Evaluation of Epidermal Growth Factor Receptor 2 Status in Gastric Cancer by CT-Based Deep Learning Radiomics Nomogram

Abstract

PurposeTo explore the role of computed tomography (CT)-based deep learning and radiomics in preoperative evaluation of epidermal growth factor receptor 2 (HER2) status in gastric cancer.Materials and methodsThe clinical data on gastric cancer patients were evaluated retrospectively, and 357 patients were chosen for this study (training cohort: 249; test cohort: 108). The preprocessed enhanced CT arterial phase images were selected for lesion segmentation, radiomics and deep learning feature extraction. We integrated deep learning features and radiomic features (Inte). Four methods were used for feature selection. We constructed models with support vector machine (SVM) or random forest (RF), respectively. The area under the receiver operating characteristics curve (AUC) was used to assess the performance of these models. We also constructed a nomogram including Inte-feature scores and clinical factors.ResultsThe radiomics-SVM model showed good classification performance (AUC, training cohort: 0.8069; test cohort: 0.7869). The AUC of the ResNet50-SVM model and the Inte-SVM model in the test cohort were 0.8955 and 0.9055. The nomogram also showed excellent discrimination achieving greater AUC (training cohort, 0.9207; test cohort, 0.9224).ConclusionCT-based deep learning radiomics nomogram can accurately and effectively assess the HER2 status in patients with gastric cancer before surgery and it is expected to assist physicians in clinical decision-making and facilitates individualized treatment planning.

Deep learning nomogram for predicting neoadjuvant chemotherapy response in locally advanced gastric cancer patients

“…However, that study only extracted features from portal venous phase images, potentially overlooking crucial information present in arterial and delayed phase images. Some researchers have noted a close association between DL features extracted from arterial phase images and the heterogeneity of gastric cancer [19]. Another study utilizing features extracted from a Densenet-121 network accurately predicted NAC response in gastric cancer, achieving AUC values ranging between 0.720 and 0.806 [35].…”

“…Deep learning models can automatically identify features and representations from raw data through an end-to-end learning process, reducing the need for manual feature extraction. Previous researchers have successfully applied this technology to predict lymph node metastasis, peritoneal metastasis, molecular subtypes, and prognosis of gastric cancer, achieving satisfactory results with good generalization in validation sets [17][18][19]. Deep learning models exhibit excellent generalization, with features and patterns learned from one dataset potentially adapting well to other datasets.…”