宏-微观辨证结合，基于"中医临床表型"构建乳腺癌化疗风险预测数学模型

Though increasing evidences show that patients could benefit from chemotherapy, serious or fatal toxic effect of chemotherapy could lead to the suspension of medication, and even some patients refuse chemotherapy because they fear the toxicity of chemotherapy. The toxic effects of chemotherapy is fairly important in optimizing individual therapeutic regimen and patients often express fear of the side effects, so this study is devoted to researching on this important aspect of chemotherapy. This study is aimed to develop an accurate predictive model of chemotherapy risk to help clinicians inform breast cancer patients with risks of chemotherapy, and thus to help the informed patient to alleviate their fear and anxiety about toxicity of chemotherapy. Furthermore, this predictive model will be built to help clinicians to assess whether serious or fatal events will happen, and determine whether the chemotherapy regimen is appropriate to the patients. And then specific treatment will be given to possible adverse reactions of chemotherapy to achieve optimal individualized treatment program and to ensure smooth implementation of chemotherapy. Firstly, we'll extract the "macro" information about physical condition, which comes from syndrome-constitutional differentiation in the light of TCM theory. Then combined with the "micro" information about pathological and biological markers, "TCM clinical phenotype" is extracted by Text Mining, and shows the full picture of the clinical features of breast cancer, which will include most of risk factors related to occurrence of adverse effects. Secondly, qualification breast cancer patients will be recruited from large cancer centers in Xi'an, to take "TCM clinical phenotype" and adverse effects of chemotherapy of breast cancer patients to establish two corresponding quantitative databases. Furthermore, predictive models will be built according to the predictors selected from an analysis on new methods of econometrics- - Granger causality and cointegration analysis between "TCM clinical phenotype" and adverse effects of chemotherapy. The research results will be used to establish effective predictive models to lay the foundation for optimizing individualized treatment program of breast cancer by TCM and Western medicine.

一旦严重或致死性化疗毒性发生，可直接导致化疗中止；恐惧化疗毒性发生，甚至使部分患者放弃化疗。优化个体化化疗方案的关键是平衡化疗风险与获益。如能准确预测个体化疗风险，就可避免选风险高的方案，拟定针对性强的防治化疗毒性策略，还能减轻患者恐惧，提高化疗依从性。项目采用文本挖掘，引入中医"宏观"辨证、辨体质来丰富对患者机体状态的认识，再结合乳腺癌"微观"信息-相关病理及生物标志物，提取乳腺癌"中医临床表型"要素。据此设计调查表，并以之为工具，在乳腺癌患者中采用心理物理学等方法开展多中心临床研究，形成乳腺癌患者中医临床表型数据库；首次应用计量经济学新方法Granger因果分析及协整分析，探寻乳腺癌中医临床表型中与化疗不良反应发生相关的"因素"；再以之构建预测乳腺癌化疗风险协整模型。避免了现有模型对"未知但可能相关危险因素"无法认知局限。项目将为优化中西医乳腺癌个体化治疗方案，确保化疗顺利实施奠定基础

一旦严重或致死性化疗毒性发生，可直接导致化疗中止；恐惧化疗毒性发生，甚至使部分患者放弃化疗。优化个体化化疗方案的关键是平衡化疗风险与获益。如能准确预测个体化疗风险，就可避免选风险高的方案，拟定针对性强的防治化疗毒性策略，还能减轻患者恐惧，提高化疗依从性。本项目从宏-微观角度研制并考评能较全面涵盖乳腺癌临床特征的中医临床表型量表。该量表能较为客观全面地涵盖乳腺癌患者宏观-微观临床特征，并具有良好的信度、效度及较高的可行性。这为后续基于乳腺癌患者临床特征，全面探寻与乳腺癌化疗敏感、不良反应的发生、肿瘤复发转移等相关的危险因素的筛选提供基础，也为具有某一类特定特征乳腺癌的患者寻找更为特异性的基因表达及现代学指标提供研究基础。项目还发现应当加强阳虚质患者血常规检测，并应加强这类患者预防白细胞减少、血小板减少症的相关治疗，加强湿热证患者在消化道不良反应方面的防护与治疗。同时阳虚的患者发生严重便秘、腹泻、发热的风险明显降低，提示化疗可能具有热邪气性质。乳腺癌患者中医临床表型中：怕冷得分越高，临床分期越高发生白细胞减少的风险越高。情绪低沉、不耐冬天寒冷的得分越高，发生严重贫血风险越低，临床分期越高，p21阳性发生严重贫血风险越高。KPS分期越高、怕冷衣服、容易患感冒比别人穿得多，这儿痛那儿痛,得分越高，严重血小板减少风险越低。 面色晦暗得分越高，严重血小板减少风险越高。口苦得分越高、PR阳性越高发生严重恶心呕吐风险越高。临床分期越高、PR、PCNA阳性发生严重腹泻风险高；胃脘部、背部或腰膝部怕怜，以及感到怕冷、衣服比别人穿得多的得分越高，发生严重腹泻风险越低。临床分期越高、PR、PCNA阳性发生严重便秘风险高；胃脘部、背部或腰膝部怕冷得分越高，发生严重便秘风险越低。临床分期越高、湿热分值越高阳性发生严重发热风险高，不耐冬天或空调的寒冷、易患感冒的分值越高，发生严重发热可能性低。本项目探寻出了乳腺癌中医临床表型中与化疗不良反应发生相关的"因素"；并初步构建预测乳腺癌化疗风险协整模型。避免了现有模型对"未知但可能相关危险因素"无法认知局限，筛选出的危险因素在国内外均是首次发现。例如研究发现高临床分期是多种化疗后多种不良反应的危险因素。该项目将为优化中西医乳腺癌个体化治疗方案，确保化疗顺利实施奠定基础。