Search

Predicting breast cancer molecular subtypes can help guide individualised clinical treatment of patients who need the rational preoperative treatment. This study aimed to investigate the efficacy of preoperative prediction of breast cancer molecular subtypes by contrast-enhanced mammography (CEM) radiomic features. This retrospective two-centre study included women with breast cancer who underwent CEM preoperatively between August 2016 and May 2022. We included 356 patients with 386 lesions, which were grouped into training (n = 162), internal test (n = 160) and external test sets (n = 64). Radiomics features were extracted from low-energy (LE) images and recombined (RC) images and selected. Three dichotomous tasks were established according to postoperative immunohistochemical results: Luminal vs. non-Luminal, human epidermal growth factor receptor (HER2)-enriched vs. non-HER2-enriched, and triple-negative breast cancer (TNBC) vs. non-TNBC. For each dichotomous task, the LE, RC, and LE+RC radiomics models were built by the support vector machine classifier. The prediction performance of the models was assessed by the area under the receiver operating characteristic curve (AUC). Then, the accuracy, sensitivity, specificity, positive predictive value, and negative predictive value were calculated for the models. DeLong's test was utilised to compare the AUCs. Radiomics models based on CEM are valuable for predicting breast cancer molecular subtypes. The LE+RC model achieved the best performance in the test set. The LE+RC model predicted Luminal, HER2-enriched, and TNBC subtypes with AUCs of 0.93, 0.89, and 0.87 in the internal test set and 0.82, 0.83, and 0.69 in the external test set, respectively. In addition, the LE model performed more satisfactorily than the RC model. CEM radiomics features can effectively predict breast cancer molecular subtypes preoperatively, and the LE+RC model has the best predictive performance. [ABSTRACT FROM AUTHOR]

Objective: To generate an image‐driven biomarker (Rad_score) to predict tumor‐infiltrating regulatory T lymphocytes (Treg) in breast cancer (BC). Methods: Overall, 928 BC patients were enrolled from the Cancer Genome Atlas (TCGA) for survival analysis; MRI (n = 71 and n = 30 in the training and validation sets, respectively) from the Cancer Imaging Archive (TCIA) were retrieved and subjected to repeat least absolute shrinkage and selection operator for feature reduction. The radiomic scores (rad_score) for Treg infiltration estimation were calculated via support vector machine (SVM) and logistic regression (LR) algorithms, and validated on the remaining patients. Results: Landmark analysis indicated Treg infiltration was a risk factor for BC patients in the first 5 years and after 10 years of diagnosis (p = 0.007 and 0.018, respectively). Altogether, 108 radiomic features were extracted from MRI images, 4 of which remained for model construction. Areas under curves (AUCs) of the SVM model were 0.744 (95% CI 0.622–0.867) and 0.733 (95% CI 0.535–0.931) for training and validation sets, respectively, while for the LR model, AUCs were 0.771 (95% CI 0.657–0.885) and 0.724 (95% CI 0.522–0.926). The calibration curves indicated good agreement between prediction and true value (p > 0.05), and DCA shows the high clinical utility of the radiomic model. Rad_score was significantly correlated with immune inhibitory genes like CTLA4 and PDCD1. Conclusions: High Treg infiltration is a risk factor for patients with BC. The Rad_score formulated on radiomic features is a novel tool to predict Treg abundance in the tumor microenvironment. [ABSTRACT FROM AUTHOR]