Background: To evaluate the diagnostic performance of radiomic signatures extracted from contrast-enhanced magnetic resonance imaging (CE-MRI) for the assessment of breast cancer receptor status and molecular subtypes. Methods: One hundred and forty-three patients with biopsy-proven breast cancer who underwent CE-MRI at 3 T were included in this IRB-approved HIPAA-compliant retrospective study. The training dataset comprised 91 patients (luminal A, n = 49; luminal B, n = 8; HER2-enriched, n = 11; triple negative, n = 23), while the validation dataset comprised 52 patients from a second institution (luminal A, n = 17; luminal B, n = 17; triple negative, n = 18). Radiomic analysis of manually segmented tumors included calculation of features derived from the first-order histogram (HIS), co-occurrence matrix (COM), run-length matrix (RLM), absolute gradient (GRA), autoregressive model (ARM), discrete Haar wavelet transform (WAV), and lesion geometry (GEO). Fisher, probability of error and average correlation (POE + ACC), and mutual information coefficients were used for feature selection. Linear discriminant analysis followed by k-nearest neighbor classification (with leave-one-out cross-validation) was used for pairwise radiomic-based separation of receptor status and molecular subtypes. Histopathology served as the standard of reference. Results: In the training dataset, radiomic signatures yielded the following accuracies > 80%: luminal B vs. luminal A, 84.2% (mainly based on COM features); luminal B vs. triple negative, 83.9% (mainly based on GEO features); luminal B vs. all others, 89% (mainly based on COM features); and HER2-enriched vs. all others, 81.3% (mainly based on COM features). Radiomic signatures were successfully validated in the separate validation dataset for luminal A vs. luminal B (79.4%) and luminal B vs. triple negative (77.1%). Conclusions: In this preliminary study, radiomic signatures with CE-MRI enable the assessment of breast cancer receptor status and molecular subtypes with high diagnostic accuracy. These results need to be confirmed in future larger studies.

We evaluated the performance of radiomics and artificial intelligence (AI) from multiparametric magnetic resonance imaging (MRI) for the assessment of breast cancer molecular subtypes. Ninety-one breast cancer patients who underwent 3T dynamic contrast-enhanced (DCE) MRI and diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) mapping were included retrospectively. Radiomic features were extracted from manually drawn regions of interest (n = 704 features per lesion) on initial DCE-MRI and ADC maps. The ten best features for subtype separation were selected using probability of error and average correlation coefficients. For pairwise comparisons with >20 patients in each group, a multi-layer perceptron feed-forward artificial neural network (MLP-ANN) was used (70% of cases for training, 30%, for validation, five times each). For all other separations, linear discriminant analysis (LDA) and leave-one-out cross-validation were applied. Histopathology served as the reference standard. MLP-ANN yielded an overall median area under the receiver-operating-characteristic curve (AUC) of 0.86 (0.77–0.92) for the separation of triple negative (TN) from other cancers. The separation of luminal A and TN cancers yielded an overall median AUC of 0.8 (0.75–0.83). Radiomics and AI from multiparametric MRI may aid in the non-invasive differentiation of TN and luminal A breast cancers from other subtypes.

Rationale: To determine whether ComBat harmonization improves ¹⁸F-FDG-PET radiomics-based tissue classification in pooled PET/MR and PET/CT datasets. Methods: Two-hundred patients who had undergone ¹⁸F-FDG-PET/MR (two scanners/vendors; 50 patients each) or -PET/CT (two scanners/vendors; 50 patients each) were retrospectively included. Grey-level histogram (GLH), co-occurrence matrix (GLCM), run-length matrix (GLRLM), size-zone matrix (GLSZM), and neighborhood grey-tone difference matrix (NGTDM) radiomic features were calculated for volumes of interest in the disease-free liver, spleen, and bone marrow. For individual feature classes and a multi-class radiomic signature, tissue classification was performed on ComBat-harmonized and unharmonized pooled data, using a multi-layer perceptron neural network. Results: Median accuracies in training/validation datasets were: GLH, 69.5/68.3% (harmonized) vs. 59.5/58.9% (unharmonized); GLCM, 92.1/86.1% vs. 53.6/50.0%; GLRLM, 84.8/82.8% vs. 62.4/58.3%; GLSZM, 87.6/85.6% vs. 56.2/52.8%; NGTDM, 79.5/77.2% vs. 54.8/53.9%, and radiomic signature, 86.9/84.4% vs. 62.9/58.3%. Conclusion: ComBat harmonization may be useful for multi-center ¹⁸F-FDG-PET radiomics studies using pooled PET/MR and PET/CT data.

The aim of our intra-individual comparison study was to investigate and compare the potential of radiomics analysis of contrast-enhanced mammography (CEM) and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) of the breast for the non-invasive assessment of tumor invasiveness, hormone receptor status, and tumor grade in patients with primary breast cancer. This retrospective study included 48 female patients with 49 biopsy-proven breast cancers who underwent pretreatment breast CEM and MRI. Radiomics analysis was performed by using MaZda software. Radiomics parameters were correlated with tumor histology (invasive vs. non-invasive), hormonal status (HR+ vs. HR−), and grading (low grade G1 + G2 vs. high grade G3). CEM radiomics analysis yielded classification accuracies of up to 92% for invasive vs. non-invasive breast cancers, 95.6% for HR+ vs. HR− breast cancers, and 77.8% for G1 + G2 vs. G3 invasive cancers. MRI radiomics analysis yielded classification accuracies of up to 90% for invasive vs. non-invasive breast cancers, 82.6% for HR+ vs. HR− breast cancers, and 77.8% for G1+G2 vs. G3 cancers. Preliminary results indicate a potential of both radiomics analysis of DCE-MRI and CEM for non-invasive assessment of tumor-invasiveness, hormone receptor status, and tumor grade. CEM may serve as an alternative to MRI if MRI is not available or contraindicated.

Dual-energy CT (DECT) iodine maps enable quantification of iodine concentrations as a marker for tissue vascularization. We investigated whether iodine map radiomic features derived from staging DECT enable prediction of breast cancer metastatic status, and whether textural differences exist between primary breast cancers and metastases. Seventy-seven treatment-naïve patients with biopsy-proven breast cancers were included retrospectively (41 non-metastatic, 36 metastatic). Radiomic features including first-, second-, and higher-order metrics as well as shape descriptors were extracted from volumes of interest on iodine maps. Following principal component analysis, a multilayer perceptron artificial neural network (MLP-NN) was used for classification (70% of cases for training, 30% validation). Histopathology served as reference standard. MLP-NN predicted metastatic status with AUCs of up to 0.94, and accuracies of up to 92.6 in the training and 82.6 in the validation datasets. The separation of primary tumor and metastatic tissue yielded AUCs of up to 0.87, with accuracies of up to 82.8 in the training, and 85.7 in the validation dataset. DECT iodine map-based radiomic signatures may therefore predict metastatic status in breast cancer patients. In addition, microstructural differences between primary and metastatic breast cancer tissue may be reflected by differences in DECT radiomic features.

Purpose: Background parenchymal enhancement (BPE), and the amount of fibroglandular tissue (FGT) assessed with magnetic resonance imaging (MRI) have been implicated as sensitive imaging biomarkers for breast cancer. The purpose of the study was to quantitatively assess the parenchymal uptake (BPU) on ¹⁸F-Fluorodeoxyglucose positron emission tomography/computed tomography (¹⁸F-FDG PET/CT) of the breast as another valuable imaging biomarker and examine its correlation with BPE, FGT, and age. Methods: There were 129 patients with normal imaging findings in the one breast (BI-RADS 1) included in this IRB-approved study, whose cases were retrospectively analyzed. All patients underwent prone ¹⁸F-FDG PET/CT and 3T CE-MRI of the breast. In all patients, Reader 1 assessed BPU quantitatively using maximum standard uptake values. Two readers assessed FGT and BPE of the normal contralateral breast by subjective visual estimation, as recommended by BI-RADS®. Reader 1 re-assessed all cases and repeated BPU measurements. Appropriate statistical tests were used to assess correlations of BPU, BPE, FGT, and age, as well as inter- and intra-reader agreement. Results: BPU on ¹⁸F-FDG PET/CT varied between patients. The mean BPU SUV_max values ± SD for patients with minimal BPE were 1.57 ± SD 0.6, for mild BPE 1.93 ± SD 0.6, for moderate BPE 2.42 ± SD 0.5, and for marked BPE 1.45 ± SD 0.3. There were significant (P<0.001) moderate to strong correlations among BPU, BPE, and FGT. BPU is directly correlated with both BPE and FGT on MRI. Patient age showed a moderate to strong indirect correlation with all three imaging-derived tissue biomarkers. The coefficient of variation for quantitative BPU measurements with SUV_max was 5.6%, indicating a high reproducibility. Inter-reader and intra-reader agreement for BPE and FGT was almost perfect, with a Kappa value of 0.860 and 0.822, respectively. Conclusion: The results of our study demonstrate that BPU varies between patients. BPU is directly correlated with both BPE and FGT on MRI, and BPU measurements are highly reproducible. Patient age showed strong inverse correlation with all three imaging-derived tissue biomarkers. These findings indicate that BPU may serve as a sensitive imaging biomarker for breast cancer prediction, prognosis, and risk assessment.