Intratumoral and Peritumoral Radiomics Based on Preoperative MRI for Evaluation of Programmed Cell Death Ligand-1 Expression in Breast Cancer

Background: Programmed cell death ligand-1 (PD-L1) is a promising target for immune checkpoint blockade therapy in breast cancer. However, the preoperative evaluation of PD-L1 expression in breast cancer is rarely explored.Purpose: To determine the ability of radiomics signatures based on preoperative dynamic contrast-enhanced (DCE) MRI to evaluate PD-L1 expression in breast cancer.Study type: Retrospective.Population: 196 primary breast cancer patients with preoperative MRI and postoperative pathological evaluation of PD-L1 expression, divided into training (n = 137, 28 PD-L1-positive) and test cohorts (n = 59, 12 PD-L1-positive).Field strength/sequence: 3.0T; volume imaging for breast assessment DCE sequence.Assessment: Radiomics features were extracted from the first phase of DCE-MRI by using the minimum redundancy maximum relevance method and least absolute shrinkage and selection operator algorithm. Three radiomics signatures were constructed based on the intratumoral, peritumoral, and combined intra- and peritumoral regions. The performance of the signatures was assessed using area under the receiver operating characteristic (ROC) curve (AUC), sensitivity, specificity, and accuracy.Statistical tests: Univariable and multivariable logistic regression analysis, t-tests, chi-square tests, Fisher exact test or Yates correction, ROC analysis, and one-way analysis of variance. P < 0.05 was considered significant.Results: In the test cohort, the combined radiomics signature (AUC, 0.853) exhibited superior performance compared to the intratumoral (AUC, 0.816; P = 0.528) and peritumoral radiomics signatures (AUC, 0.846; P = 0.905) in PD-L1 status evaluation, although the differences did not reach statistical significance.Data conclusion: Intratumoral and peritumoral radiomics signatures based on preoperative breast MRI showed some potential accuracy for the non-invasive evaluation of PD-L1 status in breast cancer.