Rofo 2021; 193(03): 252-261
DOI: 10.1055/a-1248-2556
Review

Deep Learning CT Image Reconstruction in Clinical Practice

CT-Bildrekonstruktion mit Deep Learning in der klinischen Praxis
Clemens Arndt
Department of Radiology, Jena University Hospital, Jena, Germany
,
Felix Güttler
Department of Radiology, Jena University Hospital, Jena, Germany
,
Andreas Heinrich
Department of Radiology, Jena University Hospital, Jena, Germany
,
Florian Bürckenmeyer
Department of Radiology, Jena University Hospital, Jena, Germany
,
Ioannis Diamantis
Department of Radiology, Jena University Hospital, Jena, Germany
,
Ulf Teichgräber
Department of Radiology, Jena University Hospital, Jena, Germany
› Author Affiliations

Abstract

Background Computed tomography (CT) is a central modality in modern radiology contributing to diagnostic medicine in almost every medical subspecialty, but particularly in emergency services. To solve the inverse problem of reconstructing anatomical slice images from the raw output the scanner measures, several methods have been developed, with filtered back projection (FBP) and iterative reconstruction (IR) subsequently providing criterion standards. Currently there are new approaches to reconstruction in the field of artificial intelligence utilizing the upcoming possibilities of machine learning (ML), or more specifically, deep learning (DL).

Method This review covers the principles of present CT image reconstruction as well as the basic concepts of DL and its implementation in reconstruction. Subsequently commercially available algorithms and current limitations are being discussed.

Results and Conclusion DL is an ML method that utilizes a trained artificial neural network to solve specific problems. Currently two vendors are providing DL image reconstruction algorithms for the clinical routine. For these algorithms, a decrease in image noise and an increase in overall image quality that could potentially facilitate the diagnostic confidence in lesion conspicuity or may translate to dose reduction for given clinical tasks have been shown. One study showed equal diagnostic accuracy in the detection of coronary artery stenosis for DL reconstructed images compared to IR at higher image quality levels. Consequently, a lot more research is necessary and should aim at diagnostic superiority in the clinical context covering a broadness of pathologies to demonstrate the reliability of such DL approaches.

Key Points:

  • Following iterative reconstruction, there is a new approach to CT image reconstruction in the clinical routine using deep learning (DL) as a method of artificial intelligence.

  • DL image reconstruction algorithms decrease image noise, improve image quality, and have potential to reduce radiation dose.

  • Diagnostic superiority in the clinical context should be demonstrated in future trials.

Citation Format

  • Arndt C, Güttler F, Heinrich A et al. Deep Learning CT Image Reconstruction in Clinical Practice. Fortschr Röntgenstr 2021; 193: 252 – 261

Zusammenfassung

Hintergrund Die Computertomografie (CT) ist eine zentrale Modalität der modernen Radiologie, die in nahezu allen medizinischen Fachdisziplinen, insbesondere aber in der Notfallmedizin, einen wichtigen Bestandteil zur Gesundheitsversorgung liefert. Die Berechnung bzw. Rekonstruktion von Schnittbildern aus den rohen Messwerten der CT-Untersuchungen stellt mathematisch ein inverses Problem dar. Bisher waren die gefilterte Rückprojektion und die iterative Rekonstruktion (IR) die Goldstandards, um schnell und zuverlässig Bilder zu berechnen. Aktuell gibt es im Rahmen neuerer Entwicklungen im Bereich der künstlichen Intelligenz mit dem Deep Learning (DL) einen weiteren Ansatzweg in der klinischen Routine.

Methode Dieser Übersichtsartikel erläutert die bisherigen Prinzipien der Bildrekonstruktion, das Konzept des DL und das Anwendungsprinzip zur Rekonstruktion. Anschließend werden kommerziell verfügbare Algorithmen und bisherige Studien diskutiert und Grenzen sowie Probleme dargestellt.

Ergebnis und Schlussfolgerung DL als Methode des Machine Learning nutzt im Allgemeinen ein trainiertes, künstliches, neuronales Netzwerk zur Lösung von Problemen. Aktuell sind DL-Rekonstruktionsalgorithmen von 2 Herstellern für die klinische Routine verfügbar. In den bisherigen Studien konnte eine Reduktion des Bildrauschens und eine Verbesserung der Gesamtqualität gezeigt werden. Eine Studie zeigte, bei höherer Qualität der mittels DL rekonstruierten Bilder, eine vergleichbare diagnostische Genauigkeit zur IR für die Detektion von Koronarstenosen. Weitere Studien sind notwendig und sollten vor allem auch auf eine klinische Überlegenheit abzielen, während dabei ein breiter Umfang an Pathologien umfasst werden muss.

Kernaussagen:

  • Nach der aktuell verbreiteten iterativen Rekonstruktion können CT-Schnittbilder im klinischen Alltag nun auch mittels Deep Learning (DL) als Methode der künstlichen Intelligenz rekonstruiert werden.

  • Durch DL-Rekonstruktionen können Bildrauschen vermindert, Bildqualität verbessert und möglicherweise Strahlung reduziert werden.

  • Eine diagnostische Überlegenheit im klinischen Kontext sollte in kommenden Studien gezeigt werden.



Publication History

Received: 11 May 2020

Accepted: 20 August 2020

Article published online:
10 December 2020

© 2020. Thieme. All rights reserved.

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany

 
  • References

  • 1 Goldman LW. Principles of CT: radiation dose and image quality. J Nucl Med Technol 2007; 35: 213-225 ; quiz 226–228
  • 2 Primak AN, McCollough CH, Bruesewitz MR. et al. Relationship between Noise, Dose, and Pitch in Cardiac Multi–Detector Row CT. RadioGraphics 2006; 26: 1785-1794
  • 3 Alkadhi H, Leschka S, Stolzmann P. et al Wie funktioniert CT?. Berlin, Heidelberg: Springer Berlin Heidelberg; 2011. Im Internet: http://link.springer.com/10.1007/978-3-642-17803-0
  • 4 Geyer LL, Schoepf UJ, Meinel FG. et al. State of the Art: Iterative CT Reconstruction Techniques. Radiology 2015; 276: 339-357
  • 5 Fleischmann D, Boas FE. Computed tomography – old ideas and new technology. Eur Radiol 2011; 21: 510-517
  • 6 Feldkamp LA, Davis LC, Kress JW. Practical cone-beam algorithm. J Opt Soc Am A 1984; 1: 612
  • 7 Desai GS, Uppot RN, Yu EW. et al. Impact of iterative reconstruction on image quality and radiation dose in multidetector CT of large body size adults. Eur Radiol 2012; 22: 1631-1640
  • 8 Noël PB, Walczak AM, Xu J. et al. GPU-based cone beam computed tomography. Computer Methods and Programs in Biomedicine 2010; 98: 271-277
  • 9 Gordon R, Bender R, Herman GT. Algebraic Reconstruction Techniques (ART) for three-dimensional electron microscopy and X-ray photography. Journal of Theoretical Biology 1970; 29: 471-481
  • 10 Beister M, Kolditz D, Kalender WA. Iterative reconstruction methods in X-ray CT. Physica Medica 2012; 28: 94-108
  • 11 den Harder AM, Willemink MJ, de Ruiter QMB. et al. Achievable dose reduction using iterative reconstruction for chest computed tomography: A systematic review. European Journal of Radiology 2015; 84: 2307-2313
  • 12 Kohli M, Prevedello LM, Filice RW. et al. Implementing Machine Learning in Radiology Practice and Research. American Journal of Roentgenology 2017; 208: 754-760
  • 13 Erickson BJ, Korfiatis P, Akkus Z. et al. Machine Learning for Medical Imaging. RadioGraphics 2017; 37: 505-515
  • 14 Chartrand G, Cheng PM, Vorontsov E. et al. Deep Learning: A Primer for Radiologists. RadioGraphics 2017; 37: 2113-2131
  • 15 Yamashita R, Nishio M, Do RKG. et al. Convolutional neural networks: an overview and application in radiology. Insights Imaging 2018; 9: 611-629
  • 16 Shan H, Zhang Y, Yang Q. et al. 3-D Convolutional Encoder-Decoder Network for Low-Dose CT via Transfer Learning From a 2-D Trained Network. IEEE Trans Med Imaging 2018; 37: 1522-1534
  • 17 Isola P, Zhu JY, Zhou T. et al Image-to-Image Translation with Conditional Adversarial Networks. arXiv:161107004 [cs] 2018. Im Internet: http://arxiv.org/abs/1611.07004
  • 18 McBee MP, Awan OA, Colucci AT. et al. Deep Learning in Radiology. Academic Radiology 2018; 25: 1472-1480
  • 19 Lakhani P, Sundaram B. Deep Learning at Chest Radiography: Automated Classification of Pulmonary Tuberculosis by Using Convolutional Neural Networks. Radiology 2017; 284: 574-582
  • 20 Wolterink JM, Leiner T, Viergever MA. et al. Generative Adversarial Networks for Noise Reduction in Low-Dose CT. IEEE Trans Med Imaging 2017; 36: 2536-2545
  • 21 Boedecker K. AiCE Deep Learning Reconstruction: Bringing the power of Ultra-High Resolution CT to routine imaging. 2019
  • 22 Hsieh J, Liu E, Nett B. et al. A new era of image reconstruction: TrueFidelityTM. Technical white paper on deep learning image reconstruction. 2019
  • 23 Machine Learning for Medical Image Reconstruction: Second International Workshop, MLMIR 2019, Held in Conjunction with MICCAI 2019, Shenzhen, China, October 17, 2019, Proceedings. Cham: Springer International Publishing 2019. Im Internet: http://link.springer.com/10.1007/978-3-030-33843-5
  • 24 Sprawls P. AAPM tutorial. CT image detail and noise. RadioGraphics 1992; 12: 1041-1046
  • 25 Verdun FR, Racine D, Ott JG. et al. Image quality in CT: From physical measurements to model observers. Physica Medica 2015; 31: 823-843
  • 26 Jin KH, McCann MT, Froustey E. et al. Deep Convolutional Neural Network for Inverse Problems in Imaging. IEEE Trans on Image Process 2017; 26: 4509-4522
  • 27 Umehara K, Ota J, Ishida T. Application of Super-Resolution Convolutional Neural Network for Enhancing Image Resolution in Chest CT. J Digit Imaging 2018; 31: 441-450
  • 28 Park J, Hwang D, Kim KY. et al. Computed tomography super-resolution using deep convolutional neural network. Phys Med Biol 2018; 63: 145011
  • 29 Zhang Y, Yu H. Convolutional Neural Network Based Metal Artifact Reduction in X-ray Computed Tomography. IEEE Trans Med Imaging 2018; 37: 1370-1381
  • 30 Benz DC, Benetos G, Rampidis G. et al. Validation of deep-learning image reconstruction for coronary computed tomography angiography: Impact on noise, image quality and diagnostic accuracy. Journal of Cardiovascular Computed Tomography 2020; DOI: S1934592519304642.
  • 31 Greffier J, Hamard A, Pereira F. et al Image quality and dose reduction opportunity of deep learning image reconstruction algorithm for CT: a phantom study. Eur Radiol 2020. Im Internet: http://link.springer.com/10.1007/s00330-020-06724-w
  • 32 Greffier J, Frandon J, Larbi A. et al. CT iterative reconstruction algorithms: a task-based image quality assessment. Eur Radiol 2020; 30: 487-500
  • 33 Heinrich A, Engler M, Dachoua D. et al. Deep Learning Image Reconstruction in CT: Quantification of Image Quality in Clinical Practice. Manuscript submitted for publication. 2020
  • 34 Akagi M, Nakamura Y, Higaki T. et al. Deep learning reconstruction improves image quality of abdominal ultra-high-resolution CT. Eur Radiol 2019; 29: 6163-6171
  • 35 Nakamura Y, Higaki T, Tatsugami F. et al. Deep Learning–based CT Image Reconstruction: Initial Evaluation Targeting Hypovascular Hepatic Metastases. Radiology: Artificial Intelligence 2019; 1: e180011
  • 36 Higaki T, Nakamura Y, Zhou J. et al. Deep Learning Reconstruction at CT: Phantom Study of the Image Characteristics. Academic Radiology 2020; 27: 82-87
  • 37 Narita K, Nakamura Y, Higaki T. et al Deep learning reconstruction of drip-infusion cholangiography acquired with ultra-high-resolution computed tomography. Abdom Radiol 2020. Im Internet: http://link.springer.com/10.1007/s00261-020-02508-4
  • 38 Hoeschen C. Einsatz künstlicher Intelligenz für die Bildrekonstruktion. Radiologe 2020; 60: 15-23
  • 39 Zhang H, Dong B. A Review on Deep Learning in Medical Image Reconstruction. arXiv:190610643 [physics] 2019. Im Internet: http://arxiv.org/abs/1906.10643
  • 40 Willemink MJ, Persson M, Pourmorteza A. et al. Photon-counting CT: Technical Principles and Clinical Prospects. Radiology 2018; 289: 293-312