skip to main content
research-article

Bridging the Gap between Structured and Free-form Radiology Reporting: A Case-study on Coronary CT Angiography

Authors Info & Claims
Published:15 October 2021Publication History
Skip Abstract Section

Abstract

Free-form radiology reports associated with coronary computed tomography angiography (CCTA) include nuanced and complicated linguistics to report cardiovascular disease. Standardization and interpretation of such reports is crucial for clinical use of CCTA. Coronary Artery Disease Reporting and Data System (CAD-RADS) has been proposed to achieve such standardization by implementing a strict template-based report writing and assignment of a score between 0 and 5 indicating the severity of coronary artery lesions. Even after its introduction, free-form unstructured report writing remains popular among radiologists. In this work, we present our attempts at bridging the gap between structured and unstructured reporting by natural language processing. We present machine learning models that while being trained only on structured reports, can predict CAD-RADS scores by analysis of free-text of unstructured radiology reports. The best model achieves 98% accuracy on structured reports and 92% 1-margin accuracy (difference of \(\le\)1 in the predicted and the actual scores) for free-form unstructured reports. Our model also performs well under very difficult circumstances including nuanced and widely varying terminology used for reporting cardiovascular functions and diseases, scarcity of labeled data for training our model, and uneven class label distribution.

REFERENCES

  1. [1] Abd Mohammad, Basha Alkhalik, Aly Sameh Abdelaziz, Abdel Ahmad, Ismail Azim, and Bahaaeldin Hanan A.. 2019. The validity and applicability of CAD-RADS in the management of patients with coronary artery disease. Insights Imag. 10, 117 (2019).Google ScholarGoogle Scholar
  2. [2] Abdel Ahmed, Abdel Khalek, Elrakhawy Mohamed Magdy, Yossof Mahmoud Mohamed, and Nageb Hadeer Mohamed. 2018. Inter-observer agreement of the Coronary Artery Disease Reporting and Data System (CAD-RADS TM) in patients with stable chest pain. Polish J. Radiol. (2018), 151159. Google ScholarGoogle Scholar
  3. [3] Abdel Ahmed, Abdel Khalek, Elrakhawy Mohamed Magdy, Yossof Mahmoud Mohamed, and Nageb Hadeer Mohamed. 2018. Inter observer agreement of the coronary artery disease reporting and data system (CAD-RADS TM) in patients with stable chest pain. Polish J. Radiol. 83, (2018), 151159.Google ScholarGoogle Scholar
  4. [4] Ancona Marco, Ceolini Enea, Öztireli Cengiz, and Gross Markus. 2017. Towards better understanding of gradient-based attribution methods for deep neural networks. arXiv:1711.06104 (2017).Google ScholarGoogle Scholar
  5. [5] Banerjee Imon, Ling Yuan, Chen Matthew C., Hasan Sadid A., Langlotz Curtis P., Moradzadeh Nathaniel, Chapman Brian, Amrhein Timothy, Mong David, Rubin Daniel L., et al. 2019. Comparative effectiveness of convolutional neural network (CNN) and recurrent neural network (RNN) architectures for radiology text report classification. Artif. Intell. Med. 97 (2019), 7988.Google ScholarGoogle ScholarDigital LibraryDigital Library
  6. [6] Banerjee Imon, Madhavan Sriraman, Goldman Roger Eric, and Rubin Daniel L.. 2017. Intelligent word embeddings of free-text radiology reports. In AMIA Annual Symposium Proceedings, Vol. 2017. American Medical Informatics Association.Google ScholarGoogle Scholar
  7. [7] Brown A. D. and Kachura J. R.. 2019. Natural language processing of radiology reports in patients with hepatocellular carcinoma to predict radiology resource utilization. J. Amer. Coll. Radiol. 16, 6 (2019), 840844.Google ScholarGoogle Scholar
  8. [8] Carrodeguas Emmanuel, Lacson Ronilda, Swanson Whitney, and Khorasani Ramin. 2019. Use of machine learning to identify follow-up recommendations in radiology reports. J. Amer. Coll. Radiol. 16, 3 (2019), 336343.Google ScholarGoogle Scholar
  9. [9] Chapman Brian E., Lee Sean, Kang Hyunseok Peter, and Chapman Wendy W.. 2011. Document-level classification of CT pulmonary angiography reports based on an extension of the ConText algorithm. J. Biomed. Inform. 44, 5 (2011), 728737. Google ScholarGoogle ScholarDigital LibraryDigital Library
  10. [10] Cheng Victor Y., Nakazato Ryo, Dey Damini, Gurudevan Swaminatha, Tabak Joshua, Budoff Matthew J., Karlsberg Ronald P., Min James, and Berman Daniel S.. 2009. Reproducibility of coronary artery plaque volume and composition quantification by 64-detector row coronary computed tomographic angiography: An intraobserver, interobserver, and interscan variability study. J. Cardiovasc. Comput. Tomog. 3, 5 (2009), 312320. DOI: https://doi.org/10.1016/j.jcct.2009.07.001Google ScholarGoogle Scholar
  11. [11] Claeys Marc J.. 2014. Guidelines on the management of stable coronary artery disease. Acta Cardiol. 69, 1 (2014), 5152. DOI: https://doi.org/10.2143/AC.69.1.3011345Google ScholarGoogle Scholar
  12. [12] Cury Ricardo C., Abbara Suhny, Achenbach Stephan, Agatston Arthur, Berman Daniel S., Budoff Matthew J., Dill Karin E., Jacobs Jill E., Maroules Christopher D., Rubin Geoffrey D., Rybicki Frank J., Schoepf U. Joseph, Shaw Leslee J., Stillman Arthur E., White Charles S., Woodard Pamela K., and Leipsic Jonathon A.. 2016. Coronary Artery Disease—Reporting and Data System (CAD-RADS). JACC-Cardiovasc. Imag. 9, 9 (2016), 10991113. DOI: https://doi.org/10.1016/j.jcmg.2016.05.005Google ScholarGoogle Scholar
  13. [13] Cury Ricardo C., Abbara Suhny, Achenbach Stephan, Agatston Arthur, Berman Daniel S., Budoff Matthew J., Dill Karin E., Jacobs Jill E., Maroules Christopher D., Rubin Geoffrey D., et al. 2016. CAD-RADSTM coronary artery disease—Reporting and data system.J. Cardiovasc. Comput. Tomog. 10, 4 (2016), 269281.Google ScholarGoogle Scholar
  14. [14] Devlin Jacob, Chang Ming-Wei, Lee Kenton, and Toutanova Kristina. 2018. BERT: Pre-training of deep bidirectional transformers for language understanding. arXiv:1810.04805 (2018).Google ScholarGoogle Scholar
  15. [15] Dreyer Keith J., Kalra Mannudeep K., Maher Michael M., Hurier Autumn M., Asfaw Benjamin A., Schultz Thomas, Halpern Elkan F., and Thrall James H.. 2005. Application of recently developed computer algorithm for automatic classification of unstructured radiology reports: Validation study. Radiology 234, 2 (2005), 323329.Google ScholarGoogle ScholarCross RefCross Ref
  16. [16] Faryna Khrystyna, Tushar Fakrul I., D’Anniballe Vincent M., Hou Rui, Rubin Geoffrey D., and Lo Joseph Y.. 2020. Attention-guided classification of abnormalities in semi-structured computed tomography reports. In Medical Imaging 2020: Computer-Aided Diagnosis, Vol. 11314. International Society for Optics and Photonics, 113141P.Google ScholarGoogle Scholar
  17. [17] Halliburton Sandra S., Abbara Suhny, Chen Marcus Y., Gentry Ralph, Ct R. T. R., Mahesh Mahadevappa, Raff Gilbert L., and Shaw Leslee J.. 2011. SCCT guidelines on radiation dose and dose-optimization strategies in cardiovascular CT. J. Cardiovasc. Comput. Tomog. 5 (2011), 198224. DOI: https://doi.org/10.1016/j.jcct.2011.06.001Google ScholarGoogle Scholar
  18. [18] Hassanpour Saeed and Langlotz Curtis P.. 2016. Predicting high imaging utilization based on initial radiology reports: A feasibility study of machine learning. Acad. Radiol. 23, 1 (2016), 8489.Google ScholarGoogle ScholarCross RefCross Ref
  19. [19] Hassanpour Saeed, Langlotz Curtis P., Amrhein Timothy J., Befera Nicholas T., and Lungren Matthew P.. 2017. Performance of a machine learning classifier of knee MRI reports in two large academic radiology practices: A tool to estimate diagnostic yield. Amer. J. Roentgen. 208, 4 (2017), 750753.Google ScholarGoogle Scholar
  20. [20] Hripcsak George, Austin John H. M., Alderson Philip O., and Friedman Carol. 2002. Use of natural language processing to translate clinical information from a database of 889,921 chest radiographic reports. Radiology 224, 1 (2002), 157163.Google ScholarGoogle Scholar
  21. [21] Johnson Melvin, Schuster Mike, Le Quoc V., Krikun Maxim, Wu Yonghui, Chen Zhifeng, Thorat Nikhil, Viégas Fernanda, Wattenberg Martin, Corrado Greg, et al. 2017. Google’s multilingual neural machine translation system: Enabling zero-shot translation. Trans. Assoc. Comput. Ling. 5 (2017), 339351.Google ScholarGoogle ScholarCross RefCross Ref
  22. [22] Kim Yoon. 2014. Convolutional neural networks for sentence classification. In Proceedings of Empirical Methods in Natural Language Processing (EMNLP’14). 17461751.Google ScholarGoogle ScholarCross RefCross Ref
  23. [23] Le Quoc and Mikolov Tomas. 2014. Distributed representations of sentences and documents. In Proceedings of the International Conference on Machine Learning. 11881196. Google ScholarGoogle ScholarDigital LibraryDigital Library
  24. [24] Lee Jinhyuk, Yoon Wonjin, Kim Sungdong, Kim Donghyeon, Kim Sunkyu, So Chan Ho, and Kang Jaewoo. 2020. BioBERT: A pre-trained biomedical language representation model for biomedical text mining. Bioinformatics 36, 4 (2020), 12341240.Google ScholarGoogle ScholarCross RefCross Ref
  25. [25] Leipsic Jonathon, Abbara Suhny, Achenbach Stephan, Cury Ricardo, Earls James P., John G. B., Nieman Koen, Pontone Gianluca, Raff Gilbert L., and Consultants Fairfax Radiological. 2014. ScienceDirect SCCT guidelines for the interpretation and reporting of coronary CT angiography: A report of the Society of Cardiovascular Computed Tomography Guidelines Committee. J. Cardiovasc. Comput. Tomog. 8, 5 (2014), 342358. DOI: https://doi.org/10.1016/j.jcct.2014.07.003Google ScholarGoogle Scholar
  26. [26] Maroules Christopher D., Hamilton-Craig Christian, Branch Kelley, Lee James, Cury Roberto C., Maurovich-horvat Pál, Rubinshtein Ronen, Thomas Dustin, Williams Michelle, Guo Yanshu, and Cury Ricardo C.. 2018. Coronary artery disease reporting and data system (CAD-RADS TM): Inter-observer agreement for assessment categories and modifiers?J. Cardiovasc. Comput. Tomog. 12, 2 (2018), 125130. DOI: https://doi.org/10.1016/j.jcct.2017.11.014Google ScholarGoogle Scholar
  27. [27] Miao Shumei, Xu Tingyu, Wu Yonghui, Xie Hui, Wang Jingqi, Jing Shenqi, Zhang Yaoyun, Zhang Xiaoliang, Yang Yinshuang, Zhang Xin, et al. 2018. Extraction of BI-RADS findings from breast ultrasound reports in Chinese using deep learning approaches. Int. J. Med. Inform. 119 (2018), 1721.Google ScholarGoogle Scholar
  28. [28] Mikolov Tomas, Chen Kai, Corrado Greg, and Dean Jeffrey. 2013. Efficient estimation of word representations in vector space. arXiv:1301.3781 (2013).Google ScholarGoogle Scholar
  29. [29] Mikolov Tomas, Sutskever Ilya, Chen Kai, Corrado Greg S., and Dean Jeff. 2013. Distributed representations of words and phrases and their compositionality. In Proceedings of the International Conference on Advances in Neural Information Processing Systems. 31113119. Google ScholarGoogle ScholarDigital LibraryDigital Library
  30. [30] Montalescot Gilles, Sechtem Udo, Achenbach Stephan, Andreotti Felicita, Arden Chris, Budaj Andrzej, Bugiardini Raffaele, Crea Filippo, Cuisset Thomas, Mario Carlo Di, Ferreira J. Rafael, Gersh Bernard J., Gitt Anselm K., Hulot Jean Sebastien, Marx Nikolaus, Opie Lionel H., Pfisterer Matthias, Prescott Eva, Ruschitzka Frank, Sabaté Manel, Senior Roxy, Taggart David Paul, Wall Ernst E. Van Der, Vrints Christiaan J. M., Zamorano Jose Luis, Baumgartner Helmut, Bax Jeroen J., Bueno Héctor, Dean Veronica, Deaton Christi, Erol Cetin, Fagard Robert, Ferrari Roberto, Hasdai David, Hoes Arno W., Kirchhof Paulus, Knuuti Juhani, Kolh Philippe, Lancellotti Patrizio, Linhart Ales, Nihoyannopoulos Petros, Piepoli Massimo F., Ponikowski Piotr, Sirnes Per Anton, Tamargo Juan Luis, Tendera Michal, Torbicki Adam, Wijns William, Windecker Stephan, Valgimigli Marco, Claeys Marc J., Donner-Banzhoff Norbert, Frank Herbert, Funck-Brentano Christian, Gaemperli Oliver, Gonzalez-Juanatey José R., Hamilos Michalis, Husted Steen, James Stefan K., Kervinen Kari, Kristensen Steen Dalby, Maggioni Aldo Pietro, Romeo Francesco, Rydén Lars, Simoons Maarten L., Steg Ph Gabriel, Timmis Adam, and Yildirir Aylin. 2013. 2013 ESC guidelines on the management of stable coronary artery disease. Eur. Heart J. 34, 38 (2013), 29493003. DOI: https://doi.org/10.1093/eurheartj/eht296.Google ScholarGoogle Scholar
  31. [31] Nowak Jakub, Taspinar Ahmet, and Scherer Rafał. 2017. LSTM recurrent neural networks for short text and sentiment classification. In Proceedings of the International Conference on Artificial Intelligence and Soft Computing. Springer, 553562.Google ScholarGoogle ScholarCross RefCross Ref
  32. [32] Pagali Sandeep R., Madaj Paul, Gupta Mohit, Nair Subu, Hamirani Yasmin S., Min James K., Lin Faye, and Budoff Matthew J.. 2010. Interobserver variations of plaque severity score and segment stenosis score in coronary arteries using 64 slice multidetector computed tomography: A substudy of the ACCURACY trial. J. Cardiovasc. Comput. Tomog. 4, 5 (2010), 312318. DOI: https://doi.org/10.1016/j.jcct.2010.05.018Google ScholarGoogle Scholar
  33. [33] Sousa-Uva Miguel, Neumann Franz Josef, Ahlsson Anders, Alfonso Fernando, Banning Adrian P., Benedetto Umberto, Byrne Robert A., Collet Jean Philippe, Falk Volkmar, Head Stuart J., Jüni Peter, Kastrati Adnan, Koller Akos, Kristensen Steen D., Niebauer Josef, Richter Dimitrios J., Seferovic Petar M., Sibbing Dirk, Stefanini Giulio G., Windecker Stephan, Yadav Rashmi, Zembala Michael O., Wijns William, Glineur David, Aboyans Victor, Achenbach Stephan, Agewall Stefan, Andreotti Felicita, Barbato Emanuele, Baumbach Andreas, Brophy James, Bueno Héctor, Calvert Patrick A., Capodanno Davide, Davierwala Piroze M., Delgado Victoria, Dudek Dariusz, Freemantle Nick, Funck-Brentano Christian, Gaemperli Oliver, Gielen Stephan, Gilard Martine, Gorenek Bulent, Haasenritter Joerg, Haude Michael, Ibanez Borja, Iung Bernard, Jeppsson Anders, Katritsis Demosthenes, Knuuti Juhani, Kolh Philippe, Leite-Moreira Adelino, Lund Lars H., Maisano Francesco, Mehilli Julinda, Metzler Bernhard, Montalescot Gilles, Pagano Domenico, Petronio Anna Sonia, Piepoli Massimo Francesco, Popescu Bogdan A., Sádaba Rafael, Shlyakhto Evgeny, Silber Sigmund, Simpson Iain A., Sparv David, Tavilla Giuseppe, Thiele Holger, Tousek Petr, Belle Eric Van, Vranckx Pascal, Witkowski Adam, Zamorano Jose Luis, Roffi Marco, Coca Antonio, Coman Ioan Mircea, Dean Veronica, Fitzsimons Donna, Hindricks Gerhard, Katus Hugo A., Lancellotti Patrizio, Leclercq Christophe, McDonagh Theresa A., Ponikowski Piotr, Chettibi Mohamed, Sisakian Hamayak, Äbrahimov Firdovsi, Stelmashok Valeriy I., Postadzhiyan Arman, Skoric Bosko, Eftychiou Christos, Kala Petr, Terkelsen Christian Juhl, Magdy Ahmed, Eha Jaan, Niemelä Matti, Kedev Sasko, Motreff Pascal, Aladashvili Alexander, Kanakakis Ioannis Georgios, Becker David, Gudnason Thorarinn, Peace Aaron, Romeo Francesco, Bajraktari Gani, Kerimkulova Alina, Rudzitis Ainars, Ghazzal Ziad, Kibarskis Aleksandras, Pereira Bruno, Xuereb Robert G., Hofma Sjoerd H., Steigen Terje K., Oliveira Eduardo Infante De, Mot Stefan, Duplyakov Dmitry, Zavatta Marco, Beleslin Branko, Kovar Frantisek, Bunc Matjaz, Ojeda Soledad, Witt Nils, Jeger Raban, Addad Faouzi, Akdemir Ramazan, Parkhomenko Alexander, and Henderson Robert. 2019. 2018 ESC/EACTS Guidelines on myocardial revascularization. Eur. J. Cardio-thorac. Surg. 55, 1 (2019), 490. DOI: https://doi.org/10.1093/ejcts/ezy289Google ScholarGoogle Scholar
  34. [34] Sun Chi, Qiu Xipeng, Xu Yige, and Huang Xuanjing. 2019. How to fine-tune bert for text classification? In Proceedings of the China National Conference on Chinese Computational Linguistics. Springer, 194206.Google ScholarGoogle ScholarDigital LibraryDigital Library
  35. [35] Tang Duyu, Qin Bing, and Liu Ting. 2015. Document modeling with gated recurrent neural network for sentiment classification. In Proceedings of the Conference on Empirical Methods in Natural Language Processing. 14221432.Google ScholarGoogle ScholarCross RefCross Ref
  36. [36] Taylor Allen J., Cerqueira Manuel, Hodgson John McB., Mark Daniel, Min James, O’Gara Patrick, and Rubin Geoffrey D.. 2010. ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/SCAI/SCMR 2010 appropriate use criteria for cardiac computed tomography. J. Amer. Coll. Cardiol. 56, 22 (2010), 18641894. DOI: https://doi.org/10.1016/j.jacc.2010.07.005Google ScholarGoogle Scholar
  37. [37] Wang Yequan, Huang Minlie, Zhu Xiaoyan, and Zhao Li. 2016. Attention-based LSTM for aspect-level sentiment classification. In Proceedings of the Conference on Empirical Methods in Natural Language Processing. 606615.Google ScholarGoogle ScholarCross RefCross Ref
  38. [38] Wang Yanshan, Sohn Sunghwan, Liu Sijia, Shen Feichen, Wang Liwei, Atkinson Elizabeth J., Amin Shreyasee, and Liu Hongfang. 2019. A clinical text classification paradigm using weak supervision and deep representation. BMC Medical Informatics Decis. Mak. 19, 1 (2019), 1.Google ScholarGoogle Scholar
  39. [39] WHO. 2017. World Health Organization (2017) Key facts Cardiovascular Diseases (CVDs). Retrieved from https://www.who.int/cardiovascular_diseases/en/.Google ScholarGoogle Scholar
  40. [40] Xu Kelvin, Ba Jimmy, Kiros Ryan, Cho Kyunghyun, Courville Aaron, Salakhudinov Ruslan, Zemel Rich, and Bengio Yoshua. 2015. Show, attend and tell: Neural image caption generation with visual attention. In Proceedings of the International Conference on Machine Learning. 20482057.Google ScholarGoogle ScholarDigital LibraryDigital Library

Index Terms

  1. Bridging the Gap between Structured and Free-form Radiology Reporting: A Case-study on Coronary CT Angiography

        Recommendations

        Comments

        Login options

        Check if you have access through your login credentials or your institution to get full access on this article.

        Sign in

        Full Access

        • Published in

          cover image ACM Transactions on Computing for Healthcare
          ACM Transactions on Computing for Healthcare  Volume 3, Issue 1
          January 2022
          255 pages
          ISSN:2691-1957
          EISSN:2637-8051
          DOI:10.1145/3485154
          Issue’s Table of Contents

          Copyright © 2021 Association for Computing Machinery.

          Publisher

          Association for Computing Machinery

          New York, NY, United States

          Publication History

          • Published: 15 October 2021
          • Accepted: 1 July 2021
          • Revised: 1 May 2021
          • Received: 1 July 2020
          Published in health Volume 3, Issue 1

          Permissions

          Request permissions about this article.

          Request Permissions

          Check for updates

          Qualifiers

          • research-article
          • Refereed
        • Article Metrics

          • Downloads (Last 12 months)42
          • Downloads (Last 6 weeks)6

          Other Metrics

        PDF Format

        View or Download as a PDF file.

        PDF

        eReader

        View online with eReader.

        eReader

        Full Text

        View this article in Full Text.

        View Full Text

        HTML Format

        View this article in HTML Format .

        View HTML Format
        About Cookies On This Site

        We use cookies to ensure that we give you the best experience on our website.

        Learn more

        Got it!