Radiomics features for assessing tumor-infiltrating lymphocytes correlate with molecular traits of triple-negative breast cancer

Abstract

Background Tumor-infiltrating lymphocytes (TILs) have become a promising biomarker for assessing tumor immune microenvironment and predicting immunotherapy response. However, the assessment of TILs relies on invasive pathological slides. Methods We retrospectively extracted radiomics features from magnetic resonance imaging (MRI) to develop a radiomic cohort of triple-negative breast cancer (TNBC) (n = 139), among which 116 patients underwent transcriptomic sequencing. This radiomic cohort was randomly divided into the training cohort (n = 98) and validation cohort (n = 41) to develop radiomic signatures to predict the level of TILs through a non-invasive method. Pathologically evaluated TILs in the H&E sections were set as the gold standard. Elastic net and logistic regression were utilized to perform radiomics feature selection and model training, respectively. Transcriptomics was utilized to infer the detailed composition of the tumor microenvironment and to validate the radiomic signatures. Results We selected three radiomics features to develop a TILs-predicting radiomics model, which performed well in the validation cohort (AUC 0.790, 95% confidence interval (CI) 0.638–0.943). Further investigation with transcriptomics verified that tumors with high TILs predicted by radiomics (Rad-TILs) presented activated immune-related pathways, such as antigen processing and presentation, and immune checkpoints pathways. In addition, a hot immune microenvironment, including upregulated T cell infiltration gene signatures, cytokines, costimulators and major histocompatibility complexes (MHCs), as well as more CD8+ T cells, follicular helper T cells and memory B cells, was found in high Rad-TILs tumors. Conclusions Our study demonstrated the feasibility of radiomics model in predicting TILs status and provided a method to make the features interpretable, which will pave the way toward precision medicine for TNBC.

Multi-View Radiomics Feature Fusion Reveals Distinct Immuno-Oncological Characteristics and Clinical Prognoses in Hepatocellular Carcinoma

“…Hectors et al found that texture features could reflect immune status in HCC patients, and were highly correlated with mRNA and protein expression of PDL1 and the markers of macrophages (CD68) and T cells (CD3) [ 23 ]. Su et al developed a radiomics model for tumor-infiltrating lymphocyte status based on GLCM and GLSZM texture features [ 40 ]. Ming et al also revealed that cell cycle pathway exhibited significant associations with SurfaceVolumeRatio (PRF9 and PRF10 in our study) [ 75 , 76 ].…”

“…The abundance of individual immune cell types in TIME was inferred using the CIBERSORT interface [ 39 ], which estimated the relative fraction of immune cell types based on the deconvolution method. The gene sets of immune-related molecules were compared by GSVA, including cell surface immune-related molecules (costimulators, coinhibitors, and major histocompatibility complex) and cytokines (interleukins, chemokines, interferons, and colony-stimulating factors) [ 40 ]. Tumor immunogenicity was calculated by immunophenoscore (IPS) based on the gene expression in effector cells, immunosuppressive cells, MHC molecules, and immunomodulators [ 41 ].…”

Tumor infiltrating lymphocytes and radiological picture of the tumor

“…The Radiomics TILs score (Rad-TILs score), based on selected 3 radiomics features, was then juxtaposed with transcriptomics data concerning expression of several genes related to immunity. It has been shown that implemented radiomics model could be used to reflect lesions characterized with different immune patterns [ 66 ]. However, the role of MRI in estimating TILs density requires further analysis, particularly in the light of another recent report, which has shown that the combination of MRI radiomics and pathological TILs evaluation yielded the best results in assessing the efficacy of NAC compared to these methods alone.…”

Freehand 1.5T MR-Guided Vacuum-Assisted Breast Biopsy (MR-VABB): Contribution of Radiomics to the Differentiation of Benign and Malignant Lesions

“…Moreover, it has been estimated that around 50–70% of MRI findings requiring biopsy turn out to be non-cancer cases [ 5 ]. In recent years, radiomics and artificial intelligence have been increasingly applied in breast MRI [ 6 ], not only to classify breast lesions or predict response to treatment but also to differentiate benign from malignant lesions [ 6 , 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 , 26 ]. Indeed, radiological practice includes a qualitative and subjective evaluation of the visible lesion describing different characteristics of the tumor aspect (e.g., the presence of spiculations and the presence of necrosis and microcalcifications).…”