目的 应用人工智能驱动的大数据平台辅助预测洛索洛芬钠凝胶贴膏的降解产物，并建立HPLC法测定有关物质。方法 通过ASKCOS平台人工智能辅助预测洛索洛芬钠凝胶贴膏中可能产生的降解杂质，定性确证了杂质1（洛索洛芬甘油酯）、杂质2（洛索洛芬甘油缩酮）、杂质3（洛索洛芬薄荷醇酯）。采用OMNI Phantom C₁₈色谱柱（250 mm×4.6 mm，5 µm），流动相为pH 2.5磷酸水溶液–甲醇，采用梯度洗脱，检测波长222 nm，体积流量1.0 mL/min，进样体积20 µL。采用加校正因子的主成分自身对照法对杂质进行定量。结果 杂质1、2、3分别在0.05～15.48、0.13～15.48、0.10～15.26μg/mL线性良好，平均回收率分别为101.5%、100.9%、97.6%，RSD值分别为1.42%、2.2%、1.95%。结论 方法简便快速、定量结果准确，对建立洛索洛芬钠制剂中有关物质的定量方法具有参考意义。

Objective To assist in predicting degradation products in Loxoprofen Sodium Gel Patches using an integrating AI-driven big data platforms, and to establish an HPLC method for determining related substances. Methods Through the ASKCOS platform, artificial intelligence was used to predict potential degradation impurities in Loxoprofen Sodium Gel Patches. Impurity 1 (loxoprofen glyceride), impurity 2 (loxoprofen glycerol ketone), and impurity 3 (loxoprofen menthol ester) were qualitatively confirmed. Chromatographic separation was performed on OMNI Phantom C₁₈ column (250 mm × 4.6 mm, 5 µm), with the mobile phase consisting of phosphate buffer (pH 2.5) - methanol. Gradient elution was performed at a detection wavelength of 222 nm, a flow rate of 1.0 mL/min, and an injection volume of 20 µL. Quantification of the impurities was conducted using the principal component self-control method with correction factors. Results Impurities 1, 2, and 3 all exhibited good linearity within the ranges of 0.05 – 15.48, 0.13 – 15.48, and 0.10 – 15.26 μg/mL, respectively, with average recoveries of 101.5%, 100.9%, and 97.6%, and RSD values of 1.42%, 2.2%, and 1.95%. Conclusion The method is simple, rapid, and provides accurate quantitative results, offering reference value for establishing quantitative methods for related substances in loxoprofen sodium preparations.

R927.2