Articles

CLEAR, METRICS, and their E3 extensions: a comprehensive framework for evaluating radiomics research

Oct 23, 2025

CLEAR, METRICS, and their E3 extensions: a comprehensive framework for evaluating radiomics research

Authors
Jingyu Zhong1,2, Burak Kocak3, Renato Cuocolo4

Authors’ institutions:
1 Department of Imaging, Tongren Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
2 Shanghai Key Laboratory of Flexible Medical Robotics, Tongren Hospital, Institute of Medical Robotics, Shanghai
Jiao Tong University, Shanghai, China
3 Department of Radiology, Basaksehir Cam and Sakura City Hospital, Istanbul, Turkey
4 Department of Medicine, Surgery and Dentistry, University of Salerno, Baronissi, Italy

 

The CheckList for EvaluAtion of Radiomics research (CLEAR) [1] and METhodological RadiomICs Score (METRICS) [2] are two related yet distinct documents built with international consensus, focusing respectively on reporting transparency and methodological quality of radiomics studies. A guidance that aimed to maximize the utility of CLEAR has also been published, providing Explanation and Elaboration with Examples (CLEAR-E3) [3]. Recently, a similar initiative has been published to promote the effective implementation of the METRICS (METRICS-E3) [4]. Together, these documents now provide a structured framework for evaluating radiomics research. This represents an important foundation, but it should be regarded as a work in progress, open to refinement rather than a completed mission.

In contrast to the numerous radiomics literature, the clinically adopted radiomics tool is extremely rare [5,6]. The roadblocks between radiomics academic papers and clinical practice include technical challenges in reproducibility, standardization and generalizability. The poor quality of radiomics research methodology has been one of the main reasons for the low translational rate of radiomics research [7,8], where tools for systematically assessing study rigor are needed. In 2024, METRICS was therefore established though a Delphi process to provide a quality control guidance, including 30 weighted items within 9 categories, and can be summarized as a percentage for total score categories from very low to excellent [2].

While METRICS targets methodological rigor, an even more fundamental issue is poor reporting. Without sufficient detail on how studies are conducted, methodological quality cannot be fully assessed. CLEAR was developed to address this gap in 2023 before METRICS. CLEAR, as a specialized checklist for compete radiomics study reporting, was thus developed, also with a Delphi process. The complete version of CLEAR consists of 58 items with considerations on both methodological aspect and aspects other than methodology, while the shortened version (CLEAR-S) includes 43 items with specific focus on methodology [1].

These initiatives were created to address the need for transparent reporting and robust methodological evaluation in radiomics research. Initial uptake, reflected in early citation patterns, indicates awareness. However, whether these tools improve study quality or facilitate clinical translation remains uncertain.

CLEAR and METRICS (along with their E3 documents) can be used by authors, editors, reviewers and readers of radiomics studies. For the potential authors of radiomics papers, these documents can guide the whole radiomics workflow from study design to model evaluation. With METRICS applied prior to study, the essential errors in the study methodology may be avoid, such as unreasonable reference standard, insufficient sample size, failure in control of overfitting, and inappropriate evaluation technique. Similarly, with CLEAR, the reporting of radiomics studies would lead higher transparency to allow better detection of these shortages, mutually benefiting each other.

For editors, CLEAR and METRICS may assist the initial assessment of radiomics studies. The standardized evaluation of radiomics studies can be helpful for treating the increasing amount of radiomics submissions to select the well-reported and high-quality ones for peer review, meanwhile telling the authors how they can improve their desk rejected submissions. Given the time constraints in editorial workflows, there may also be potential for large language models to assist with preliminary screening, though this remains exploratory and untested [9,10]. There has been so far no journal include these two documents as a part of submission guideline [11]. With these tools, the editorial process can be smoother with a preliminary quality control step, which may finally improve the study quality of the journal [12].

For the reviewers and readers, CLEAR and METRICS may also serve as tools for structural information extraction. The reviewers can pay more attention to the clinical questions investigated in the study rather than the basic technical issues. The experience of the reviewers can be used more efficiently and wisely, releasing from repeating shared inefficiencies in different submissions. The readers can also rapidly detect the advantages and disadvantages of radiomics studies using these two documents and decide at degree the study is technically reliable.

In all usage scenarios, both CLEAR and METRICS should be applied with care, as early evaluations suggest that inappropriate use may undermine their purpose [13]. As a result, the potential value of the respective explanation and elaboration documents (CLEAR-E3 and METRICS-E3) becomes evident [3,4].

CLEAR and METRICS are expected to be applied as quality assessment tools for evidence block establishment. The attempts for the evidence level rating for the use of radiomics in specific clinical questions are mainly according to quantitative analysis results [5]. The criteria for up- or down-grading the evidence level of radiomics use is lacking. It is hard to introduce radiomics into clinical practice without supporting from evidence block. These tools may play a role in strengthening the evidence base required for potential clinical adoption of radiomics models, though this will also depend on broader validation.

The Chinese translation and interpretation of CLEAR and METRICS has been completed with endorsement of CLEAR and METRICS working groups to enable easier use by Chinese radiomics researchers [14,15]. Translated version of CLEAR and METRICS in various languages are expected to gain the attention form the local radiological communities and promote the acceptance of them. Further, there is a need to assess the influence of the CLEAR and METRICS on the editorial and peer review process and final paper quality [12]. If there is sufficient data demonstrating that the use of CLEAR and METRICS can shorten the editorial assessment process or improve the radiomics paper, they are recommended to be included as a routine for radiomics submission [11]. Importantly, both CLEAR and METRICS are conceived as living documents, open to feedback and iterative refinement as the field evolves. .  All stakeholders should see themselves as contributors to their continuous refinement, ensuring these tools evolve alongside the radiomics community.

In summary, CLEAR and METRICS accompanying with E3 documents (CLEAR-E3 and METRICS-E3) not only provide guidance for radiomics stakeholders alongside the radiomics workflow but also provide a structured pathway for strengthening the quality and transparency of radiomics studies. Their consistent use could help narrow the gap between experimental research and clinically practicable tools.

 

 

References:

  1. Kocak B, Baessler B, Bakas S et al (2023) CheckList for EvaluAtion of Radiomics research (CLEAR): a step-by-step reporting guideline for authors and reviewers endorsed by ESR and EuSoMII. Insights Imaging 14(1):75. https://doi.org/10.1186/s13244-023-01415-8
  2. Kocak B, Akinci D’Antonoli T, Mercaldo N et al (2024) METhodological RadiomICs Score (METRICS): a quality scoring tool for radiomics research endorsed by EuSoMII. Insights Imaging 15(1):8. https://doi.org/10.1186/s13244-023-01572-w
  3. Kocak B, Borgheresi A, Ponsiglione A et al (2024) Explanation and Elaboration with Examples for CLEAR (CLEAR-E3): an EuSoMII Radiomics Auditing Group Initiative. Eur Radiol Exp 8(1):72. https://doi.org/10.1186/s41747-024-00471-z
  4. Kocak B, Ammirabile A, Ambrosini I et al (2025) Explanation and Elaboration with Examples for METRICS (METRICS-E3): an initiative from the EuSoMII Radiomics Auditing Group. Insights Imaging 16(1):175. https://doi.org/10.1186/s13244-025-02061-y
  5. Zhong J, Lu J, Zhang G et al (2023) An overview of meta-analyses on radiomics: more evidence is needed to support clinical translation. Insights Imaging 14:111. https://doi.org/10.1186/s13244-023-01437-2
  6. Kocak B, Pinto dos Santos D, Dietzel M (2025) The widening gap between radiomics research and clinical translation: rethinking current practices and shared responsibilities. Eur J Radiol Artif Intell 1:100004. https://doi.org/10.1016/j.ejrai.2025.100004
  7. Barry N, Kendrick J, Molin K et al (2025) Evaluating the impact of the Radiomics Quality Score: a systematic review and meta-analysis. Eur Radiol 35(3):1701-1713. https://doi.org/10.1007/s00330-024-11341-y
  8. Kocak B, Keles A, Kose F, Sendur A (2025) Quality of radiomics research: comprehensive analysis of 1574 unique publications from 89 reviews. Eur Radiol 35(4):1980-1992. https://doi.org10.1007/s00330-024-11057-z
  9. de Almeida JG, Papanikolaou N (2025) Auto-METRICS: LLM-assisted scientific quality control for radiomics research. Eur J Radiol 191:112358. https://doi.org/10.1016/j.ejrad.2025.112358
  10. Mese I, Kocak B. Large language models in methodological quality evaluation of radiomics research based on METRICS: ChatGPT vs NotebookLM vs radiologist. Eur J Radiol 184:111960. https://doi.org/10.1016/j.ejrad.2025.111960
  11. Zhong J, Xing Y, Lu J et al (2023) The endorsement of general and artificial intelligence reporting guidelines in radiological journals: a meta-research study. BMC Med Res Methodol 23(1):292. https://doi.org/10.1186/s12874-023-02117-x
  12. Stahl AC, Tietz AS, Dewey M, Kendziora B (2023) Has the quality of reporting improved since it became mandatory to use the Standards for Reporting Diagnostic Accuracy? Insights Imaging 14(1):85. https://doi.org/10.1186/s13244-023-01432-7
  13. Kocak B, Ponsiglione A, Stanzione A et al (2024) CLEAR guideline for radiomics: Early insights into current reporting practices endorsed by EuSoMII. Eur J Radiol 181:111788. https://doi.org/10.1016/j.ejrad.2024.111788
  14. Zhong J, Xing Y, Hu Y et al; CLEAR working group (2025) Interpretation of CheckList for EvaluAtion of Radiomics research (CLEAR). Chinese Journal of Evidenced-Based Medicine 25(8):932-942. https://doi.org/10.7507/1672-2531.202503173 [Article in Chinese]
  15. Zhong J, Xing Y, Hu Y et al; METRICS working group (2025) Interpretation of METhodological RadiomICs Score (METRICS). Chinese Journal of Evidenced-Based Medicine 25(8):943-951. https://doi.org/10.7507/1672-2531.202504047 [Article in Chinese]