Hierarchical embedding attention for overall survival prediction in lung cancer from unstructured EHRs

• Published in: BMC medical informatics and decision making Vol. 25; no. 1; pp. 169 - 16
• Main Authors: Paolo, Domenico, Greco, Carlo, Cortellini, Alessio, Ramella, Sara, Soda, Paolo, Bria, Alessandro, Sicilia, Rosa
• Format: Journal Article
• Language: English
• Published: London BioMed Central 18.04.2025; BioMed Central Ltd; BMC
• Subjects: Attention mechanism; Automation; Carcinoma, Non-Small-Cell Lung - mortality; Cell survival; Decision making; Deep learning; Effectiveness; Electronic health records; Electronic Health Records - statistics & numerical data; Electronic medical records; Electronic records; Embedding; Feature extraction; Female; Health aspects; Health Informatics; Humans; Information processing; Information Systems and Communication Service; Lung cancer; Lung cancer, Non-small cell; Lung Neoplasms - mortality; Male; Management of Computing and Information Systems; Medical history; Medical prognosis; Medical records; Medicine; Medicine & Public Health; Methods; Natural Language Processing; Natural language processing in medical informatics; NER; Non-small cell lung carcinoma; Patient outcomes; Patients; Prediction models; Prognosis; Representations; Small cell lung carcinoma; Statistical analysis; Statistical models; Subject specialists; Survival; Survival Analysis; Transformer; Unstructured data; Unstructured EHRs; Italy; Attention mechanism; Survival analysis; NER; Unstructured EHRs; Transformer; Lung cancer
• ISSN: 1472-6947; 1472-6947
• DOI: 10.1186/s12911-025-02998-6

The automated processing of Electronic Health Records (EHRs) poses a significant challenge due to their unstructured nature, rich in valuable, yet disorganized information. Natural Language Processing (NLP), particularly Named Entity Recognition (NER), has been instrumental in extracting structured information from EHR data. However, existing literature primarly focuses on extracting handcrafted clinical features through NLP and NER methods without delving into their learned representations. In this work, we explore the untapped potential of these representations by considering their contextual richness and entity-specific information. Our proposed methodology extracts representations generated by a transformer-based NER model on EHRs data, combines them using a hierarchical attention mechanism, and employs the obtained enriched representation as input for a clinical prediction model. Specifically, this study addresses Overall Survival (OS) in Non-Small Cell Lung Cancer (NSCLC) using unstructured EHRs data collected from an Italian clinical centre encompassing 838 records from 231 lung cancer patients. Whilst our study is applied on EHRs written in Italian, it serves as use case to prove the effectiveness of extracting and employing high level textual representations that capture relevant information as named entities. Our methodology is interpretable because the hierarchical attention mechanism highlights the information in EHRs that the model considers the most crucial during the decision-making process. We validated this interpretability by measuring the agreement of domain experts on the importance assigned by the hierarchical attention mechanism to EHRs information through a questionnaire. Results demonstrate the effectiveness of our method, showcasing statistically significant improvements over traditional manually extracted clinical features.