Page 80 - 《中国药房》2025年17期
P. 80

PD-1抑制剂用于晚期食管鳞状细胞癌的疗效预测模型构建                                                                Δ



                                                   1
                           1
                                           1
                                   1
                                                           2
                 1*
          吴珊珊 ,黄晓杰 ,谢晓纯 ,黄少楷 ,黄丽娜 ,王晓芬(1.揭阳市人民医院药学部,广东 揭阳 522000;2.揭阳
          市人民医院肿瘤内科,广东 揭阳 522000)
          中图分类号  R969;R979.1      文献标志码  A      文章编号  1001-0408(2025)17-2154-06
          DOI  10.6039/j.issn.1001-0408.2025.17.12
          摘   要  目的  构建晚期食管鳞状细胞癌(ESCC)患者接受程序性死亡受体1(PD-1)抑制剂治疗获得持久临床获益(DCB)的预测
          模型。方法  回顾性收集2020年1月至2023年12月在揭阳市人民医院接受PD-1抑制剂治疗的晚期ESCC患者的临床资料,通过
          最小绝对收缩和选择算子(Lasso)回归模型筛选预测变量,建立多因素Logistic回归模型预测患者DCB,并基于该模型绘制列线
          图。通过Bootstrap法对预测模型进行内部验证,采用受试者操作特征曲线、校准曲线、决策曲线分析对预测模型进行评价。结果
          共纳入91例晚期ESCC患者。Lasso回归与Logistic回归联合分析的结果表明,基线淋巴细胞/单核细胞比值(LMR)[比值比(OR)
          为 1.97,95% 置信区间(CI)为 1.15~3.36,P=0.013]、白蛋白(ALB)含量(OR=1.35,95%CI 为 1.13~1.60,P<0.001)、体重指数
         (BMI)1(正常 vs. 低:OR=0.28,95%CI 为 0.09~0.96,P=0.042)、BMI2(超重~肥胖 vs. 低:OR=0.08,95%CI 为 0.01~0.59,P=
          0.013)及治疗方案(免疫单药vs.免疫联合其他:OR=0.07,95%CI为0.01~0.50,P=0.008)是晚期ESCC患者接受PD-1抑制剂获
          得DCB的预测因素。基于上述指标构建预测模型,经Bootstrap法内部验证,其曲线下面积为0.831(95%CI为0.746~0.904),特异
          度为74.4%,灵敏度为75.0%。Hosmer-Lemeshow拟合优度检验的χ =9.930,P=0.270,校准曲线的斜率接近1。决策曲线分析结
                                                              2
          果显示该模型的风险阈值范围在0.1~1.0时具有良好的临床应用价值。结论  通过基线LMR、ALB含量、BMI和治疗方案成功构
          建了预测模型,其在评估晚期ESCC患者接受PD-1抑制剂治疗效果上具有良好的预测能力和临床实用性。
          关键词  食管鳞状细胞癌;PD-1抑制剂;持久临床获益;疗效;预测模型;Lasso回归;多因素Logistic回归;列线图

          Construction  of  a  prediction  efficacy  model  for  PD-1  inhibitor  in  advanced  esophageal  squamous  cell
          carcinoma
                                                                      1
                                                     1
                                       1
                                                                                     1
                                                                                                    2
                       1
          WU Shanshan ,HUANG Xiaojie ,XIE Xiaochun ,HUANG Shaokai ,HUANG Lina ,WANG Xiaofen(1.  Dept.
          of  Pharmacy,  Jieyang  People’s  Hospital,  Guangdong  Jieyang  522000,  China;2.  Dept.  of  Oncology,  Jieyang
          People’s Hospital, Guangdong Jieyang 522000, China)
          ABSTRACT    OBJECTIVE  To  develop  a  prediction  model  for  durable  clinical  benefit (DCB)  in  patients  with  advanced
          esophageal  squamous  cell  carcinoma (ESCC)  receiving  programmed  death-1 (PD-1)  inhibitor.  METHODS  The  clinical  data  of
          patients  with  advanced  ESCC  who  received  PD-1  inhibitor  in  Jieyang  People’s  Hospital  were  retrospectively  collected  between
          January  2020  to  December  2023.  Predictors  were  screened  by  least  absolute  shrinkage  and  selection  operator (Lasso)  regression,
          and  a  multivariable  Logistic  regression  model  was  developed  to  predict  DCB. A  nomogram  was  constructed  based  on  the  model.
          Internal  validation  of  the  prediction  model  was  performed  by  using  the  Bootstrap  method,  and  the  model  was  evaluated  by  the
          receiver  operating  characteristic (ROC)  curve,  calibration  curve,  and  decision  curve  analysis.  RESULTS  A  total  of  91  patients
          with  advanced  ESCC  were  included.  The  results  of  Lasso  regression  combined  with  Logistic  regression  analysis  indicated  that  the
          baseline  lymphocyte  monocyte  ratio (LMR)  [odds  ratio (OR)=1.97,  95%  confidence  interval (CI):  1.15-3.36,  P=0.013],
          albumin (ALB)  content (OR=1.35,  95%CI:  1.13-1.60,  P<0.001),  body  mass  index (BMI)  category  1  [normal  vs.  low:  OR=
          0.28,  95%CI (0.09-0.96),  P=0.042],  BMI  category  2  [overweight-obesity  vs.  low:  OR=0.08,  95%CI (0.01-0.59),  P=0.013],
          and  treatment  regimen  [monotherapy  vs.  monotherapy  combination  therapy:  OR=0.07,  95%CI (0.01-0.50),  P=0.008]  were
          predictive  factors  for  patients  with  advanced  ESCC  to  achieve  DCB  when  treated  with  PD-1  inhibitor.  A  prediction  model  was
          constructed  based  on  the  above  indicators.  Internal  validation  of  the  model  using  the  Bootstrap  method  showed  an  area  under  the
          curve of 0.831 (95%CI: 0.746-0.904), with specificity of 74.4% and sensitivity of 75.0%. The Hosmer-Lemeshow test yielded χ =
                                                                                                           2
          9.930, P=0.270, and the calibration curve slope was close to 1. The decision curve analysis demonstrated that the model exhibited
          good  clinical  utility  within  a  threshold  range  of  0.1  to  1.0.  CONCLUSIONS  The  prediction  model  based  on  baseline  LMR, ALB
          content,  BMI,  and  treatment  regimen  demonstrates  robust  predictive  performance  and  clinical  utility  for  assessing  therapeutic
          efficacy of PD-1 inhibitor in the treatment of advanced ESCC.
                                                              KEYWORDS    esophageal  squamous  cell  carcinoma;  PD-1
              Δ 基金项目 揭阳市卫生医疗科技创新项目(No.ylws2024043)            inhibitor;  durable  clinical  benefit;  therapeutic  efficacy;
             *第一作者 副主任药师,硕士。研究方向:临床药学。E-mail:                 prediction  model;  Lasso  regression;  multivariable  Logistic
          ss8311@163.com                                      regression; nomogram


          · 2154 ·    China Pharmacy  2025 Vol. 36  No. 17                            中国药房  2025年第36卷第17期
   75   76   77   78   79   80   81   82   83   84   85