Deep learning for liver tumor diagnosis part II: convolutional neural network interpretation using radiologic imaging features.

Objectives: To develop a proof-of-concept "interpretable" deep learning prototype that justifies aspects of its predictions from a pre-trained hepatic lesion classifier.Methods: A convolutional neural network (CNN) was engineered and trained to classify six hepatic tumor entities using 494 lesions o...

Full description

Bibliographic Details
Published in:European Radiology Vol. 29; no. 7; pp. 3348 - 3358
Main Authors: Wang, Clinton J., Hamm, Charlie A., Savic, Lynn J., Ferrante, Marc, Schobert, Isabel, Schlachter, Todd, Lin, MingDe, Weinreb, Jeffrey C., Duncan, James S., Chapiro, Julius, Letzen, Brian
Format: Journal Article
Published: Springer Nature Jul2019
Online Access:View this record in EBSCOhost
fields @attributes:
  recordID: 1
pdfLink:
plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=136842447&site=ehost-live
header:
  @attributes:
    shortDbName: ccm
    uiTerm: 136842447
    longDbName: CINAHL Complete
    uiTag: AN
  controlInfo:
    bkinfo:
    dissinfo:
    jinfo:
      jid:
        09387994
        NPH
      jtl: European Radiology
      issn: 09387994
      maglogo: N
    pubinfo:
      dt: Jul2019
      vid: 29
      iid: 7
      pid: 237
      pub: Springer Nature
      place: New York, New York
    artinfo:
      ui:
        136842447
        136842447
        NLM31093705
        10.1007/s00330-019-06214-8
        NLM31093705
        136842447
      ppf: 3348
      ppct: 10
      formats:
        fmt:
          – @attributes:
              type: T
          – @attributes:
              type: P
      tig:
        atl: Deep learning for liver tumor diagnosis part II: convolutional neural network interpretation using radiologic imaging features.
      aug:
        au:
          Wang, Clinton J.
          Hamm, Charlie A.
          Savic, Lynn J.
          Ferrante, Marc
          Schobert, Isabel
          Schlachter, Todd
          Lin, MingDe
          Weinreb, Jeffrey C.
          Duncan, James S.
          Chapiro, Julius
          Letzen, Brian
        affil: Department of Radiology and Biomedical Imaging, Yale School of Medicine, 333 Cedar Street, 06520, New Haven, CT, USA
      sug:
        subj:
          Liver Neoplasms
          Carcinoma, Hepatocellular
          Neural Networks (Computer)
          Male
          Cholangiocarcinoma
          Predictive Value of Tests
          Algorithms
          Adult
          Aged
          Retrospective Design
          Bile Ducts
          Bile Duct Neoplasms
          Image Interpretation, Computer Assisted Methods
          Female
          Middle Age
          Magnetic Resonance Imaging Methods
          Psychological Tests
          Scales
          Adult: 19-44 years
          Aged: 65+ years
          Middle Aged: 45-64 years
          Male
          Female
      ab: Objectives: To develop a proof-of-concept "interpretable" deep learning prototype that justifies aspects of its predictions from a pre-trained hepatic lesion classifier.Methods: A convolutional neural network (CNN) was engineered and trained to classify six hepatic tumor entities using 494 lesions on multi-phasic MRI, described in Part 1. A subset of each lesion class was labeled with up to four key imaging features per lesion. A post hoc algorithm inferred the presence of these features in a test set of 60 lesions by analyzing activation patterns of the pre-trained CNN model. Feature maps were generated that highlight regions in the original image that correspond to particular features. Additionally, relevance scores were assigned to each identified feature, denoting the relative contribution of a feature to the predicted lesion classification.Results: The interpretable deep learning system achieved 76.5% positive predictive value and 82.9% sensitivity in identifying the correct radiological features present in each test lesion. The model misclassified 12% of lesions. Incorrect features were found more often in misclassified lesions than correctly identified lesions (60.4% vs. 85.6%). Feature maps were consistent with original image voxels contributing to each imaging feature. Feature relevance scores tended to reflect the most prominent imaging criteria for each class.Conclusions: This interpretable deep learning system demonstrates proof of principle for illuminating portions of a pre-trained deep neural network's decision-making, by analyzing inner layers and automatically describing features contributing to predictions.Key Points: • An interpretable deep learning system prototype can explain aspects of its decision-making by identifying relevant imaging features and showing where these features are found on an image, facilitating clinical translation. • By providing feedback on the importance of various radiological features in performing differential diagnosis, interpretable deep learning systems have the potential to interface with standardized reporting systems such as LI-RADS, validating ancillary features and improving clinical practicality. • An interpretable deep learning system could potentially add quantitative data to radiologic reports and serve radiologists with evidence-based decision support.
      pubtype: Academic Journal
      doctype: Journal Article
      ougenre: Article
    language: English
    refInfo:
    holdings:
      @attributes:
        islocal: N