目的 针对现行红花Carthami Flos质量评价中体外指标与活血功效关联不足以及性状鉴别依赖主观经验的2大瓶颈，基于质量标志物（quality marker，Q-Marker）理论与电子感官技术，构建红花“内在药效物质-外在感官品质”一体化评价体系。方法 采用UPLC-Q-TOF-MS技术鉴定红花化学成分及入血成分；通过对比正常和急性血瘀模型大鼠药动学行为差异，筛选Q-Marker。建立UPLC-MS/MS多指标含量测定方法，测定26批不同产地红花中Q-Marker的含量。进一步利用电子眼与电子鼻分别获取色泽及气味数字化参数，结合多元统计分析，通过Spearman相关分析与偏最小二乘回归（partial least squares regression，PLSR）解析Q-Marker含量与感官特征的关联。结果 共解析出红花药材中的16种化学成分，其中羟基红花黄色素A、咖啡酸、对香豆酸、阿魏酸、山柰酚-3-O-芸香糖苷、山柰酚6种为入血成分；与正常组相比，急性血瘀模型大鼠体内咖啡酸、山柰酚的曲线下面积（area under curve，AUC）显著升高，山柰酚-3-O-芸香糖苷在病理状态下呈现更明显的脱糖代谢倾向，提示活血效应成分在血瘀病理状态下暴露增强。综合Q-Marker“特有性-可测性-有效性-传递与溯源”的原则，确定羟基红花黄色素A、咖啡酸、对香豆酸、阿魏酸、山柰酚-3-O-芸香糖苷、山柰酚6个成分为红花活血的Q-Marker。基于电子感官参数的主成分分析（principal component analysis，PCA）显示，电子眼与电子鼻可有效区分红花品质优劣。Spearman与PLSR相关分析表明，电子眼色泽参数L^*、b^*及电子鼻传感器W6S、W3S信号与Q-Marker含量呈显著相关，可作为表征内在品质的关键指标。结论 将病理状态药动学差异引入红花Q-Marker筛选，据此建立的Q-Marker-电子感官联用评价体系，实现了红花“化学物质组-药效暴露-感官表型”的三维关联，为红花药材品质快速无损的电子感官鉴别提供了科学依据和新的研究范式。

Objective To address the current bottlenecks in the quality evaluation of Honghua (Carthami Flos, CF), namely the insufficient correlation between in vitro indicators and blood-activating efficacy, and the reliance on subjective experience for morphological identification, this study constructed an integrated evaluation system of “intrinsic pharmacodynamic substances–extrinsic sensory quality” based on the quality marker (Q-Marker) theory and electronic sensory technologies. Methods UPLC-Q-TOF-MS was employed to identify the chemical constituents and blood-absorbed components of CF. Candidate Q-Markers related to blood-activating efficacy were screened by comparing the pharmacokinetic differences between normal rats and rats with acute blood-stasis syndrome (ABS). A UPLC-MS/MS multi-component quantification method was established to determine the contents of six Q-Markers in 26 batches of CF from different origins. Furthermore, an electronic eye and an electronic nose were utilized to obtain digital color parameters and odor profiles. Multivariate statistical analyses were performed to distinguish genuine and inferior samples, and Spearman correlation analysis combined with partial least squares regression (PLSR) was conducted to elucidate the quantitative relationships between Q-Marker contents and sensory features. Results Sixteen chemical constituents were identified in CF, of which hydroxysafflor yellow A, caffeic acid, p-coumaric acid, ferulic acid, kaempferol-3-O-rutinoside, and kaempferol were blood-absorbed constituents. Compared with the normal group, the area under the curve (AUC) of caffeic acid and kaempferol in the ABS model rats was significantly increased, and kaempferol-3-O-rutinoside showed a more obvious tendency of deglycosylation metabolism in the pathological state, suggesting that the exposure of blood-activating components was enhanced in the pathological state of blood stasis. Based on the Q-Marker principles of “specificity, measurability, effectiveness, and transmission and traceability”, six components including hydroxysafflor yellow A, caffeic acid, p-coumaric acid, ferulic acid, kaempferol-3-O-rutinoside and kaempferol were determined as Q-Markers for the blood-activating efficacy of CF. Principal component analysis (PCA) based on electronic sensory parameters demonstrated that the electronic eye and electronic nose can effectively distinguish high-quality and inferior-quality CF. Spearman and PLSR correlation analysis showed that the electronic eye color parameters L^*, b^* and the signals from the electronic nose sensors W6S and W3S were significantly correlated with Q-Marker content, which could be used as key indicators to characterize the internal quality. Conclusion In this study, the difference of pharmacokinetics in pathological state was introduced into the screening of CF Q-Marker. Based on this, the Q-Marker-electronic sensory evaluation system was established, which realized the three-dimensional correlation of safflower "chemical substance group-pharmacodynamic exposure-sensory phenotype”, and provided a scientific basis and a new research paradigm for the rapid and non-destructive electronic sensory identification of CF quality.

R283.6