The value of CT-based radiomics in early assessment of chemotherapeutic effect in patients with advanced lung adenocarcinoma: a preliminary study

• Published in: Acta radiologica (1987) Vol. 64; no. 2; p. 524
• Main Authors: Lin, Xi, Wang, Huaming, Chen, Jiayou, Lu, Tao, Zheng, Dechun, Chen, Ying, Chen, Yunbin
• Format: Journal Article
• Language: English
• Published: England 01.02.2023
• Subjects: Adenocarcinoma of Lung - diagnostic imaging; Adenocarcinoma of Lung - drug therapy; Adenocarcinoma of Lung - pathology; Humans; Lung Neoplasms - diagnostic imaging; Lung Neoplasms - drug therapy; Lung Neoplasms - pathology; Retrospective Studies; ROC Curve; Tomography, X-Ray Computed - methods; lung cancer; radiomics; response evaluation; Computed tomography; chemotherapy; texture analysis
• ISSN: 1600-0455
• DOI: 10.1177/02841851221078290

Computed tomography (CT) is the preferred method for evaluating the therapeutic effect of lung cancer. Radiomics parameters can provide a lot of supplementary information for clinical diagnosis and treatment. To investigate the value of radiomics features of CT imaging to predict and evaluate the early efficacy of chemotherapy in patients with advanced lung adenocarcinoma. A total of 101 patients with advanced lung adenocarcinoma were enrolled. Patients were classified into a response group and non-response group according to RECIST 1.1 standard. All patients underwent chest CT examination before and after two cycles of chemotherapy. A total of 293 radiomics features were calculated. The features between response group and non-response group were compared before and after chemotherapy. The diagnostic efficacy was evaluated using the receiver operating characteristic curve. The six pre-chemotherapy radiomics features were selected, with area under the curve (AUC), sensitivity, and specificity at 0.720, 68.3%, and 69.0% in the training group and 0.573, 50.0%, and 76.9% in the test group, respectively. The eleven post-chemotherapy radiomics features were selected, with AUC, sensitivity, specificity at 0.789, 75.6%, and 75.9% in the training group and 0.718, 61.1%, and 76.9% in the test group, respectively. The prognostic value of △f8, △f16, %f8, and %f16 were higher than the other features with AUCs of 0.787, 0.837, 0.763, and 0.877, respectively. Radiomics is expected to provide more valuable information for evaluating the chemotherapy efficacy of lung adenocarcinoma.