Incorporating radiomics into clinical trials: expert consensus endorsed by the European Society of Radiology on considerations for data-driven compared to biologically driven quantitative biomarkers

Autor: Anna Caroli, James P B O'Connor, Xavier Golay, Nicolas Michoux, Nathalie Lassau, Marc Dewey, Marius E. Mayerhoefer, Laure Fournier, Daniel C. Sullivan, Rik Achten, Olivier Clément, Edwin H.G. Oei, Karen Rosendahl, Lena Costaridou, Egesta Lopci, Aad van der Lugt, Christian Loewe, Anders Persson, Nandita M. deSouza, Ronald Boellaard, Wolfgang G. Kunz, Manuela França, Lioe-Fee de Geus-Oei, Rashindra Manniesing, Christophe Deroose, Daniela E. Oprea-Lager, Marion Smits, Angel Alberich-Bayarri, Luc Bidaut, Nancy A. Obuchowski, Caroline Caramella, Frédéric Lecouvet, Elmar Kotter
Přispěvatelé: Hôpital Européen Georges Pompidou [APHP] (HEGP), Assistance publique - Hôpitaux de Paris (AP-HP) (AP-HP)-Hôpitaux Universitaires Paris Ouest - Hôpitaux Universitaires Île de France Ouest (HUPO), INSERM, Paris Cardiovasc Res Ctr PARCC UMR970, F-75015 Paris, France, Partenaires INRAE, Université de Paris - UFR Médecine Paris Centre [Santé] (UP Médecine Paris Centre), Université de Paris (UP), ANR-19-P3IA-0001,PRAIRIE,PaRis Artificial Intelligence Research InstitutE(2019), UCL - SSS/IREC/IMAG - Pôle d'imagerie médicale, UCL - (SLuc) Service de radiologie, Radiology & Nuclear Medicine
Jazyk: angličtina
Rok vydání: 2021
Předmět:
medicine.medical_specialty
Consensus
Standardization
Process (engineering)
Vascular damage Radboud Institute for Health Sciences [Radboudumc 16]
Feature selection
Overfitting
Machine learning
computer.software_genre
030218 nuclear medicine & medical imaging
Data-driven
03 medical and health sciences
0302 clinical medicine
Image Processing
Computer-Assisted
medicine
[INFO.INFO-IM]Computer Science [cs]/Medical Imaging
Humans
Radiology
Nuclear Medicine and imaging
ComputingMilieux_MISCELLANEOUS
business.industry
Radiology
Statistics and numerical data
Validation studies
Clinical trial
General Medicine
Women's cancers Radboud Institute for Health Sciences [Radboudumc 17]
Imaging Informatics and Artificial Intelligence
Sample size determination
Feature (computer vision)
030220 oncology & carcinogenesis
Artificial intelligence
Radiologi och bildbehandling
Tomography
X-Ray Computed
business
computer
Biomarkers
Radiology
Nuclear Medicine and Medical Imaging
Zdroj: European Radiology
European Radiology, Springer Verlag, 2021, 31 (8), pp.6001-6012. ⟨10.1007/s00330-020-07598-8⟩
European Radiology, 31, 6001-6012
European radiology, Vol. 31, no.8, p. 6001-6012 (2021)
European Radiology, 31(8), 6001-6012. Springer-Verlag
European Radiology, 31, 8, pp. 6001-6012
ISSN: 0938-7994
1432-1084
DOI: 10.1007/s00330-020-07598-8⟩
Popis: Abstract Existing quantitative imaging biomarkers (QIBs) are associated with known biological tissue characteristics and follow a well-understood path of technical, biological and clinical validation before incorporation into clinical trials. In radiomics, novel data-driven processes extract numerous visually imperceptible statistical features from the imaging data with no a priori assumptions on their correlation with biological processes. The selection of relevant features (radiomic signature) and incorporation into clinical trials therefore requires additional considerations to ensure meaningful imaging endpoints. Also, the number of radiomic features tested means that power calculations would result in sample sizes impossible to achieve within clinical trials. This article examines how the process of standardising and validating data-driven imaging biomarkers differs from those based on biological associations. Radiomic signatures are best developed initially on datasets that represent diversity of acquisition protocols as well as diversity of disease and of normal findings, rather than within clinical trials with standardised and optimised protocols as this would risk the selection of radiomic features being linked to the imaging process rather than the pathology. Normalisation through discretisation and feature harmonisation are essential pre-processing steps. Biological correlation may be performed after the technical and clinical validity of a radiomic signature is established, but is not mandatory. Feature selection may be part of discovery within a radiomics-specific trial or represent exploratory endpoints within an established trial; a previously validated radiomic signature may even be used as a primary/secondary endpoint, particularly if associations are demonstrated with specific biological processes and pathways being targeted within clinical trials. Key Points • Data-driven processes like radiomics risk false discoveries due to high-dimensionality of the dataset compared to sample size, making adequate diversity of the data, cross-validation and external validation essential to mitigate the risks of spurious associations and overfitting. • Use of radiomic signatures within clinical trials requires multistep standardisation of image acquisition, image analysis and data mining processes. • Biological correlation may be established after clinical validation but is not mandatory.
Databáze: OpenAIRE
Externí odkaz: