Scholarlink

Background <p>The concept of field cancerization highlights spatially diffuse, pre-malignant changes in carcinogen-exposed lung tissue, yet current screening rarely captures such effects regionally. This exploratory study aims to quantify field cancerization via lobar-level radiomic clustering and assess its association with lung cancer risk in high-risk smokers.</p> Materials and methods <p>A total of 10,280 male current or former smokers (mean age, 62.1 ± 6.34 years) were enrolled from a high-risk population undergoing lung cancer screening. Unsupervised clustering of CT-derived radiomic features was performed for each lobe. A logistic regression-derived weighted spatial risk score was developed to quantify cumulative lobar risk. Cancer incidence associations were assessed after adjusting for polygenic risk and epidemiological factors, with stratified analyses by genetic risk and smoking duration.</p> Results <p>Distinct radiomic clusters were observed across all lobes, with the right upper lobe demonstrating a significantly higher lung cancer incidence in the high-risk cluster (0.997% vs 0.593%, <i>p</i> = 0.020). The weighted spatial risk score was independently associated with cancer risk (OR = 1.09, 95% CI: 1.02–1.16, <i>p</i> = 0.010). There was no significant interaction between the score and PRS (<i>p</i> = 0.938), suggesting a genetic background-independent effect. In stratified analysis, the score was significantly associated with lung cancer among long-term smokers (OR = 1.08, 95% CI: 1.00–1.17, <i>p</i> = 0.049), with a similar but nonsignificant trend in short-term smokers.</p> Conclusion <p>Unsupervised clustering of lobar radiomics reveals background pulmonary alterations, supporting field cancerization and the “seed-and-soil” hypothesis, and offers an exploratory imaging-based framework for cancer risk stratification in screening.</p> Critical relevance statement <p>Unsupervised clustering of radiomic features at the pulmonary lobe level identified distinct background patterns associated with nodule and cancer incidence. A derived weighted lobe score was independently associated with lung cancer risk, especially among individuals with prolonged smoking histories.</p> Key Points <p><UnorderedList Mark="Bullet"> <ItemContent> <p>Lobar radiomic clustering identifies background lung changes related to cancer risk.</p> </ItemContent> <ItemContent> <p>Weighted lobe spatial risk score predicts lung cancer risk independently of genetic background.</p> </ItemContent> <ItemContent> <p>Longer smoking duration strengthens the link between lung damage and cancer risk.</p> </ItemContent> </UnorderedList></p> Graphical Abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Lobar-level radiomic clustering reveals background lung changes associated with lung cancer risk: a new perspective for early screening

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月20日

查看全文
  • Yihuan Wang,
  • Chen Zhu,
  • Zhenzhen Lu,
  • Xianglan Zhou,
  • Chengting Lin,
  • Yuwei Li,
  • Liting Shi,
  • Lili Wu,
  • Hongxia Ma,
  • Meng Zhu,
  • Jia Chen,
  • Junwei Lv,
  • Lingying Zhu,
  • Lingbin Du,
  • Chen Ji,
  • Honglun Ren,
  • Enyu Wang,
  • Lei Shi

摘要

Background

The concept of field cancerization highlights spatially diffuse, pre-malignant changes in carcinogen-exposed lung tissue, yet current screening rarely captures such effects regionally. This exploratory study aims to quantify field cancerization via lobar-level radiomic clustering and assess its association with lung cancer risk in high-risk smokers.

Materials and methods

A total of 10,280 male current or former smokers (mean age, 62.1 ± 6.34 years) were enrolled from a high-risk population undergoing lung cancer screening. Unsupervised clustering of CT-derived radiomic features was performed for each lobe. A logistic regression-derived weighted spatial risk score was developed to quantify cumulative lobar risk. Cancer incidence associations were assessed after adjusting for polygenic risk and epidemiological factors, with stratified analyses by genetic risk and smoking duration.

Results

Distinct radiomic clusters were observed across all lobes, with the right upper lobe demonstrating a significantly higher lung cancer incidence in the high-risk cluster (0.997% vs 0.593%, p = 0.020). The weighted spatial risk score was independently associated with cancer risk (OR = 1.09, 95% CI: 1.02–1.16, p = 0.010). There was no significant interaction between the score and PRS (p = 0.938), suggesting a genetic background-independent effect. In stratified analysis, the score was significantly associated with lung cancer among long-term smokers (OR = 1.08, 95% CI: 1.00–1.17, p = 0.049), with a similar but nonsignificant trend in short-term smokers.

Conclusion

Unsupervised clustering of lobar radiomics reveals background pulmonary alterations, supporting field cancerization and the “seed-and-soil” hypothesis, and offers an exploratory imaging-based framework for cancer risk stratification in screening.

Critical relevance statement

Unsupervised clustering of radiomic features at the pulmonary lobe level identified distinct background patterns associated with nodule and cancer incidence. A derived weighted lobe score was independently associated with lung cancer risk, especially among individuals with prolonged smoking histories.

Key Points

Lobar radiomic clustering identifies background lung changes related to cancer risk.

Weighted lobe spatial risk score predicts lung cancer risk independently of genetic background.

Longer smoking duration strengthens the link between lung damage and cancer risk.

Graphical Abstract