A Radiological-Radiomics model for differentiation between minimally invasive adenocarcinoma and invasive adenocarcinoma less than or equal to 3 cm: A two-center retrospective study

Hao Dong ¹, Yuzhen Xi ², Kai Liu ³

⁰Department of Radiology, The First People's Hospital of Xiaoshan District, Xiaoshan Affiliated Hospital of Wenzhou Medical University, Hangzhou, Zhejiang, China.

European Journal of Radiology +

|

May 31, 2024

View abstract on PubMed

Summary

A novel Radiological-Radiomics (R-R) combined model effectively differentiates minimal invasive adenocarcinoma (MIA) from invasive adenocarcinoma (IA) in lung adenocarcinoma (LUAD). This R-R model demonstrates excellent diagnostic performance, aiding clinical decisions and surgical planning for LUAD patients.

Area Of Science

Oncology
Radiology
Medical Imaging
Machine Learning in Medicine

Background

Lung adenocarcinoma (LUAD) classification into minimal invasive adenocarcinoma (MIA) and invasive adenocarcinoma (IA) is crucial for treatment.
Accurate differentiation between MIA and IA impacts patient prognosis and therapeutic strategies.

Purpose Of The Study

To develop and evaluate a combined Radiological-Radiomics (R-R) model for distinguishing MIA from IA in LUAD.
To assess the predictive performance of the R-R model using clinical, pathological, and imaging data.

Main Methods

Retrospective analysis of 509 LUAD patients (522 lesions) from two medical centers.
Feature selection using the Least Absolute Shrinkage and Selection Operator (LASSO) method.
Development of Radiological, Radiomics, and R-R models using logistic regression; performance evaluated via ROC curves (AUC, sensitivity, specificity, accuracy).

Main Results

The R-R model achieved high diagnostic performance across training, validation, and external test sets.
In the external test set, the R-R model reported an AUC of 0.894, sensitivity of 84.8%, specificity of 78.6%, and accuracy of 83.3%.
The R-R model demonstrated excellent diagnostic performance in differentiating MIA and IA.

Conclusions

The developed R-R model exhibits excellent diagnostic performance for differentiating MIA and IA in LUAD.
This model can serve as a valuable reference for clinical diagnosis and surgical treatment planning in LUAD management.

Keywords:

Invasive adenocarcinoma Lung adenocarcinoma Minimally invasive adenocarcinoma Radiological Radiomics