Skip to main content

Prediction of Occult Invasive Disease in Ductal Carcinoma in Situ Using Deep Learning Features.

Publication ,  Journal Article
Shi, B; Grimm, LJ; Mazurowski, MA; Baker, JA; Marks, JR; King, LM; Maley, CC; Hwang, ES; Lo, JY
Published in: J Am Coll Radiol
March 2018

PURPOSE: The aim of this study was to determine whether deep features extracted from digital mammograms using a pretrained deep convolutional neural network are prognostic of occult invasive disease for patients with ductal carcinoma in situ (DCIS) on core needle biopsy. METHODS: In this retrospective study, digital mammographic magnification views were collected for 99 subjects with DCIS at biopsy, 25 of which were subsequently upstaged to invasive cancer. A deep convolutional neural network model that was pretrained on nonmedical images (eg, animals, plants, instruments) was used as the feature extractor. Through a statistical pooling strategy, deep features were extracted at different levels of convolutional layers from the lesion areas, without sacrificing the original resolution or distorting the underlying topology. A multivariate classifier was then trained to predict which tumors contain occult invasive disease. This was compared with the performance of traditional "handcrafted" computer vision (CV) features previously developed specifically to assess mammographic calcifications. The generalization performance was assessed using Monte Carlo cross-validation and receiver operating characteristic curve analysis. RESULTS: Deep features were able to distinguish DCIS with occult invasion from pure DCIS, with an area under the receiver operating characteristic curve of 0.70 (95% confidence interval, 0.68-0.73). This performance was comparable with the handcrafted CV features (area under the curve = 0.68; 95% confidence interval, 0.66-0.71) that were designed with prior domain knowledge. CONCLUSIONS: Despite being pretrained on only nonmedical images, the deep features extracted from digital mammograms demonstrated comparable performance with handcrafted CV features for the challenging task of predicting DCIS upstaging.

Duke Scholars

Altmetric Attention Stats
Dimensions Citation Stats

Published In

J Am Coll Radiol

DOI

EISSN

1558-349X

Publication Date

March 2018

Volume

15

Issue

3 Pt B

Start / End Page

527 / 534

Location

United States

Related Subject Headings

  • Retrospective Studies
  • Prognosis
  • Nuclear Medicine & Medical Imaging
  • Neural Networks, Computer
  • Neoplasm Staging
  • Monte Carlo Method
  • Middle Aged
  • Mammography
  • Humans
  • Female
 

Citation

APA
Chicago
ICMJE
MLA
NLM
Shi, B., Grimm, L. J., Mazurowski, M. A., Baker, J. A., Marks, J. R., King, L. M., … Lo, J. Y. (2018). Prediction of Occult Invasive Disease in Ductal Carcinoma in Situ Using Deep Learning Features. J Am Coll Radiol, 15(3 Pt B), 527–534. https://doi.org/10.1016/j.jacr.2017.11.036
Shi, Bibo, Lars J. Grimm, Maciej A. Mazurowski, Jay A. Baker, Jeffrey R. Marks, Lorraine M. King, Carlo C. Maley, E Shelley Hwang, and Joseph Y. Lo. “Prediction of Occult Invasive Disease in Ductal Carcinoma in Situ Using Deep Learning Features.J Am Coll Radiol 15, no. 3 Pt B (March 2018): 527–34. https://doi.org/10.1016/j.jacr.2017.11.036.
Shi B, Grimm LJ, Mazurowski MA, Baker JA, Marks JR, King LM, et al. Prediction of Occult Invasive Disease in Ductal Carcinoma in Situ Using Deep Learning Features. J Am Coll Radiol. 2018 Mar;15(3 Pt B):527–34.
Shi, Bibo, et al. “Prediction of Occult Invasive Disease in Ductal Carcinoma in Situ Using Deep Learning Features.J Am Coll Radiol, vol. 15, no. 3 Pt B, Mar. 2018, pp. 527–34. Pubmed, doi:10.1016/j.jacr.2017.11.036.
Shi B, Grimm LJ, Mazurowski MA, Baker JA, Marks JR, King LM, Maley CC, Hwang ES, Lo JY. Prediction of Occult Invasive Disease in Ductal Carcinoma in Situ Using Deep Learning Features. J Am Coll Radiol. 2018 Mar;15(3 Pt B):527–534.

Published In

J Am Coll Radiol

DOI

EISSN

1558-349X

Publication Date

March 2018

Volume

15

Issue

3 Pt B

Start / End Page

527 / 534

Location

United States

Related Subject Headings

  • Retrospective Studies
  • Prognosis
  • Nuclear Medicine & Medical Imaging
  • Neural Networks, Computer
  • Neoplasm Staging
  • Monte Carlo Method
  • Middle Aged
  • Mammography
  • Humans
  • Female