AUTHOR=Yang Ran , Hui Dongming , Li Xing , Wang Kun , Li Caiyong , Li Zhichao 

TITLE=Prediction of single pulmonary nodule growth by CT radiomics and clinical features — a one-year follow-up study

JOURNAL=Frontiers in Oncology

VOLUME=Volume 12 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.1034817

DOI=10.3389/fonc.2022.1034817

ISSN=2234-943X

ABSTRACT=Background:
More and more pulmonary nodules have been found. Some of them may gradually grow and develop into lung cancer, while others may stay stable. Accurately predicting the growth of them in advance is of great clinical significance for early treatment. The purpose of this study was to establish a predictive model using radiomics and to study its value in predicting the growth of pulmonary nodules.
Methods:
According to the inclusion and exclusion conditions, 228 pulmonary nodules in 228 subjects were included in the study. 69/228 nodules grew larger, and 159/228 nodules remained stable. All the nodules were divided into training groups and validation groups at 7:3. The T test, Chi-square test and Fisher exact test were employed to analyze the gender, age and nodule location between the two groups. Two radiologists independently delineated the ROI of the nodule to extract the radiomics characteristics using Pyradiomics. After dimension reduction by LASSO algorithm, logistic regression analysis was performed on ten selected radiological features and ages, and a prediction model was established and tested in the validation group. SVM, RF, MLP and AdaBoost models were also established, and the prediction effect was evaluated by ROC analysis.
Results:
There was a significant difference in age between the growth group and the stable group (P < 0.05), but no significant difference in gender and nodule location (P > 0.05). After dimension reduction by LASSO algorithm, ten radiomic features were selected. The logistics model combining age and radiomics features achieved an AUC of 0.87, accuracy of 0.82 in the training group, and an AUC of 0.82, accuracy of 0.84 in the verification group for the prediction of nodule growth. In the training group, the AUC of SVM, RF,MLP and boost models are respectively 0.95, 1.0, 1.0 and 1.0. In the validation group, the AUC are respectively 0.81, 0.77, 0.81,0.71.
Conclusions:
In this study, the logistic regression model combining age and imaging parameters has the best accuracy and generalization. This is very helpful for the early treatment of pulmonary nodules and has important clinical significance.