Correlation Between Mammographic Radiomics Features and the Level of Tumor-Infiltrating Lymphocytes in Patients With Triple-Negative Breast Cancer

Abstract

Objectives: Tumor-Infiltrating lymphocytes (TILs) have been determined as an important prognostic indicator of neoadjuvant and immunotherapy response in patients with triple-negative breast cancer (TNBC). The purpose of this study was to evaluate the associations between TIL levels and mammographic radiomic features in patients with TNBC. Methods: we retrospectively enrolled 43 patient with surgically confirmed triple-negative breast cancer who underwent preoperative mammogram from January to December 2018. A total of 204 quantitative radiomic features, including morphologic, grayscale statistic, and texture features, were extracted from the segmented lesion area using Pyradiomic software. The most statistically significant radiological features were screened by Mann-Whitney U test and Pearson correlation coefficient to evaluate the correlation between radiological features and TIL levels. The TIL level of tumors were divided into two groups: those with a TIL level of less than 50%, and those with a TIL of 50% or more. Correlation between TIL levels and clinicopathological characteristics were assessed by Chi-square test or Fisher’s exact test. Mann-Whitney U test and Pearson correlation coefficient were used to analyze the significance test and Pearson correlation coefficient of clinical pathological features, age and radiological features. Results: Of the 43 patients, 32 (74.4%) were low TIL and 11 (25.6%) were high TIL. The low TIL group was higher than the high TIL group (p=0.043) histologically. High TIL group have more negative threshold Ki-67 level (＜14%) than low TIL groups (p=0.017). The 6 top-class radiomic features (uniformity, variance, GLCM correlation, GLCM autocorrelation, GLDM low gray level emphasis, and NGTDM contrast) representing qualitative mammographic image characteristics demonstrated a statistical significance (p＜0.05) in the distinction of low and high TIL levels groups. Tumors in the high TIL levels groups had a more nonuniform density and a smoother gradient of tumor patterns than low TIL levels groups. The significance test of ki-67, age, epidermal growth factor receptor (EGFR), radiological characteristics and Pearson correlation coefficient were statistically significant (p<0.05). Conclusion: Mammography features can distinguish high and low TIL levels in TNBC patients and can be used as imaging biomarkers to improve the ability to diagnose and treat these patients.