Preprints with The Lancet is a collaboration between The Lancet Group of journals and SSRN to facilitate the open sharing of preprints for early engagement, community comment, and collaboration. Preprints available here are not Lancet publications or necessarily under review with a Lancet journal. These preprints are early-stage research papers that have not been peer-reviewed. The usual SSRN checks and a Lancet-specific check for appropriateness and transparency have been applied. The findings should not be used for clinical or public health decision-making or presented without highlighting these facts. For more information, please see the FAQs.
Lung CT Radiomic Signatures as Potential Biomarkers: A Soil-Seed Hypothesis for Cancer Onset
Background: Radiomic features based on CT images can capture tissue phenotypes at different spatial scales. Different lung lobes exhibit anatomical and physiological heterogeneity, which may be reflected in radiomics. We aimed to segment each lobe of the entire lung and extract radiomic features for clustering, exploring the prognostic relationships within the clustered subgroups.
Methods: Participants were incorporated from a lung cancer screening project. A total of 1,470 quantitative radiomics features were extracted from CT images from each lobe per patient. Unsupervised clustering was conducted in each lobe The top 200 contributing features and their corresponding image types for each lobe clustering process were summarized. Epidemiological and clinical differences were compared between subgroups.
Findings: Totally 11,159 male participants were enrolled. The two upper lobes and the right middle lobe were clustered into three subgroups each, while the two lower lobes were clustered into two subgroups individually. The incidence of lung cancer between subgroups was statistically significant in the right upper and right middle lobes, whereas the proportions of different smoking statuses and the incidence of lung nodules were significant across clustered subgroups in any lung lobe. The more high-risk lung lobes an individual had, the higher their incidence of lung cancer and lung nodules. Additionally, the type of features contributing the most to clustering varied across different lung lobes.
Interpretation: Unsupervised clustering can utilize radiomic features at the lobar level to identify and preliminarily differentiate changes in the pulmonary background, which is associated with epidemiological and clinical outcomes.
Funding: L. S is supported by The Zhejiang Provincial Medical and Health Science and Technology Plan (WKJ-ZJ-2330). L. Z is supported by Wenling Science and Technology Bureau (2021S00080).
Declaration of Interest: The authors have no conflicts of interest to declare.
Ethical Approval: This prospective single-center cohort study was approved by Taizhou Cancer Hospital (IRB-2019-2).
Wang, Yihuan and Zhu, Chen and Lu, Zhenzhen and Zhou, Xianglan and Lin, Chengting and Li, Yuwei and Shi, Liting and Wu, Lili and Ma, Hongxia and Zhu, Meng and Chen, Jia and Lv, Junwei and Zhu, Lingying and Du, Lingbin and Ji, Chen and Ren, Honglun and Wang, Enyu and Shi, Lei, Lung CT Radiomic Signatures as Potential Biomarkers: A Soil-Seed Hypothesis for Cancer Onset. Available at SSRN: https://ssrn.com/abstract=4986325 or http://dx.doi.org/10.2139/ssrn.4986325
Download This Paper