A simple PubMed search using the term “radiomics” on February 28, 2023, reveals 7857 publications at the time of the writing, with no time filter applied. Despite this level of research activity, there is a substantial gap between the number of radiomics-based publications and their actual clinical use [1]. A typical radiomics workflow consists of numerous steps, each of which may be influenced by a variety of factors [2]. This ultimately leads to variability that has a significant impact on reproducibility, and in turn clinical translation, resulting in the well-known “reproducibility crisis” in radiomics [3].

According to the same PubMed search, European Radiology (435/7,857; 5.5%) is one of the leading journals in terms of the number of radiomics-related publications. Considering all journal categories, it ranks second after Frontiers in Oncology (729/7,857; 9.3%) and is followed by Cancers (Basel) (309/7,857; 3.9%).

In order to improve the quality of radiomics publications, the European Radiology editorial board members in the Imaging Informatics and Artificial Intelligence section propose 13 consensus recommendations for future radiomics submissions, which relate to design, data, radiomics methodology, metrics, and reporting of research (Table 1). These are listed briefly below for the authors, but reviewers and readers may also find them helpful.