Intratumoral and peritumoral radiomics for the pretreatment prediction of pathological complete response to neoadjuvant chemotherapy based on breast DCE-MRI

Abstract

Background: In this study, we evaluated the ability of radiomic textural analysis of intratumoral and peritumoral regions on pretreatment breast cancer dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to predict pathological complete response (pCR) to neoadjuvant chemotherapy (NAC).

Methods: A total of 117 patients who had received NAC were retrospectively analyzed. Within the intratumoral and peritumoral regions of T1-weighted contrast-enhanced MRI scans, a total of 99 radiomic textural features were computed at multiple phases. Feature selection was used to identify a set of top pCR-associated features from within a training set (n = 78), which were then used to train multiple machine learning classifiers to predict the likelihood of pCR for a given patient. Classifiers were then independently tested on 39 patients. Experiments were repeated separately among hormone receptor-positive and human epidermal growth factor receptor 2-negative (HR⁺, HER2^-) and triple-negative or HER2⁺ (TN/HER2⁺) tumors via threefold cross-validation to determine whether receptor status-specific analysis could improve classification performance.

Results: Among all patients, a combined intratumoral and peritumoral radiomic feature set yielded a maximum AUC of 0.78 ± 0.030 within the training set and 0.74 within the independent testing set using a diagonal linear discriminant analysis (DLDA) classifier. Receptor status-specific feature discovery and classification enabled improved prediction of pCR, yielding maximum AUCs of 0.83 ± 0.025 within the HR⁺, HER2^- group using DLDA and 0.93 ± 0.018 within the TN/HER2⁺ group using a naive Bayes classifier. In HR⁺, HER2^- breast cancers, non-pCR was characterized by elevated peritumoral heterogeneity during initial contrast enhancement. However, TN/HER2⁺ tumors were best characterized by a speckled enhancement pattern within the peritumoral region of nonresponders. Radiomic features were found to strongly predict pCR independent of choice of classifier, suggesting their robustness as response predictors.

Conclusions: Through a combined intratumoral and peritumoral radiomics approach, we could successfully predict pCR to NAC from pretreatment breast DCE-MRI, both with and without a priori knowledge of receptor status. Further, our findings suggest that the radiomic features most predictive of response vary across different receptor subtypes.

Keywords: Imaging; MRI; Neoadjuvant chemotherapy; Personalized medicine; Radiomics; Treatment response.

Fig. 1

Top: Patient selection flowchart for experiments 1–3. Bottom: Radiomic pathological complete response (pCR) prediction pipeline. DCE-MRI Dynamic contrast-enhanced magnetic resonance imaging, ER Estrogen receptor, HER2 Human epidermal growth factor receptor 2, HR Hormone receptor, NAC Neoadjuvant chemotherapy, TN Triple-negative

Fig. 2

Consensus clustering using combined peritumoral and intratumoral radiomics, intratumoral radiomics, peritumoral radiomics, and pharmacokinetic parameter feature sets. Combination of intratumoral and peritumoral features yielded clusters with the best consensus and correlation to pathological complete response (pCR) status

Fig. 3

a, c, e Feature expression maps for top radiomic features. b, d, f Corresponding hematoxylin and eosin-stained images at × 100 original magnification taken from the original diagnostic core biopsy specimen before neoadjuvant chemotherapy. All-comers: a Co-occurrence of Local Anisotropic Gradient Orientations (CoLlAGe) information measure of correlation 1 is elevated in pathological complete response (pCR) tumors intratumorally during the initial postcontrast phase. b For a patient who experienced pCR, the corresponding histology shows a high percentage of stromal tumor-infiltrating lymphocytes (TILs) present relatively uniformly within the invasive carcinoma. The histopathological image from the non-pCR patient on the right shows a heterogeneous mix of tumor cells, necrosis, and sclerosis. HR ⁺ /HER2 ⁻ : c Peritumoral initial CoLlAGe entropy is increased among HR⁺, HER2⁻ nonresponders. d Corresponding histology for peritumoral regions of HR⁺/HER2⁻ patients with and without a pCR. The image on the left shows a brisk lymphocytic response at the periphery of the tumor. The image on the right from the non-pCR patient shows tumor cells infiltrating the adipose tissue at the periphery of the lesion without a significant stromal response. TN/HER2 ⁺ : e Peritumoral peak Laws level-ripple is elevated in non-pCR tumors. f Peritumoral regions from patients with TN breast cancer with and without a pCR. Once again, there is a brisk lymphocytic response in the peritumoral region and numerous stromal TILs within the tumor on the left. The image of the non-pCR on the right is from a patient with TN breast cancer with a matrix-producing metaplastic carcinoma. The tumor cells with associated chondroid matrix are dissecting through the adipose tissue at the periphery of the lesion. HER2 Human epidermal growth factor receptor 2, HR Hormone receptor, TN Triple-negative