Machine learning based natural language processing of radiology reports in orthopaedic trauma

Olthof, A; Shouche, P; Fennema, Eelco M; IJpma, Frank F. A.; Koolstra, R H C; Stirler, Vincent; Ooijen, Peter M. A. van; Cornelissen, Ludo J.

doi:10.1016/j.cmpb.2021.106304

Cited by 47 publications

(21 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Current NLP approaches are limited to a few languages like English or Dutch. 142 , 143 The research can be expanded to include other languages to make the best use of the plethora of available data in regional languages. This can be useful in expanding the reach of NLP systems and improving the performance of the current state-of-the-art algorithms.…”

Section: Discussionmentioning

confidence: 99%

How can natural language processing help model informed drug development?: a review

et al. 2022

View full text Add to dashboard Cite

ObjectiveTo summarize applications of natural language processing (NLP) in model informed drug development (MIDD) and identify potential areas of improvement.Materials and MethodsPublications found on PubMed and Google Scholar, websites and GitHub repositories for NLP libraries and models. Publications describing applications of NLP in MIDD were reviewed. The applications were stratified into 3 stages: drug discovery, clinical trials, and pharmacovigilance. Key NLP functionalities used for these applications were assessed. Programming libraries and open-source resources for the implementation of NLP functionalities in MIDD were identified.ResultsNLP has been utilized to aid various processes in drug development lifecycle such as gene-disease mapping, biomarker discovery, patient-trial matching, adverse drug events detection, etc. These applications commonly use NLP functionalities of named entity recognition, word embeddings, entity resolution, assertion status detection, relation extraction, and topic modeling. The current state-of-the-art for implementing these functionalities in MIDD applications are transformer models that utilize transfer learning for enhanced performance. Various libraries in python, R, and Java like huggingface, sparkNLP, and KoRpus as well as open-source platforms such as DisGeNet, DeepEnroll, and Transmol have enabled convenient implementation of NLP models to MIDD applications.DiscussionChallenges such as reproducibility, explainability, fairness, limited data, limited language-support, and security need to be overcome to ensure wider adoption of NLP in MIDD landscape. There are opportunities to improve the performance of existing models and expand the use of NLP in newer areas of MIDD.ConclusionsThis review provides an overview of the potential and pitfalls of current NLP approaches in MIDD.

show abstract

Section: Discussionmentioning

confidence: 99%

How can natural language processing help model informed drug development?: a review

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Zheng et al developed an NLP algorithm to identify pulmonary nodules and the associated characteristics with high accuracy ( 18 ). Furthermore, a recent study compared different machine learning NLP methods to classify radiology reports in orthopedic trauma for injuries and found that BERT NLP outperformed traditional machine learning models and rule-based classifiers for Dutch radiology reports in orthopedic trauma ( 19 ). However, no NLP-based deep learning algorithm has been reported in the field of cardiovascular disease research.…”

Section: Discussionmentioning

confidence: 99%

Using Text Content From Coronary Catheterization Reports to Predict 5-Year Mortality Among Patients Undergoing Coronary Angiography: A Deep Learning Approach

Lee

Chen

et al. 2022

Front. Cardiovasc. Med.

View full text Add to dashboard Cite

BackgroundCurrent predictive models for patients undergoing coronary angiography have complex parameters which limit their clinical application. Coronary catheterization reports that describe coronary lesions and the corresponding interventions provide information of the severity of the coronary artery disease and the completeness of the revascularization. This information is relevant for predicting patient prognosis. However, no predictive model has been constructed using the text content from coronary catheterization reports before.ObjectiveTo develop a deep learning model using text content from coronary catheterization reports to predict 5-year all-cause mortality and 5-year cardiovascular mortality for patients undergoing coronary angiography and to compare the performance of the model to the established clinical scores.MethodThis retrospective cohort study was conducted between January 1, 2006, and December 31, 2015. Patients admitted for coronary angiography were enrolled and followed up until August 2019. The main outcomes were 5-year all-cause mortality and 5-year cardiovascular mortality. In total, 11,576 coronary catheterization reports were collected. BioBERT (bidirectional encoder representations from transformers for biomedical text mining), which is a BERT-based model in the biomedical domain, was utilized to construct the model. The area under the receiver operating characteristic curve (AUC) was used to assess model performance. We also compared our results to the residual SYNTAX (SYNergy between PCI with TAXUS and Cardiac Surgery) score.ResultsThe dataset was divided into the training (60%), validation (20%), and test (20%) sets. The mean age of the patients in each dataset was 65.5 ± 12.1, 65.4 ± 11.2, and 65.6 ± 11.2 years, respectively. A total of 1,411 (12.2%) patients died, and 664 (5.8%) patients died of cardiovascular causes within 5 years after coronary angiography. The best of our models had an AUC of 0.822 (95% CI, 0.790–0.855) for 5-year all-cause mortality, and an AUC of 0.858 (95% CI, 0.816–0.900) for 5-year cardiovascular mortality. We randomly selected 300 patients who underwent percutaneous coronary intervention (PCI), and our model outperformed the residual SYNTAX score in predicting 5-year all-cause mortality (AUC, 0.867 [95% CI, 0.813–0.921] vs. 0.590 [95% CI, 0.503–0.684]) and 5-year cardiovascular mortality (AUC, 0.880 [95% CI, 0.873–0.925] vs. 0.649 [95% CI, 0.535–0.764]), respectively, after PCI among these patients.ConclusionsWe developed a predictive model using text content from coronary catheterization reports to predict the 5-year mortality in patients undergoing coronary angiography. Since interventional cardiologists routinely write reports after procedures, our model can be easily implemented into the clinical setting.

show abstract

“…A study on natural language processing of radiology reports in orthopaedic trauma was carried out by [16]. This study compared different machine learning approach to classify presence of injury(ies) from radiology reports in orthopaedic trauma.…”

Section: Related Workmentioning

confidence: 99%

Design and Implementation of Diagnosis System for Cardiomegaly from Clinical Chest X-ray Reports

Ogungbe¹,

Iyanda²,

Aderibigbe³

2022

IJEM

View full text Add to dashboard Cite

With the increasingly broadening adoption of Electronic Health Record (EHR) worldwide, there is a growing need to widen the use of EHR to support clinical decision making and research particularly in radiology. A number of studies on generation, analysis and presentation of chest x-ray reports from digital images to detect abnormalities have been well documented in the literature but studies on automatic analysis of chest x-ray reports have not been well represented. Interestingly, there is a large amount of unstructured electronic chest x-ray notes that need to be organized and processed in such a way that it can be automated for the purpose of giving urgent attention to abnormal radiographs in clinical findings to allow for quicker report analysis and decision making. This study developed a system to automate this analysis in order to prioritize findings from chest x-rays using support vector machine and Lagrange Multiplier for the constraint optimization. The classification model was implemented using Python programming language and Django framework. The developed system was evaluated based on precision, recall, f1-score, negative predictive value (NPV). Expert's knowledge was also used as gold standard and comparison with the existing system. The result showed a precision of 96.04%, recall of 95.10%, f1-score of 95.57%, specificity of 86.21%, negative predictive value of 83.33% and an accuracy of 93.13%. The study revealed that a limited but important number of relevant attributes provided an effective and efficient model for the detection of cardiomegaly in clinical chest x-ray reports. From the evaluation result, it is evident that this system can help the clinicians to quickly prioritize findings from chest x-ray reports, thereby reducing the delay in attending to patients. Hence, the developed system could be used for the analysis of chest x-ray reports with the purpose of diagnosing the patient for cardiomegaly. Chest X-ray reports are usually textual, therefore, further studies can introduce spell checker to the system to provide higher sensitivity.

show abstract

Machine learning based natural language processing of radiology reports in orthopaedic trauma

Cited by 47 publications

References 35 publications

How can natural language processing help model informed drug development?: a review

How can natural language processing help model informed drug development?: a review

Using Text Content From Coronary Catheterization Reports to Predict 5-Year Mortality Among Patients Undergoing Coronary Angiography: A Deep Learning Approach

Design and Implementation of Diagnosis System for Cardiomegaly from Clinical Chest X-ray Reports

Contact Info

Product

Resources

About