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

• DOI: 10.1007/s00330-020-07598-8
• Publication Date: 2021-01-25T11:06:46Z
• Authors: Laure Fournier; Lena Costaridou; Luc Bidaut; Nicolas Michoux; Frederic E. Lecouvet; Lioe-Fee de Geus-Oei; Ronald Boellaard; Daniela E. Oprea-...; Nancy A Obuchowski; Anna Caroli; Wolfgang G. Kunz; Edwin H. Oei; James P. B. O’Connor; Marius E. Mayerho...; Manuela Franca; Angel Alberich-Ba...; Christophe M. Der...; Christian Loewe; Rashindra Manniesing; Caroline Caramella; Egesta Lopci; Nathalie Lassau; Anders Persson; Rik Achten; Karen Rosendahl; Olivier Clement; Elmar Kotter; Xavier Golay; Marion Smits; Marc Dewey; Daniel C. Sullivan; Aad van der Lugt; Nandita M. deSouza; European Society...
• Keywords: Biological data

Abstract Existing quantitative imaging biomarkers (QIBs) are associated with known biological tissue characteristics and follow a well-understood path of technical, clinical validation before incorporation into trials. In radiomics, novel data-driven processes extract numerous visually imperceptible statistical features from the data no priori assumptions on their correlation processes. The selection relevant (radiomic signature) trials therefore requires additional considerations to ensure meaningful endpoints. Also, number radiomic tested means that power calculations would result in sample sizes impossible achieve within This article examines how process standardising validating differs those based associations. Radiomic signatures best developed initially datasets represent diversity acquisition protocols as well disease normal findings, rather than standardised optimised this risk being linked pathology. Normalisation through discretisation feature harmonisation essential pre-processing steps. Biological may be performed after technical validity signature is established, but not mandatory. Feature part discovery radiomics-specific trial or exploratory endpoints an established trial; previously validated even used primary/secondary endpoint, particularly if associations demonstrated specific pathways targeted Key Points • Data-driven like radiomics false discoveries due high-dimensionality dataset compared size, making adequate data, cross-validation external mitigate risks spurious overfitting. Use multistep standardisation image acquisition, analysis mining