CLEAR item#50

“Previous works with differences from the current study. Provide the most important and relevant previous works. Mention the most prominent differences between the current study and the previous works.” [1] (from the article by Kocak et al.; licensed under CC BY 4.0)

Reporting examples for CLEAR item#50

Example#1. “A multi-center study focusing on Stage IV EGFR-mutant NSCLC patients treated with EGFR-TKI therapy reported that the nomogram integrating pretherapy CT-based signature improved the prediction accuracy of PFS significantly. A noteworthy difference from our work was the introduction of additional timepoint and our signature reflected the dynamic change of tumor heterogeneity. Tumor heterogeneity not only changes over time owing to subclonal diversities and genomic instability [36] but also alters between pre- and post-treatment under potent selective pressures arising from antineoplastic therapy, especially in targeted therapies. […] Among the ten features constituting our radiomics signature, seven delta-radiomics features contributed crucial information associated with patient-specific outcome prediction. Previous studies presented controversy over the calculation and application of these delta-radiomics features. Nardone et al discovered delta-texture features were more robust than texture features through a phantom study. A retrospective study of 107 NSCLC patients under radiotherapy revealed that pretreatment imaging features were not prognostic, whereas delta-radiomics features had a statistically significant impact on estimation for local recurrence. Contrarily in another smaller study consisting of 48 EGFR-mutant LUAD patients who received EGFR-TKIs, follow-up features were more predictive for PFS than the percentage change. We suspected that different duration between therapy initiation and first follow-up imaging among patients could create bias. Therefore, we adjusted such factor by calculating the unit time (per day) percentage change between the two timepoints, and our data indeed proved the prognostic value of these dynamic characteristics.” [2] (from the article by Zhang et al.; licensed under CC BY 4.0)

Example#2. “Unlike previous studies that mostly cataloged radiomic features to delineate PCNSL from other tumors such as glioblastoma, our study identified clinically relevant subgroups within PCNSL. In line with recent work, textural features were particularly informative. Of note, in the present cohort, the tumor surroundings contributed relevant information to the model, which is important to consider since those regions are not necessarily included in radiomic studies. Moreover, we evaluated the radiomic risk score in a multicenter setting across different MRI scanning conditions and distinct patient populations, suggesting resilience and generalizability of informative imaging-based features in patients with PCNSL.” [3] (from the article by Nenning et al.; licensed under CC BY-NC 4.0)

Example#3. “Radiomics, utilizing diverse machine learning techniques to construct predictive models, can non-invasively reflect the internal heterogeneity of lesions for early diagnosis, differential diagnosis, and prognostic predictions. This approach has been extensively investigated in various cancer studies. AVM related hematomas are more likely to be irregular and heterogeneous due to the presence of calcification and malformed vasculatures embedded in the hematomas, whereas hypertensive intracerebral hematomas are more likely to have a uniform shape. These imaging features can be reflected by radiomics features. So far, radiomics analysis of vascular diseases has primarily focused on identifying and stratifying the stability of vessel plaques, predicting cerebral hematomas expansion, and identifying tumorous intracerebral hemorrhages. These previous studies have demonstrated that radiomics features may have the potential to objectively quantify the shape of hematomas and the heterogeneity of hematomas. In this study, three radiomics features belonging to the shape and many radiomics features belonging to the first order, GLCM, GLDM, GLRLM, GLSZM, and NGTDM that described the heterogeneity of the hematoma may be associated with AVM related hematomas (Supplementary Table S2). The non-enhanced CT based radiomics model may present potential benefits such as reduced contrast and radiation exposure, as well as faster treatment times that could lead to improved outcomes. In our study, we found that …” [4] (from the article by Xie et al.; licensed under CC BY 4.0)

Explanation and elaboration of CLEAR item#50

The discussion section focuses on the interpretation of results, particularly on the differences with previous work. The results presented in the manuscript must be contextualized and integrated with the previous evidence [5]. That evidence could include similar studies with similar models, outcomes, and results, studies with different models with similar outcomes, or studies with different outcomes that can be relevant. The analytic comparison with studies is important to highlight the strengths and weaknesses of the presented model [6, 7]. The comparison of the presented model with previous studies should include systematic reviews or meta-analyses if available. Otherwise, the authors should provide a review of the literature, and in particular, they should refer to papers already presented in the introduction section [5, 7, 8]. The authors should focus on differences in model building, predictors included, performance, populations, validation process, and generalizability: these points can be used to reinforce the claims made in the manuscript [7]. In particular, the elucidation of discrepancies with previous studies is encouraged since they can be the most captivating points of the discussion section [5].

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 14:75. https://doi.org/10.1186/s13244-023-01415-8
  2. Zhang X, Lu B, Yang X, et al (2023) Prognostic analysis and risk stratification of lung adenocarcinoma undergoing EGFR-TKI therapy with time-serial CT-based radiomics signature. Eur Radiol 33:825–835. https://doi.org/10.1007/s00330-022-09123-5
  3. Nenning K-H, Gesperger J, Furtner J, et al (2023) Radiomic features define risk and are linked to DNA methylation attributes in primary CNS lymphoma. Neuro-Oncol Adv 5:vdad136. https://doi.org/10.1093/noajnl/vdad136
  4. Xie H, Dong F, Zhang R, et al (2023) Building nonenhanced CT based radiomics model in discriminating arteriovenous malformation related hematomas from hypertensive intracerebral hematomas. Front Neurosci 17. https://doi.org/10.3389/fnins.2023.1284560
  5. Bannas P, Reeder SB (2017) How to write an original radiological research manuscript. Eur Radiol 27:4455–4460. https://doi.org/10.1007/s00330-017-4879-8
  6. Siontis GCM, Tzoulaki I, Siontis KC, Ioannidis JPA (2012) Comparisons of established risk prediction models for cardiovascular disease: systematic review. BMJ 344:e3318. https://doi.org/10.1136/bmj.e3318
  7. Moons KGM, Altman DG, Reitsma JB, et al (2015) Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): explanation and elaboration. Ann Intern Med 162:W1-73. https://doi.org/10.7326/M14-0698
  8. Clarke M, Chalmers I (1998) Discussion Sections in Reports of Controlled Trials Published in General Medical JournalsIslands in Search of Continents? JAMA 280:280–282. https://doi.org/10.1001/jama.280.3.280

Back Next