Komenti: A semantic text mining framework

Karwath

Williams

et al. 2021

Preprint

Self Cite

Ontology-based phenotype profiles have been utilised for the purpose of differential diagnosis of rare genetic diseases, and for decision support in specific disease domains. Particularly, semantic similarity facilitates diagnostic hypothesis generation through comparison with disease phenotype profiles. However, the approach has not been applied for differential diagnosis of common diseases, or generalised clinical diagnostics from uncurated text-derived phenotypes. In this work, we describe the development of an approach for deriving patient phenotype profiles from clinical narrative text, and apply this to text associated with MIMIC-III patient visits. We then explore the use of semantic similarity with those text-derived phenotypes to classify primary patient diagnosis, comparing the use of patient-patient similarity and patient-disease similarity using phenotype-disease profiles previously mined from literature. The results reveal a potentially powerful approach that can be applied to a variety of clinical tasks, such as differential diagnosis, cohort discovery, document and text classification, and outcome prediction.

Section: Methodsmentioning

confidence: 99%

Towards Similarity-based Differential Diagnostics For Common Diseases

Karwath

Williams

et al. 2021

Preprint

Self Cite

“…The algorithm was developed using the Komenti semantic text mining framework, which is available under an open source licence at http://github.com/reality/Komenti [16]. It makes use of Stanford CoreNLP [7], and particularly the RegexNER annotator for identifying entities in text, and the openIE algorithm for extracting triples from text.…”

Section: Methodsmentioning

confidence: 99%

“…The triple mining and ontology construction modules are available as part of the Komenti semantic text mining framework, which is available under an open source licence at https://github.com/reality/komenti [16]. Files used for the construction of the MIMIC Adapted Disease Ontology (MADO), and for the automated validation are available at https://github.com/reality/mado_tripulate.…”

Section: /11mentioning

confidence: 99%

“…For the vocabulary construction and annotation stages, our approach uses Komenti's existing pipeline for vocabulary construction and text annotation [16]. The vocabulary construction process uses the AberOWL reasoner-based ontology access framework [17] to build a list of labels and synonyms for a given input ontology.…”

Section: Vocabulary Construction and Annotationmentioning

confidence: 99%

See 1 more Smart Citation

Exploring Binary Relations for Ontology Extension and Improved Adaptation to Clinical Text

Hoehndorf

Karwath

et al. 2020

Preprint

Self Cite

BackgroundThe controlled domain vocabularies provided by ontologies make them an indispensable tool for text mining. Ontologies also include semantic features in the form of taxonomy and axioms, which make annotated entities in text corpora useful for semantic analysis. Extending those semantic features may improve performance for characterisation and analytic tasks. Ontology learning techniques have previously been explored for novel ontology construction from text, though most recent approaches have focused on literature, with applications in information retrieval or human interaction tasks. We hypothesise that extension of existing ontologies using information mined from clinical narrative text may help to adapt those ontologies such that they better characterise those texts, and lead to improved classification performance.ResultsWe develop and present a framework for identifying new classes in text corpora, which can be integrated into existing ontology hierarchies. To do this, we employ the Stanford Open Information Extraction algorithm and integrate its implementation into the Komenti semantic text mining framework. To identify whether our approach leads to better characterisation of text, we present a case study, using the method to learn an adaptation to the Disease Ontology using text associated with a sample of 1,000 patient visits from the MIMIC-III critical care database. We use the adapted ontology to annotate and classify shared first diagnosis on patient visits with semantic similarity, revealing an improved performance over use of the base Disease Ontology on the set of visits the ontology was constructed from. Moreover, we show that the adapted ontology also improved performance for the same task over two additional unseen samples of 1,000 and 2,500 patient visits.ConclusionsWe report a promising new method for ontology learning and extension from text. We demonstrate that we can successfully use the method to adapt an existing ontology to a textual dataset, improving its ability to characterise the dataset, and leading to improved analytic performance, even on unseen portions of the dataset.

“…Figure 1 depicts the seven steps that make up the information extraction pipeline. We implemented the pipeline using the Komenti semantic text-mining framework 14 , a natural language processing program developed in-house. This is available under an open-source licence at https://github.com/reality/Komenti.…”

Section: Information Extraction Frameworkmentioning

confidence: 99%

Making Words Count with Computerised Identification of Hypertrophic Cardiomyopathy Patients

Bradlow

Desai

et al. 2021

Preprint

Self Cite

Background: The traditional outpatient model in hypertrophic cardiomyopathy (HCM) is under pressure. Population health management based on an accurate patient record provides an efficient, cost-effective alternative. Methods: To improve the accuracy of the HCM patient list in a single hospital, we developed a rule-based information extraction natural language processing (NLP) framework. The framework employed ontological expansion of vocabulary and exclusion-first annotation, and received training by an 'expert in the loop'. The output stratified patients with atrial fibrillation (AF) and heart failure (HF), those without active cardiology care and likely screened individuals. Results: The algorithm was validated against multiple data sources, including manual validation, for HCM, AF and HF and family history of the disease. Overall precision and recall were 0.854 and 0.865 respectively. The pipeline found 25,356 documents featuring HCM-related terms belonging to 11,083 patients. Excluding scanned documents resulted in 17,178 letters from 3,120 patients. Subsequent categorisation identified 1,753 real cases, of whom 357 had AF and 205 had HF. There were 696 likely screened individuals. Adjusting for 304 false-negative patients, the total HCM cohort was 2,045 patients. 214 were not under a cardiologist. NLP uncovered 709 patients who were absent in the registry or hospital disease codes. Conclusion: This novel NLP framework generated a hospital-wide record of patients with HCM and defined various cohorts, including the small set of HCM patients lacking current cardiology input. Existing data sources inadequately described this population, spotlighting NLP's essential role for clinical teams planning to move to a population health management model of care. Keywords: text mining, natural language processing, hypertrophic cardiomyopathy, population health management