Detecting insertion, substitution, and deletion errors in radiology reports using neural sequence-to-sequence models

Zech, John R.; Forde, Jessica Zosa; Titano, J.; Kaji, Deepak; Costa, Anthony; Oermann, Eric K.

doi:10.21037/atm.2018.08.11

Cited by 10 publications

(3 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This category is for publications that have a primary technical aim that is not focused on radiology report outcome, e.g. detecting negation in reports, spelling correction [106], fact checking [107,108] methods for sample selection, crowd source annotation [109]. This category did not occur in Pons' earlier review.…”

Section: Technical Nlpmentioning

confidence: 99%

A systematic review of natural language processing applied to radiology reports

Casey

Davidson

Poon

et al. 2021

BMC Med Inform Decis Mak

119

View full text Add to dashboard Cite

Background Natural language processing (NLP) has a significant role in advancing healthcare and has been found to be key in extracting structured information from radiology reports. Understanding recent developments in NLP application to radiology is of significance but recent reviews on this are limited. This study systematically assesses and quantifies recent literature in NLP applied to radiology reports. Methods We conduct an automated literature search yielding 4836 results using automated filtering, metadata enriching steps and citation search combined with manual review. Our analysis is based on 21 variables including radiology characteristics, NLP methodology, performance, study, and clinical application characteristics. Results We present a comprehensive analysis of the 164 publications retrieved with publications in 2019 almost triple those in 2015. Each publication is categorised into one of 6 clinical application categories. Deep learning use increases in the period but conventional machine learning approaches are still prevalent. Deep learning remains challenged when data is scarce and there is little evidence of adoption into clinical practice. Despite 17% of studies reporting greater than 0.85 F1 scores, it is hard to comparatively evaluate these approaches given that most of them use different datasets. Only 14 studies made their data and 15 their code available with 10 externally validating results. Conclusions Automated understanding of clinical narratives of the radiology reports has the potential to enhance the healthcare process and we show that research in this field continues to grow. Reproducibility and explainability of models are important if the domain is to move applications into clinical use. More could be done to share code enabling validation of methods on different institutional data and to reduce heterogeneity in reporting of study properties allowing inter-study comparisons. Our results have significance for researchers in the field providing a systematic synthesis of existing work to build on, identify gaps, opportunities for collaboration and avoid duplication.

show abstract

Section: Technical Nlpmentioning

confidence: 99%

A systematic review of natural language processing applied to radiology reports

Casey

Davidson

Poon

et al. 2021

BMC Med Inform Decis Mak

119

View full text Add to dashboard Cite

show abstract

“…This category is for publications that have a primary technical aim that is not focused on radiology report outcome, e.g. detecting negation in reports, spelling correction [Zech et al, 2019], fact checking , Steinkamp et al, 2019 methods for sample selection, crowd source annotation [Cocos et al, 2017]. This category did not occur in Pons' earlier review.…”

Section: Technical Nlpmentioning

confidence: 99%

A Systematic Review of Natural Language Processing Applied to Radiology Reports

Casey,

Davidson,

Poon

et al. 2021

Preprint

View full text Add to dashboard Cite

Background Natural language processing (NLP) has a significant role in advancing healthcare and has been found to be key in extracting structured information from radiology reports. Understanding recent developments in NLP application to radiology is of significance but recent reviews on this are limited. This study systematically assesses and quantifies recent literature in NLP applied to radiology reports.Methods We conduct an automated literature search yielding 4,799 results using automated filtering, metadata enriching steps and citation search combined with manual review. Our analysis is based on 21 variables including radiology characteristics, NLP methodology, performance, study, and clinical application characteristics.Results We present a comprehensive analysis of the 164 publications retrieved with publications in 2019 almost triple those in 2015. Each publication is categorised into one of 6 clinical application categories. Deep learning use increases in the period but conventional machine learning approaches are still prevalent. Deep learning remains challenged when data is scarce and there is little evidence of adoption into clinical practice. Despite 17% of studies reporting greater than 0.85 F1 scores, it is hard to comparatively evaluate these approaches given that most of them use different datasets. Only 14 studies made their data and 15 their code available with 10 externally validating results.Conclusions Automated understanding of clinical narratives of the radiology reports has the potential to enhance the healthcare process and we show that research in this field continues to grow. Reproducibility and explainability of models

show abstract

“…These cases are often called nonwords [83] and also have typical characteristics [33]. The high number of misspells in radiologic reports is also already established internationally [168].…”

Section: Automatic Correction Of the Textmentioning

confidence: 99%

Utilization of Underlying Semantic Information in Textual Data

Kicsi

View full text Add to dashboard Cite

Ever since my childhood, I knew that the world is a wonderful place where one can find magic in nearly everything, only the right viewpoint is needed. This can even apply to a doctoral thesis. I can promise, however, that whatever viewpoint the reader adopts, Magic will still be encountered through these pages in one way or another.I was always fascinated with stories. They make us think, motivate us, and each of us stars in many little stories of our own every single day. Our stories consist of a cavalcade of exciting words, and together they form our personality. We are going to embark on a journey spanning pharmaceutical supply system features, finding targets for software tests, and even dealing with human spine disorders, but the underlying theme is always the versatility and usefulness of words. Modern technology enables us to teach computers to work with these, extract their meaning, or even use them to construct their own understanding. And this is a wonder by itself, even if we can explain it.It is my most sincere conviction that other people are what truly make our life worth living. I am incredibly grateful that I have the good fortune of knowing so many great people. I would like to extend an extra special thanks to my family, who have given me a brain and a heart to face the world, and have continuously supported me through it. I would also like to thank my love and my wonderful friends, who truly give merit to my life. Thus, if we assess life in such a way, then mine is truly blessed.There are so many who helped me along the journey of my studies that I couldn't possibly name them all. My utmost gratitude goes out to my supervisor, László Vidács, who, in my opinion, is the best supervisor I could have ever wished for, and not just for my doctoral studies, but also my work. I will be forever grateful for all the guidance and care he provided through the years. I do not consider myself easily inspired by outside sources, and yet, to this day, he can always manage this. I am exceptionally thankful for my amazing co-authors and my dear colleagues who greatly contributed to the success of our research. To name a few of them, I would like to thank Viktor Csuvik, Klaudia Szabó Ledenyi, Péter Pusztai, and Ferenc Horváth without whom I would have much less to write about now, as their devoted work was indispensable for my own. I would also like to thank Tibor Gyimóthy for providing me with an offer for doctoral studies at the Department of Software Engineering, and many interesting research opportunities ever since. It was a privilege to conduct my studies in such company. 1

show abstract

Detecting insertion, substitution, and deletion errors in radiology reports using neural sequence-to-sequence models

Cited by 10 publications

References 23 publications

A systematic review of natural language processing applied to radiology reports

A systematic review of natural language processing applied to radiology reports

A Systematic Review of Natural Language Processing Applied to Radiology Reports

Utilization of Underlying Semantic Information in Textual Data

Contact Info

Product

Resources

About