The Role of AI in Clinical Trials

  • Irene Mayorga-RuizEmail author
  • Ana Jiménez-Pastor
  • Belén Fos-Guarinos
  • Rafael López-González
  • Fabio García-Castro
  • Ángel Alberich-Bayarri


Medical imaging is increasingly being used in clinical trials. With the introduction of imaging in the clinical trial environment, AI tools and imaging biomarker computation are consequently introduced in order to reduce times in the radiological reading and add objectivity to the evaluation of new treatment response. In order to integrate artificial intelligence (AI) techniques and imaging biomarker (IB) analysis pipelines in clinical trials and improve quality and accuracy in the conclusions of the study, medical image acquisition should be harmonized and standardized across imaging centers. Since all the imaging biomarkers to be extracted from the images rely on the image quality, attention should be given to the design of the image acquisition protocols followed by the theoretical and technical validation of the site. Site’s validation should be performed by the acquisition of dummy run studies to evaluate equipment performance and by a cross-calibration of the different acquisition equipment involved in the trial, since there is a growing trend to integrate quantitative measures in clinical trials beyond lesion diameter. Artificial intelligence and its implementation through machine learning, and specifically deep learning techniques, bring many benefits to medical image processing by allowing to automatize tasks that in the past were performed manually, like organ, tissue, and structure segmentation. Also, the use of this automatic AI tools minimizes the human interaction, reducing the human-induced variability that may bias the results, decreasing the number of patients needed by the increased statistical power, and therefore accelerating the time-to-market of new molecules.


Image standardization Image storage Central reading Core laboratory Treatment response Real-time clinical trial Cross-calibration Imaging biomarkers Automatic segmentation 


  1. 1.
    FDA Guidance for Industry. Clinical trials imaging endpoints process standard. 2018. Accessed 24 May 2018.
  2. 2.
  3. 3.
    NCI. NCI-CQIE qualification materials. Accessed 12 May 2018.
  4. 4.
    Food and Drug Administration. FDA official web. Accessed 10 May 2018.
  5. 5.
    European Medicine Agency. EMA official web. Accessed 10 May 2018.
  6. 6.
    Shin HC, Roth HR, Gao M, et al. Deep convolutional neural networks for computer-aided detection: CNN architectures, dataset characteristics and transfer learning. IEEE Trans Med Imaging. 35(5):1285–98.CrossRefGoogle Scholar
  7. 7.
    Perez L, Wang J. The effectiveness of data augmentation in image classification using deep learning. 2017.Google Scholar
  8. 8.
    DeVries T, Taylor GW. Dataset augmentation in feature space, 2017. arXiv:1702.05538v1. Google Scholar
  9. 9.
    Cheatsheet ML. Official web. Accessed 17 Jul 18.
  10. 10.
    França M, Alberich-Bayarri Á, Martí-Bonmatí L, et al. Accurate simultaneous quantification of liver steatosis and iron overload in diffuse liver diseases with MRI. Abdom Radiol. 2017;42:1434–43.CrossRefGoogle Scholar
  11. 11.
    Bruynseels K, Santoni de Sio F, van den Hoven J. Digital twins in health care: ethical implications of an emerging engineering paradigm. Front Genet. 2018;9:31.CrossRefGoogle Scholar
  12. 12.
    Viceconti M, Henney A, Morley-Fletcher E, editors. In silico clinical trials: how computer simulation will transform the biomedical industry. Brussels: Avicenna Consortium; 2016.Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Irene Mayorga-Ruiz
    • 1
    Email author
  • Ana Jiménez-Pastor
    • 1
  • Belén Fos-Guarinos
    • 1
  • Rafael López-González
    • 1
  • Fabio García-Castro
    • 1
  • Ángel Alberich-Bayarri
    • 1
  1. 1.Quantitative Imaging Biomarkers in MedicineValenciaSpain

Personalised recommendations