In:
Journal of Clinical Oncology, American Society of Clinical Oncology (ASCO), Vol. 40, No. 16_suppl ( 2022-06-01), p. e20574-e20574
Abstract:
e20574 Background: Recent evidences have suggested potential applications of radiomics in early diagnosis, prognostic stratification and treatment outcome prediction of Non-Small Cell Lung Cancer (NSCLC) patients. The purpose of this study is to evaluate the ability of radiomic analysis to discriminate between different clinical-pathological conditions in patients with stage III NSCLC. Methods: Baseline CT studies from 59 patients with stage III NSCLC referred to our Institution from 2010 and 2020 were retrospectively reviewed, and the segmentation of the main lung lesion and the extraction of 517 radiomic features performed using a commercial software. The number of features was reduced to 46 by means of principal component analysis applied using the R package “RadAR” (Radiomics Analysis with R). The Kruskal-Wallis test was applied to all the radiomic features in order to evaluate which of them can discriminate between 7 clinical dichotomous characteristics: tumor stage, type, presence of mutation, treatment response, relapse free survival (RFS), smoking habit, patient outcome. P 〈 0.05 means that there is a statistically significant difference between the two subgroups. Results: The median age at diagnosis was 69 years (range 43-83). Most patients were males (40/59 = 67.8%) and heavy smokers (36/59 = 61.0%). Adenocarcinoma was the most common histology (41/59 = 70.7%), while cases were almost equally splitted between stage IIIA (45.8%) and stage IIIB or IIIC (54.2%). Most selected radiomic features (29/46 = 63.0%) showed a statistically significant difference between patients with and without mutations. Ten (10/46 = 21.7%) radiomic features were associated with patient sex. Seven features (7/45 = 15.2%) were “sensitive” to the tumor clinical stage (stage IIIA vs. stage IIIB+IIIC), 4 (4/46 = 8.7%) to the histological type, and 2 (2/46 = 4.3%) to the patient outcome. None of the selected radiomic features was able to discriminate between responder and non-responder patients, current/previous smokers and never smokers, and patients with RFS lower than 12 months versus RFS equal or higher than 12 months. Conclusions: This preliminary analysis showed that radiomics has the potential of identifying mathematical features associated with clinical and histopathological characteristics in stage III NSCLC patients, which might feed multiparametric predictive models. Larger datasets and further analysis are necessary in order to confirm initial results.
Type of Medium:
Online Resource
ISSN:
0732-183X
,
1527-7755
DOI:
10.1200/JCO.2022.40.16_suppl.e20574
Language:
English
Publisher:
American Society of Clinical Oncology (ASCO)
Publication Date:
2022
detail.hit.zdb_id:
2005181-5
Bookmarklink