernestina Menasalvas
  • Juan Cristobal Sanchez Gonzalez
  • Mariano Provencio
  • Víctor Robles

Research Organisations

External Research Organisations

  • Technical University of Madrid (UPM)
  • Scientific Data Management Group
  • German National Library of Science and Technology (TIB)
  • Puerta de Hierro Majadahonda University Hospital
View graph of relations

Details

Original languageEnglish
Number of pages24
JournalMedical and Biological Engineering and Computing
Early online date1 Nov 2024
Publication statusE-pub ahead of print - 1 Nov 2024

Abstract

Abstract: Accurate recognition and linking of oncologic entities in clinical notes is essential for extracting insights across cancer research, patient care, clinical decision-making, and treatment optimization. We present the Neuro-Symbolic System for Cancer (NSSC), a hybrid AI framework that integrates neurosymbolic methods with named entity recognition (NER) and entity linking (EL) to transform unstructured clinical notes into structured terms using medical vocabularies, with the Unified Medical Language System (UMLS) as a case study. NSSC was evaluated on a dataset of clinical notes from breast cancer patients, demonstrating significant improvements in the accuracy of both entity recognition and linking compared to state-of-the-art models. Specifically, NSSC achieved a 33% improvement over BioFalcon and a 58% improvement over scispaCy. By combining large language models (LLMs) with symbolic reasoning, NSSC improves the recognition and interoperability of oncologic entities, enabling seamless integration with existing biomedical knowledge. This approach marks a significant advancement in extracting meaningful information from clinical narratives, offering promising applications in cancer research and personalized patient care. Graphical abstract: (Figure presented.)

Keywords

    Breast cancer, EHR, EL, LLM, NER, Neuro-symbolic

ASJC Scopus subject areas

Sustainable Development Goals

Cite this

NSSC: a neuro-symbolic AI system for enhancing accuracy of named entity recognition and linking from oncologic clinical notes. / García-Barragán, Álvaro; Sakor, Ahmad; Vidal, Maria Esther et al.
In: Medical and Biological Engineering and Computing, 01.11.2024.

Research output: Contribution to journalArticleResearchpeer review

García-Barragán, Á., Sakor, A., Vidal, M. E., Menasalvas, E., Gonzalez, J. C. S., Provencio, M., & Robles, V. (2024). NSSC: a neuro-symbolic AI system for enhancing accuracy of named entity recognition and linking from oncologic clinical notes. Medical and Biological Engineering and Computing. Advance online publication. https://doi.org/10.1007/s11517-024-03227-4
García-Barragán Á, Sakor A, Vidal ME, Menasalvas E, Gonzalez JCS, Provencio M et al. NSSC: a neuro-symbolic AI system for enhancing accuracy of named entity recognition and linking from oncologic clinical notes. Medical and Biological Engineering and Computing. 2024 Nov 1. Epub 2024 Nov 1. doi: 10.1007/s11517-024-03227-4
Download
@article{29d339e613d34f5ea7d813f4aa904f8e,
title = "NSSC: a neuro-symbolic AI system for enhancing accuracy of named entity recognition and linking from oncologic clinical notes",
abstract = "Abstract: Accurate recognition and linking of oncologic entities in clinical notes is essential for extracting insights across cancer research, patient care, clinical decision-making, and treatment optimization. We present the Neuro-Symbolic System for Cancer (NSSC), a hybrid AI framework that integrates neurosymbolic methods with named entity recognition (NER) and entity linking (EL) to transform unstructured clinical notes into structured terms using medical vocabularies, with the Unified Medical Language System (UMLS) as a case study. NSSC was evaluated on a dataset of clinical notes from breast cancer patients, demonstrating significant improvements in the accuracy of both entity recognition and linking compared to state-of-the-art models. Specifically, NSSC achieved a 33% improvement over BioFalcon and a 58% improvement over scispaCy. By combining large language models (LLMs) with symbolic reasoning, NSSC improves the recognition and interoperability of oncologic entities, enabling seamless integration with existing biomedical knowledge. This approach marks a significant advancement in extracting meaningful information from clinical narratives, offering promising applications in cancer research and personalized patient care. Graphical abstract: (Figure presented.)",
keywords = "Breast cancer, EHR, EL, LLM, NER, Neuro-symbolic",
author = "{\'A}lvaro Garc{\'i}a-Barrag{\'a}n and Ahmad Sakor and Vidal, {Maria Esther} and Ernestina Menasalvas and Gonzalez, {Juan Cristobal Sanchez} and Mariano Provencio and V{\'i}ctor Robles",
note = "Publisher Copyright: {\textcopyright} The Author(s) 2024.",
year = "2024",
month = nov,
day = "1",
doi = "10.1007/s11517-024-03227-4",
language = "English",
journal = "Medical and Biological Engineering and Computing",
issn = "0140-0118",
publisher = "Peter Peregrinus, Ltd.",

}

Download

TY - JOUR

T1 - NSSC

T2 - a neuro-symbolic AI system for enhancing accuracy of named entity recognition and linking from oncologic clinical notes

AU - García-Barragán, Álvaro

AU - Sakor, Ahmad

AU - Vidal, Maria Esther

AU - Menasalvas, Ernestina

AU - Gonzalez, Juan Cristobal Sanchez

AU - Provencio, Mariano

AU - Robles, Víctor

N1 - Publisher Copyright: © The Author(s) 2024.

PY - 2024/11/1

Y1 - 2024/11/1

N2 - Abstract: Accurate recognition and linking of oncologic entities in clinical notes is essential for extracting insights across cancer research, patient care, clinical decision-making, and treatment optimization. We present the Neuro-Symbolic System for Cancer (NSSC), a hybrid AI framework that integrates neurosymbolic methods with named entity recognition (NER) and entity linking (EL) to transform unstructured clinical notes into structured terms using medical vocabularies, with the Unified Medical Language System (UMLS) as a case study. NSSC was evaluated on a dataset of clinical notes from breast cancer patients, demonstrating significant improvements in the accuracy of both entity recognition and linking compared to state-of-the-art models. Specifically, NSSC achieved a 33% improvement over BioFalcon and a 58% improvement over scispaCy. By combining large language models (LLMs) with symbolic reasoning, NSSC improves the recognition and interoperability of oncologic entities, enabling seamless integration with existing biomedical knowledge. This approach marks a significant advancement in extracting meaningful information from clinical narratives, offering promising applications in cancer research and personalized patient care. Graphical abstract: (Figure presented.)

AB - Abstract: Accurate recognition and linking of oncologic entities in clinical notes is essential for extracting insights across cancer research, patient care, clinical decision-making, and treatment optimization. We present the Neuro-Symbolic System for Cancer (NSSC), a hybrid AI framework that integrates neurosymbolic methods with named entity recognition (NER) and entity linking (EL) to transform unstructured clinical notes into structured terms using medical vocabularies, with the Unified Medical Language System (UMLS) as a case study. NSSC was evaluated on a dataset of clinical notes from breast cancer patients, demonstrating significant improvements in the accuracy of both entity recognition and linking compared to state-of-the-art models. Specifically, NSSC achieved a 33% improvement over BioFalcon and a 58% improvement over scispaCy. By combining large language models (LLMs) with symbolic reasoning, NSSC improves the recognition and interoperability of oncologic entities, enabling seamless integration with existing biomedical knowledge. This approach marks a significant advancement in extracting meaningful information from clinical narratives, offering promising applications in cancer research and personalized patient care. Graphical abstract: (Figure presented.)

KW - Breast cancer

KW - EHR

KW - EL

KW - LLM

KW - NER

KW - Neuro-symbolic

UR - http://www.scopus.com/inward/record.url?scp=85208104355&partnerID=8YFLogxK

U2 - 10.1007/s11517-024-03227-4

DO - 10.1007/s11517-024-03227-4

M3 - Article

AN - SCOPUS:85208104355

JO - Medical and Biological Engineering and Computing

JF - Medical and Biological Engineering and Computing

SN - 0140-0118

ER -