Radiomics nomogram integrating intratumoural and peritumoural features to predict lymph node metastasis and prognosis in clinical stage IA non-small cell lung cancer: a two-centre study

To investigate the value of a radiomics nomogram integrating intratumoural and peritumoural features in predicting lymph node metastasis and overall survival (OS) in patients with clinical stage IA non-small-cell lung cancer (NSCLC). This study retrospectively enrolled 199 patients (training cohort:...

Full description

Saved in:
Bibliographic Details
Published in:Clinical radiology 2023-05, Vol.78 (5), p.e359-e367
Main Authors: Chen, Q.-L., Li, M.-M., Xue, T., Peng, H., Shi, J., Li, Y.-Y., Duan, S.-F., Feng, F.
Format: Article
Language:English
Subjects:
Carcinoma, Non-Small-Cell Lung - diagnostic imaging
Humans
Lung Neoplasms - diagnostic imaging
Lymph Nodes
Lymphatic Metastasis - diagnostic imaging
Nomograms
Prognosis
Retrospective Studies
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:To investigate the value of a radiomics nomogram integrating intratumoural and peritumoural features in predicting lymph node metastasis and overall survival (OS) in patients with clinical stage IA non-small-cell lung cancer (NSCLC). This study retrospectively enrolled 199 patients (training cohort: 71 patients from Affiliated Tumour Hospital of Nantong University; internal validation cohort: 46 patients from Affiliated Tumour Hospital of Nantong University; external validation cohort: 82 patients from the public database). CT radiomics models were constructed based on four volumes of interest: gross tumour volume (GTV), gross and 3 mm peritumoural volume (GPTV3), gross and 6 mm peritumoural volume (GPTV6), and gross and 9 mm peritumoural volume (GPTV9). The optimal radiomics signature was further combined with independent clinical predictors to develop a nomogram. Univariable and multivariable Cox regression analysis were applied to determine the relationship between factors and OS. GPTV6 radiomics yielded better performance than GTV, GPTV3, and, GPTV9 radiomics in the training (area under the curve [AUC], 0.81), internal validation (AUC, 0.79), and external validation cohorts (AUC, 0.71), respectively. The nomogram integrating GPTV6 radiomics and spiculation improved predictive ability, with AUCs of 0.85, 0.80, and 0.74 in three cohorts, respectively. Pathological lymph node metastasis, nomogram-predicted lymph node metastasis, and pleural indentation were independent risk predictors of OS (p
ISSN:0009-9260
1365-229X
DOI:10.1016/j.crad.2023.02.004