基于乳腺MRI组学及背景特征评估三阴性乳腺癌基因分型及疗效预测的研究

Triple-negative breast cancer (TNBC), as a special type of breast tumour, is identified different subtypes and associated with BRCA gene mutation. Since a short median survival time and poor prognosis of TNBC, there is essential clinical value on comprehensive assessment of early tumour response. Radiomics of MRI has been proved to predictive the early prognosis in advanced TNBC after neoadjuvant chemotherapy (NAC). However, some problem still has been reported, such as unable to assess subtype of TNBC and the BRCA gene mutation, not consider the effect of microenvironments and so on. Our previous study found that breast parenchymal enhancement (BPE) as one of imaging indicators to affect microenvironment of normal tissue can also predict the efficacy of tumour response. Furthermore, we hypothesize that BPE may also reflect the BRCA gene status according to the evaluation of breast epithelial blood supply. Thus, put forward the hypothesis that combined with radiomics of breast MRI and background features, can further assess TNBC biological heterogeneity of different molecular subtypes, evaluate BRCA mutations to guide the clinical decision at the same time, and improve the early prediction of tumour. response in TNBC. Therefore, this topic will use the quantitative analysis of both radiomics and BPE in breast MRI to combine with the tumour heterogeneity and tissue microenvironment. We hope to investigate the biological heterogeneity of different TNBC subtypes, assess the impact on BRCA mutations and homologous recombination deficient （HRD）mutations, thus build a comprehensive and noninvasive imaging platform to predict the whole biology characteristics and assess the efficacy of early prognosis after NAC of TNBC. in order to provide a theoretical basis for individualized precise treatment.

三阴性乳腺癌（TNBC）作为特殊类型乳腺肿瘤，中位生存期短、预后差，早期综合评估疗效具有重要的临床价值。已证实乳腺MRI组学可评估进展期TNBC早期疗效，但未涉及TNBC分型及BRCA突变，未考虑腺体微环境影响等。我们前期研究结果发现背景实质强化（BPE），作为腺体微环境影像指标，同样可以预测乳腺肿瘤疗效，且推测BPE对乳腺上皮血供评估可间接反映BRCA状态。由此提出假说：联合乳腺MRI组学及背景特征，可进一步评估TNBC分子亚型的生物学异质性，同时识别BRCA突变以指导临床决策，进而提高对TNBC早期疗效预测。因此本研究拟通过乳腺MRI组学及BPE的量化分析，结合肿瘤本身特质与腺体微环境状态，分析TNBC不同亚型的生物学异质性，关联BRCA及相应同源重组缺陷（HRD）突变对TNBC异质性的影响，进而构建全面无创评估TNBC整体生物学特征及疗效预测的影像平台，为个体化精准治疗提供理论依据。

三阴性乳腺癌（TNBC）作为特殊类型乳腺肿瘤，中位生存期短、预后差，早期综合评估疗效具有重要的临床价值。本研究构建包含乳腺MRI组学数据的TNBC多组学队列，共纳入202名TNBC女性患者。针对TNBC瘤体特质与瘤周微环境状态,联合瘤体及瘤周背景的MRI组学特征构建模型。本研究在实现乳腺、肿瘤及纤维腺体背景的自动化分割基础上，发现（1）MRI组学模型可有效鉴别TNBC亚型，且联合IHC可进一步提高预测效能；（2）MRI组学可无创预测TNBC关键靶点，包括与BRCA突变密切相关的同源重组缺陷（HRD）及与免疫治疗相关的肿瘤淋巴浸润细胞（Tils）水平；（3）影像组学模型具有显著的疗效及预后预测价值，且具有较高瘤周背景异质性提示预后不良。本研究构建的全面无创评估TNBC整体生物学特征及疗效预测的MRI影像平台，为TNBC个体化精准治疗提供理论依据。