Fei-Hong Yu1, Jian-Xiang Wang1, Xin-Hua Ye1, Jing Deng1, Jing Hang1, Bin Yang2. 1. Department of Ultrasound, The First Affiliated Hospital of Nanjing Medical University, Nanjing, China. 2. Department of Ultrasound, Jinling Clinical Medical College, Nanjing Medical University, Nanjing, China. Electronic address: yangbin12yb@126.com.
Abstract
PURPOSE: To establish a radiomics nomogram integrating clinical factors and radiomics features from ultrasound for the preoperative diagnosis axillary lymph node (ALN) status in patients with early-stage invasive breast cancer (EIBC). MATERIALS AND METHODS: Between September 2016 and December 2018, four hundred twenty-six ultrasound manually segmented images of patients with EIBC were enrolled in our retrospective study, which were divided into a primary cohort (n = 300) and a validation cohort (n = 126). A radiomics signature was built with the least absolute shrinkage and selection operator (LASSO) algorithm in the primary cohort. Multivariable logistic regression analysis was used to establish a radiomics nomogram model based on radiomics signature and clinical variables. The performance of nomogram was quantified with respect to discrimination and calibration. The radiomics model was further evaluated in the internal validation cohort. RESULTS: The radiomics signature, consisted of fourteen selected ALN-status-related features, achieved moderate prediction efficacy with an area under the curve (AUC) of 0.78 and 0.71 in the primary and validation cohorts respectively. The radiomics nomogram, comprising tumor size, US-reported LN status and radiomics signature, showed good calibration and favorite performance for ALN detection (AUC 0.84 and 0.81 in the primary and validation cohort). The decision curve which was demonstrated the radiomics nomogram displayed good clinical utility. CONCLUSION: The radiomics nomogram could hold promise as a non-invasive and reliable tool in predicting ALN metastasis and may facilitate to develop more effective preoperative decision-making.
PURPOSE: To establish a radiomics nomogram integrating clinical factors and radiomics features from ultrasound for the preoperative diagnosis axillary lymph node (ALN) status in patients with early-stage invasive breast cancer (EIBC). MATERIALS AND METHODS: Between September 2016 and December 2018, four hundred twenty-six ultrasound manually segmented images of patients with EIBC were enrolled in our retrospective study, which were divided into a primary cohort (n = 300) and a validation cohort (n = 126). A radiomics signature was built with the least absolute shrinkage and selection operator (LASSO) algorithm in the primary cohort. Multivariable logistic regression analysis was used to establish a radiomics nomogram model based on radiomics signature and clinical variables. The performance of nomogram was quantified with respect to discrimination and calibration. The radiomics model was further evaluated in the internal validation cohort. RESULTS: The radiomics signature, consisted of fourteen selected ALN-status-related features, achieved moderate prediction efficacy with an area under the curve (AUC) of 0.78 and 0.71 in the primary and validation cohorts respectively. The radiomics nomogram, comprising tumor size, US-reported LN status and radiomics signature, showed good calibration and favorite performance for ALN detection (AUC 0.84 and 0.81 in the primary and validation cohort). The decision curve which was demonstrated the radiomics nomogram displayed good clinical utility. CONCLUSION: The radiomics nomogram could hold promise as a non-invasive and reliable tool in predicting ALN metastasis and may facilitate to develop more effective preoperative decision-making.