乳腺癌居女性恶性肿瘤发病率和死亡率的首位。癌细胞无限增殖，进而导致的侵袭转移，是肿瘤致死的根本原因。Ki67抗原是最常使用的增殖标记物，在乳腺癌诊疗中起重要作用。当前Ki67计数方法存在主观性强、定量不准确、判读结果变异大等缺点，不利于乳腺癌精准治疗。申请人在前期研究中，建立了量子点多光谱成像和人工定义计算机识别技术，实现了乳腺癌Ki67的定量评估和多分子的协同研究。本课题拟联合量子点多光谱成像和计算机智能识别技术，建立乳腺癌Ki67自动化评估体系，通过构建具有优良光学性能的精准量子点探针，建立多分子量子点实时多色成像，特异性标Ki67评估相关的多个靶标，同时提取多靶标的光谱信息，开发计算机智能识别技术，识别和分析上述靶标的时空表达、量化表达和交互表达，最后研究乳腺癌Ki67表达水平、表达模式、表达部位与患者预后关系，建立客观、准确、多维的Ki67评估体系，为乳腺癌精准治疗提供决策依据。

Breast cancer (BC) is the leading malignant cancer for tumor incidence and mortality of females. The causes of cancer-related deaths are high cancer cells proliferation, migration and metastasis. As the most common used proliferation marker, Ki67 plays key in BC prognosis. However, high subjectivity, inaccurate counting and large variation in the currently Ki67 counting method are huge challenges for BC precision medicine. Quantum dots (QDs)-based multispectral imaging have been established in our previous studies to research quantified evaluation of Ki67 and co-expressions of multiple molecules in BC. This study is aimed to establish an automatic Ki67 counting system by combining QDs-based multispectral imaging with computer intelligent recognition. Firstly, QDs-based multispectral in-situ, real-time and quantify imaging will be established by the constructed QDs with excellent optical properties to specifically label multiple molecules in BC tissues. The multispectral analysis software is used to above spectral information of multiple molecules. Then, the computer intelligent recognition technology is developed to high throughput, objectively and accurately recognize and analysis space-time expressions, quantification expressions and co-expressions of above multiple molecules. Finally, by analyzing the association between expression levels, patterns and locations of Ki67 and prognosis of BC patients, the objective, accurate and multidimensional Ki67 evaluation system is established to provide decision evidences for BC precision medicine.