Advertisement

AutoDVT: Joint Real-Time Classification for Vein Compressibility Analysis in Deep Vein Thrombosis Ultrasound Diagnostics

  • Ryutaro Tanno
  • Antonios Makropoulos
  • Salim Arslan
  • Ozan Oktay
  • Sven Mischkewitz
  • Fouad Al-Noor
  • Jonas Oppenheimer
  • Ramin Mandegaran
  • Bernhard Kainz
  • Mattias P. Heinrich
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11071)

Abstract

We propose a dual-task convolutional neural network (CNN) to fully automate the real-time diagnosis of deep vein thrombosis (DVT). DVT can be reliably diagnosed through evaluation of vascular compressibility at anatomically defined landmarks in streams of ultrasound (US) images. The combined real-time evaluation of these tasks has never been achieved before. As proof-of-concept, we evaluate our approach on two selected landmarks of the femoral vein, which can be identified with high accuracy by our approach. Our CNN is able to identify if a vein fully compresses with a F1 score of more than 90% while applying manual pressure with the ultrasound probe. Fully compressible veins robustly rule out DVT and such patients do not need to be referred to further specialist examination. We have evaluated our method on 1150 5–10 s compression image sequences from 115 healthy volunteers, which results in a data set size of approximately 200k labelled images. Our method yields a theoretical inference frame rate of more than 500 fps and we thoroughly evaluate the performance of 15 possible configurations.

References

  1. 1.
    Beckman, M.G., et al.: Venous thromboembolism: a public health concern. Am. J. Prev. Med. 38(4, Supplement), S495–S501 (2010)CrossRefGoogle Scholar
  2. 2.
    Stein, P.D., et al.: D-dimer for the exclusion of acute venous thrombosis and pulmonary embolism: a systematic review. Ann. Intern. Med. 140(8), 589–602 (2004)CrossRefGoogle Scholar
  3. 3.
    Schellong, S.M., et al.: Complete compression ultrasonography of the leg veins as a single test for the diagnosis of deep vein thrombosis. Thromb. Haemost. 89(2), 228–234 (2003)CrossRefGoogle Scholar
  4. 4.
    Guerrero, J., et al.: System for deep venous thrombosis detection using objective compression measures. IEEE Trans. Biomed. Eng. 53(5), 845–854 (2006)CrossRefGoogle Scholar
  5. 5.
    Friedland, N., Adam, D.: Automatic ventricular cavity boundary detection from sequential ultrasound images using simulated annealing. IEEE Trans. Med. Imaging 8(4), 344–353 (1989)CrossRefGoogle Scholar
  6. 6.
    Guerrero, J., et al.: Deep venous thrombosis identification from analysis of ultrasound data. Int. J. Comput. Assist. Radiol. Surg. 10(12), 1963–1971 (2015)CrossRefGoogle Scholar
  7. 7.
    Baumgartner, C.F., et al.: SonoNet: real-time detection and localisation of fetal standard scan planes in freehand ultrasound. IEEE Trans. Med. Imaging 36(11), 2204–2215 (2017)CrossRefGoogle Scholar
  8. 8.
    Prevost, R., Salehi, M., Sprung, J., Bauer, R., Wein, W.: Deep learning for sensorless 3D freehand ultrasound imaging. In: Descoteaux, M., Maier-Hein, L., Franz, A., Jannin, P., Collins, D.L., Duchesne, S. (eds.) MICCAI 2017. LNCS, vol. 10434, pp. 628–636. Springer, Cham (2017).  https://doi.org/10.1007/978-3-319-66185-8_71CrossRefGoogle Scholar
  9. 9.
    Smistad, E., Løvstakken, L.: Vessel detection in ultrasound images using deep convolutional neural networks. In: Carneiro, G. (ed.) LABELS/DLMIA-2016. LNCS, vol. 10008, pp. 30–38. Springer, Cham (2016).  https://doi.org/10.1007/978-3-319-46976-8_4CrossRefGoogle Scholar
  10. 10.
    Lensing, A., et al.: A comparison of compression ultrasound with color doppler ultrasound for the diagnosis of symptomless postoperative deep vein thrombosis. Arch. Intern. Med. 157(7), 765–768 (1997)CrossRefGoogle Scholar
  11. 11.
    Caruana, R.: Multitask learning: a knowledge-based source of inductive bias. In: International Conference on Machine Learning (ICML) (1993)Google Scholar
  12. 12.
    Selvaraju, R.R., et al.: Grad-Cam: visual explanations from deep networks via gradient-based localization. CoRR abs/1610.02391 v3 7(8) (2016)Google Scholar
  13. 13.
    Crimi, A., et al.: Automatic measurement of venous pressure using B-mode ultrasound. IEEE Trans. Biomed. Eng. 63(2), 288–299 (2016)CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  • Ryutaro Tanno
    • 1
    • 2
  • Antonios Makropoulos
    • 1
  • Salim Arslan
    • 1
    • 3
  • Ozan Oktay
    • 1
    • 3
  • Sven Mischkewitz
    • 1
  • Fouad Al-Noor
    • 1
  • Jonas Oppenheimer
    • 1
  • Ramin Mandegaran
    • 1
    • 4
  • Bernhard Kainz
    • 1
    • 3
  • Mattias P. Heinrich
    • 1
    • 5
  1. 1.ThinkSono Ltd.LondonUK
  2. 2.Department of Computer ScienceUniversity College LondonLondonUK
  3. 3.Department of ComputingImperial College LondonLondonUK
  4. 4.Department of RadiologyGuy’s and St Thomas’ NHS Foundation TrustLondonUK
  5. 5.Institute of Medical InformaticsUniversity of LübeckLübeckGermany

Personalised recommendations