Advertisement

Journal of Digital Imaging

, Volume 31, Issue 2, pp 245–251 | Cite as

Automatic Determination of the Need for Intravenous Contrast in Musculoskeletal MRI Examinations Using IBM Watson’s Natural Language Processing Algorithm

  • Hari Trivedi
  • Joseph Mesterhazy
  • Benjamin Laguna
  • Thienkhai Vu
  • Jae Ho Sohn
Article

Abstract

Magnetic resonance imaging (MRI) protocoling can be time- and resource-intensive, and protocols can often be suboptimal dependent upon the expertise or preferences of the protocoling radiologist. Providing a best-practice recommendation for an MRI protocol has the potential to improve efficiency and decrease the likelihood of a suboptimal or erroneous study. The goal of this study was to develop and validate a machine learning-based natural language classifier that can automatically assign the use of intravenous contrast for musculoskeletal MRI protocols based upon the free-text clinical indication of the study, thereby improving efficiency of the protocoling radiologist and potentially decreasing errors. We utilized a deep learning-based natural language classification system from IBM Watson, a question-answering supercomputer that gained fame after challenging the best human players on Jeopardy! in 2011. We compared this solution to a series of traditional machine learning-based natural language processing techniques that utilize a term-document frequency matrix. Each classifier was trained with 1240 MRI protocols plus their respective clinical indications and validated with a test set of 280. Ground truth of contrast assignment was obtained from the clinical record. For evaluation of inter-reader agreement, a blinded second reader radiologist analyzed all cases and determined contrast assignment based on only the free-text clinical indication. In the test set, Watson demonstrated overall accuracy of 83.2% when compared to the original protocol. This was similar to the overall accuracy of 80.2% achieved by an ensemble of eight traditional machine learning algorithms based on a term-document matrix. When compared to the second reader’s contrast assignment, Watson achieved 88.6% agreement. When evaluating only the subset of cases where the original protocol and second reader were concordant (n = 251), agreement climbed further to 90.0%. The classifier was relatively robust to spelling and grammatical errors, which were frequent. Implementation of this automated MR contrast determination system as a clinical decision support tool may save considerable time and effort of the radiologist while potentially decreasing error rates, and require no change in order entry or workflow.

Keywords

IBM Watson Machine learning Artificial intelligence Deep learning Natural language processing (NLP) Imaging protocol Workflow efficiency Quality improvement 

Abbreviations

IRB

Institutional Review Board

MRI

Magnetic resonance imaging

NC

Non-contrast

NLP

Natural language processing

WC

With contrast

Notes

Funding Information

HT was supported by an NIH T32 Fellowship, 5T32EB001631-10. JHS was supported by the NVIDIA academic grant program.

Compliance with Ethical Standards

Conflict of Interest

The authors declare that they have no conflict of interest.

Supplementary material

10278_2017_21_MOESM1_ESM.docx (16 kb)
ESM 1 (DOCX 15 kb)

References

  1. 1.
    Boland GW, Duszak, Jr R, Kalra M: Protocol design and optimization. Journal of the American College of Radiology. 11(5):440–441, 2014.  https://doi.org/10.1016/j.jacr.2014.01.021 CrossRefPubMedGoogle Scholar
  2. 2.
    Ginat DT, Uppuluri P, Christoforidis G, Katzman G, Lee S-K: Identification of neuroradiology MRI protocol errors via a quality-driven categorization approach. J Am Coll Radiol. 13(5):545–548, 2016.  https://doi.org/10.1016/j.jacr.2015.08.027 CrossRefPubMedGoogle Scholar
  3. 3.
    Bairstow PJ, Persaud J, Mendelson R, Nguyen L: Reducing inappropriate diagnostic practice through education and decision support. International Journal for Quality in Health Care. 22(3):194–200, 2010.  https://doi.org/10.1093/intqhc/mzq016 CrossRefPubMedGoogle Scholar
  4. 4.
    Garg AX, Adhikari NKJ, McDonald H et al.: Effects of computerized clinical decision support systems on practitioner performance and patient outcomes. JAMA. 293(10):1223, 2005.  https://doi.org/10.1001/jama.293.10.1223 CrossRefPubMedGoogle Scholar
  5. 5.
    Blackmore CC, Castro A: Improving the quality of imaging in the emergency department. Acad Emerg Med 22(12):1385–1392, 2015  https://doi.org/10.1111/acem.12816
  6. 6.
    Kim, Yoon: Convolutional Neural Networks for Sentence Classification. Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2014.  https://doi.org/10.3115/v1/d14-1181
  7. 7.
    Pons E, Braun LMM, Hunink MGM, Kors JA: Natural language processing in radiology: a systematic review. Radiology. 279(2):329–343, 2016.  https://doi.org/10.1148/radiol.16142770 CrossRefPubMedGoogle Scholar
  8. 8.
    Hassanpour S, Bay G, Langlotz CP: Characterization of change and significance for clinical findings in radiology reports through natural language processing. J Digit Imaging 30(3):314-322, 2017.  https://doi.org/10.1007/s10278-016-9931-8
  9. 9.
    Huang M-W, Chen C-W, Lin W-C, Ke S-W, Tsai C-F: SVM and SVM ensembles in breast cancer prediction. PLOS ONE. 12(1):e0161501, 2017.  https://doi.org/10.1371/journal.pone.0161501 CrossRefPubMedPubMedCentralGoogle Scholar
  10. 10.
    Lakhani P, Langlotz CP: Automated detection of radiology reports that document non-routine communication of critical or significant results. J Digit Imaging. 23(6):647–657, 2010.  https://doi.org/10.1007/s10278-009-9237-1 CrossRefPubMedGoogle Scholar
  11. 11.
    Hassanpour S, Langlotz CP: Information extraction from multi-institutional radiology reports. Artif Intell Med. 66:29–39, 2016.  https://doi.org/10.1016/j.artmed.2015.09.007 CrossRefPubMedGoogle Scholar
  12. 12.
    Cheng LTE, Zheng J, Savova GK, Erickson BJ: Discerning tumor status from unstructured MRI reports-completeness of information in existing reports and utility of automated natural language processing. J Digit Imaging. 23(2):119–132, 2010.  https://doi.org/10.1007/s10278-009-9215-7 CrossRefPubMedGoogle Scholar
  13. 13.
    Meystre SM, Savova GK, Kipper-Schuler KC, Hurdle JF: Extracting information from textual documents in the electronic health record: a review of recent research. Yearb Med Inform. 35(8):128–144, 2008.Google Scholar
  14. 14.
    Chapman WW, Bridewell W, Hanbury P, Cooper GF, Buchanan BG: A simple algorithm for identifying negated findings and diseases in discharge summaries. J Biomed Inform. 34(5):301–310, 2001.  https://doi.org/10.1006/jbin.2001.1029 CrossRefPubMedGoogle Scholar
  15. 15.
    LeCun Y, Bengio Y, Hinton G: Deep learning. Nature. 521(7553):436–444, 2015.  https://doi.org/10.1038/nature14539 CrossRefPubMedGoogle Scholar
  16. 16.
    Ferrucci D, Levas A, Bagchi S, Gondek D, Mueller ET: Watson: beyond jeopardy! Artificial Intelligence. 199:93–105, 2013.  https://doi.org/10.1016/j.artint.2012.06.009 CrossRefGoogle Scholar
  17. 17.
    R Core Team (2017). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL: https://www.R-project.org/
  18. 18.
    Jurka TP, Collingwood L, Boydstun AE, Grossman E, van Atteveldt W: RTextTools: a supervised learning package for text classification. R Journal. 5(1):6–12, 2013Google Scholar
  19. 19.
    Jurka T: MAXENT: an R package for low-memory multinomial logistic regression with support for semi-automated text classification. R J 4(1):56, 2012Google Scholar
  20. 20.
    Friedman J, Hastie T, Tibshirani R: Regularization paths for generalized linear models via coordinate descent. J Stat Soft 33(1):1–968, 2010.  https://doi.org/10.1109/TPAMI.2005.127
  21. 21.
    Liaw A, Wiener M: Classification and regression by randomForest. R News 2(1):18, 2002Google Scholar
  22. 22.
    Feinerer I, Hornik K, Meyer D: Text Mining Infrastructure in R. J Stat Softw 25(5):1–54, 2008.Google Scholar
  23. 23.
    Peters A, Hothorn T, Lausen B: Ipred: improved predictors. R News, 2002. Available at https://cran.r-project.org/web/packages/ipred/vignettes/ipred-examples.pdf. Accessed 12 Sept 2017
  24. 24.
    Tuszynski J.: caTools: tools: moving window statistics, GIF, Base64, ROC AUC, Etc. R package version, 2008. Available at https://cran.r-project.org/web/packages/caTools/caTools.pdf. Accessed 12 Sept 2017
  25. 25.
    Ripley B.: Classification and regression trees. Available at https://cran.r-project.org/web/packages/tree/tree.pdf. Accessed 13 Sept 2017
  26. 26.
    Feinerer I, Hornik K, Meyer D: Text mining infrastructure in R. Journal of Statistical Software 25(5):1–54, 2008.  10.18637/jss.v025.i05 CrossRefGoogle Scholar
  27. 27.
    Ferrucci DA: Introduction to “this is Watson”. IBM Journal of Research and Development 56(3.4):1:1–1:15, 2012.  https://doi.org/10.1147/JRD.2012.2184356 CrossRefGoogle Scholar

Copyright information

© Society for Imaging Informatics in Medicine 2017

Authors and Affiliations

  1. 1.Radiology & Biomedical ImagingUCSF Medical CenterSan FranciscoUSA

Personalised recommendations