目的 建立非小細胞肺癌（NSCLC）常用靶向藥物的用量預(yù)測模型，為指導(dǎo)醫(yī)療機構(gòu)抗腫瘤靶向藥物的采購和庫存管理提供數(shù)據(jù)支撐。方法 參考天津市腫瘤醫(yī)院2019年各月的靶向藥物用量數(shù)據(jù)，以4種臨床常用的NSCLC靶向藥物（吉非替尼、?？颂婺?、奧希替尼和克唑替尼）為例，建立多元回歸模型、GM（1，1）灰色模型以及多元回歸-灰色組合模型，并對3種預(yù)測模型進行評價和驗證。結(jié)果 多元回歸模型在描述靶向藥物用量的波動變化方面具有優(yōu)勢，灰色模型可以更好地描述靶向藥物用量的增長趨勢，而組合模型兼?zhèn)涿枋鲇昧孔兓厔莺投唐诓▌拥哪芰Α＜翘婺?、埃克替尼、奧希替尼和克唑替尼4種靶向藥物運用組合模型得到的的預(yù)測值與實際值誤差分別為4.30%、2.87%、3.62%、4.42%。結(jié)論 多元回歸-灰色組合模型運行良好，與單一模型相比表現(xiàn)出更高的精準度，可應(yīng)用于醫(yī)療機構(gòu)靶向藥物的用量預(yù)測，從而實現(xiàn)靶向藥物的精準化管理。

Objective To establish a prediction model of the commonly used targeted drugs for non-small cell lung cancer (NSCLC), providing data support for the procurement and management of targeted anti-tumor drugs in medical institutions. Methods The multiple regression model, gray model and the combined model were established with four commonly used targeted drugs for non-small cell lung cancer (gefitinib, icotinib, osimertinib and crizotinib), based on monthly usage of Tianjin Medical University Cancer Institute and Hospital in 2019, and the prediction models were evaluated and verified. Results The advantage of the multiple regression model is to describe the fluctuation of the usage of the targeted drugs. The gray model can reflect the increasing trend, while the combined model has the characteristics of describing the increasing trend and short-term fluctuations. The error between the predicted value and the actual value of the combined prediction model for gefitinib, icotinib, osimertinib and crizotinib was 4.30%, 2.87%, 3.62% and 4.42%, respectively. Conclusion Compared with the single model, the multiple regression-gray combined prediction model works better and shows higher accuracy. It can be applied to the prediction of the usage of targeted drugs in medical institutions, which is conducive to the accurate management for targeted drugs.

R979.1

國家自然科學(xué)基金青年基金資助項目（81703454）