|
A logistic regression model utilizing computed tomography features for predicting the degree of malignancy in pure ground-glass nodule lung adenocarcinoma: A clinical investigation
|
W. Zhang   , Y. Dong , X. Wei |
| Department of Radiology, Pinghu Hospiral of Traditional Chinese Medicine, Pinghu, Zhejiang Province, China , 19957485499@163.com |
|
|
Abstract: (12 Views) |
Background: Differentiating the benign or malignant nature and the degree of infiltration of pure ground-glass nodules (pGGNs) is challenging and represents a clinical difficulty in the precise management of pulmonary nodules. It was to fabricate and validate a logistic regression (LR) model under high-resolution computed tomography (HRCT) features for the non-invasive preoperative prediction of the malignant infiltration degree in pGGN-type lung adenocarcinoma. Materials and Methods: A total of 180 patients (with 279 lesions) presenting with pure ground-glass nodules (pGGNs), defined as homogeneous ground-glass opacities without solid components on preoperative high-resolution computed tomography (HRCT), were retrospectively enrolled and randomly assigned to a model development group and a model validation group. HRCT imaging features were analyzed, including lesion size, mean CT value, margin characteristics, vacuole or cavity sign, and the relationship between vessels and the lesion. LR analyses identified independent predictors of invasive adenocarcinoma (IAC), to build the prediction model. Model discriminatory ability was evaluated. Results: Multivariate analysis identified lesion size, margin characteristics (irregular), vacuole or cavity sign, and type III vascular relationship as independent risk factors for predicting IAC. The model demonstrated good predictive performance in the development group, with area under the curve (AUC) of 0.810 (95%CI: 0.728-0.952), sensitivity of 0.782, and specificity of 0.722. The model also had robust performance in the validation group. Conclusion: The LR model under CT features effectively differentiated between pre-invasive lesions and invasive lung adenocarcinoma in pGGNs, providing a valuable non-invasive quantitative tool for formulating individualized clinical management strategies. |
|
| Keywords: Logistic regression, predictive model, pure ground-glass nodule, lung adenocarcinoma, CT, invasion degree. |
|
|
Full-Text [PDF 1160 kb]
(3 Downloads)
|
|
Type of Study: Original Research |
Subject:
Radiation Biology
|
|
|
|
|
|
|
| Send email to the article author |
|
|
|
| Add your comments about this article |
|
|
|
