Machine Learning-Based Prediction of Pathological Responses and Prognosis After Neoadjuvant Chemotherapy for Non–Small-Cell Lung Cancer: A Retrospective Study

Background Neoadjuvant chemotherapy has variable efficacy in patients with non-small cell lung cancer (NSCLC), yet reliable noninvasive predictive markers are lacking. This study aimed to develop a radiomics model predicting pathological complete response(pCR) and post–neoadjuvant chemotherapy survival in NSCLC. Methods Retrospective data collection involved 130 patients with NSCLC who underwent neoadjuvant chemotherapy and surgery. Patients were randomly divided into training and independent testing sets. Nine radiomics features from pre-chemotherapy CT images were extracted from intratumoral and peritumoral regions. An auto-encoder (AE) model was constructed, and it's performance was evaluated. X-tile software classified patients into high and low-risk groups based on their predicted probabilities. survival of patients in different risk groups and the role of postoperative adjuvant chemotherapy were examined. Results The model demonstrated area under the receiver operating characteristic (ROC) curve (AUC) of 0.874 (training set) and 0.876 (testing set). The higher the AUC value, the better the model performance. Calibration curve and decision curve analysis(DCA) indicated excellent model calibration (Hosmer-Lemeshow test, P = 0.763, the higher the P-value, the better the model fit) and potential clinical applicability. Survival analysis revealed significant differences in overall survival (OS, P = 0.011) and disease-free survival (DFS, P = 0.017) between different risk groups. Adjuvant chemotherapy significantly improved survival in the low-risk group (P = 0.041) but not high-risk group (P = 0.56). Conclusions This study represents the first successful prediction of pCR achievement after neoadjuvant chemotherapy for NSCLC, as well as the patients' survival, utilizing intratumoral and peritumoral radiomics features. Micro Abstract We constructed a radiomics model predicting pathological complete response and survival after neoadjuvant chemotherapy in NSCLC, utilizing intratumoral and peritumoral radiomics features of 130 people. And the model demonstrated the AUCs in training set and testing set are 0.874 and 0.876. So, this study may provide new ideas for decision-making about individualized treatment.