基于乳腺DCE-MRI的肿瘤内和肿瘤周围放射线检查预处理预测新辅助化疗的病理学完全缓解(英文)

2017-05-18 12:00:38 admin 9

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

Breast Cancer Res. 2017 May 18

Braman NM1, Etesami M2, Prasanna P3, Dubchuk C2, Gilmore H2, Tiwari P3, Plecha D2, Madabhushi A3.

Author information

    1.Department of Biomedical Engineering, Case Western Reserve University, Cleveland, OH, 44106, USA. nathaniel.braman@case.edu.

    2.University Hospitals Case Medical Center, Cleveland, OH, 44106, USA.

    3.Department of Biomedical Engineering, Case Western Reserve University, Cleveland, OH, 44106, USA.

Abstract

BACKGROUND


In this study, we evaluated the ability of radiomic textural analysis of intratumoral and peritumoral regions on pretreatmentbreast 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


慧影医疗科技(北京)有限公司

地点:北京市海淀区中关村东升科技园B2-C103

电话:400-890-9020

邮箱:radcloud@huiyihuiying.com

关闭
图片
图片
  • 人工智能诊断云平台