Radiomics in cone-beam breast CT for the prediction of axillary lymph node metastasis in breast cancer: a multi-center multi-device study

Abstract

ObjectivesTo develop a radiomics model in contrast-enhanced cone-beam breast CT (CE-CBBCT) for preoperative prediction of axillary lymph node (ALN) status and metastatic burden of breast cancer.Methods Two hundred and seventy-four patients who underwent CE-CBBCT examination with two scanners between 2012 and 2021 from two institutions were enrolled. The primary tumor was annotated in each patient image, from which 1781 radiomics features were extracted with PyRadiomics. After feature selection, support vector machine models were developed to predict ALN status and metastatic burden. To avoid overfitting on a specific patient subset, 100 randomly stratified splits were made to assign the patients to either training/fine-tuning or test set. Area under the receiver operating characteristic curve (AUC) of these radiomics models was compared to those obtained when training the models only with clinical features and combined clinical-radiomics descriptors. Ground truth was established by histopathology. ResultsOne hundred and eighteen patients had ALN metastasis (N + (≥ 1)). Of these, 74 had low burden (N + (1-2)) and 44 high burden (N + (≥ 3)). The remaining 156 patients had none (N0). AUC values across the 100 test repeats in predicting ALN status (N0/N + (≥ 1)) were 0.75 ± 0.05 (0.67-0.93, radiomics model), 0.68 ± 0.07 (0.53-0.85, clinical model), and 0.74 ± 0.05 (0.67-0.88, combined model). For metastatic burden prediction (N + (1-2)/N + (≥ 3)), AUC values were 0.65 ± 0.10 (0.50-0.88, radiomics model), 0.55 ± 0.10 (0.40-0.80, clinical model), and 0.64 ± 0.09 (0.50-0.90, combined model), with all the ranges spanning 0.5. In both cases, the radiomics model was significantly better than the clinical model (both p < 0.01) and comparable with the combined model (p = 0.56 and 0.64).Conclusions Radiomics features of primary tumors could have potential in predicting ALN metastasis in CE-CBBCT imaging. Clinical relevance statementThe findings support potential clinical use of radiomics for predicting axillary lymph node metastasis in breast cancer patients and addressing the limited axilla coverage of cone-beam breast CT.

Performance evaluation of ML models for preoperative prediction of HER2-low BC based on CE-CBBCT radiomic features: A prospective study

“…Although the exact biological characteristics captured by these 2 features cannot be determined, it is hypothesized that these subtle details may be associated with slight changes in microvessel density and the tumor microenvironment. [ 20 ]…”

“…CE-CBBCT has shown promising results in lesion identification and differentiation, [ 14 , 15 ] molecular subtyping, [ 16 , 17 ] assessment of disease extent, [ 12 , 18 ] and prediction of axillary lymph node metastasis. [ 19 , 20 ]…”

Pre-treatment Contrast-enhanced Cone Beam Breast CT Imaging Features Combining with Clinicopathological Characteristics to Predict the Response of Neoadjuvant Chemotherapy: A Preliminary Feasibility Study

“…Meta-analysis has shown that the diagnostic sensitivity of CBBCT was higher than that of DM and ultrasound, and CE-CBBCT has comparable sensitivity but slightly lower speci city compared to MRI [17,18]. In addition, CBBCT has shown promise in preoperative assessment, including determining the extent of the primary tumor [19,20] and predicting axillary lymph node metastasis [21,22]. Some studies have also suggested that CE-CBBCT features can assist in characterizing IHC receptor status and distinguishing molecular subtypes of breast cancer [13,23,24].…”