基于临床-影像学特征构建模型预测抗肿瘤坏死因子-α单克隆抗体维持治疗小肠克罗恩病的疗效

A predictive model based on clinical and computed tomography enterography features to evaluate maintenance efficacy of anti-tumor necrosis factor-αmonoclonal antibody in Crohn′s disease patients with small intestinal involvement

目的构建抗肿瘤坏死因子(TNF)-α单克隆抗体(简称单抗)维持治疗小肠受累的活动期克罗恩病(CD)的疗效预测模型。方法采用回顾性队列研究法,连续纳入2017年1月至2020年12月在同济大学附属第十人民医院住院的98例接受抗TNF-α单抗诱导治疗有效且治疗期间随访达1年的小肠受累的CD患者,所有患者在治疗前接受计算机断层扫描小肠成像(CTE)检查。根据维持期是否发生药物治疗方案调整,将患者分为药物优化组和药物维持组,单因素分析比较两组间临床及影像特征的差异,多因素Logistic回归分析筛选需行药物优化的危险因素,建立预测模型,绘制列线图并行内部验证。结果药物优化组患者40例,药物维持组患者58例。单因素分析发现,与药物维持组比较,药物优化组发病年龄更大[(35.7±14.3)岁比(29.6±12.3)岁,P=0.027],诊断年龄更大[(37.7±17.8)岁比(30.6±11.1)岁,P=0.006],血红蛋白水平更低[(112.9±23.2)g/L比(126.9±26.5)g/L,P=0.008],血白蛋白水平更低[(38.1±5.0)g/L比(42.5±4.9)g/L,P<0.001],而两组间肠壁强化程度(轻度:40.0%比74.1%,中重度:60.0%比25.9%,P=0.001)以及肠腔狭窄(无或可疑:47.5%比87.9%,轻度:17.5%比5.2%,中重度:35.0%比6.9%,P<0.001)的差异均有统计学意义。多因素Logistic分析发现,诊断年龄(OR=1.051,95%CI:1.009~1.096,P=0.018)、肠壁强化程度(OR=3.807,95%CI:1.268~11.428,P=0.017)和肠腔中重度狭窄(OR=6.550,95%CI:1.640~26.165,P=0.008)是需行药物优化的危险因素,而血白蛋白水平为保护因素(OR=0.841,95%CI:0.747~0.946,P=0.004)。所构建预测模型受试者操作曲线下面积为0.856(95%CI:0.779~0.933,P<0.001),灵敏度为82.5%,特异度81.0%,准确性95.9%。结论基于CTE提示的肠腔狭窄程度和肠壁强化程度联合诊断年龄和血白蛋白水平构建模型,可预测抗TNF-α单抗治疗小肠受累CD患者维持期药物的疗效.

Objective To establish a model to predict the maintenance efficacy of anti-tumor necrosis factor(TNF)-αmonoclonal antibody in active Crohn′s disease(CD)patients with small intestinal involvement.Methods A retrospective cohort study was carried out.Ninety-eight CD patients with small intestinal involvement admitted in the Tenth People′s Hospital of Tongji University from January 2017 to December 2020 were consecutively included.All the patients received anti-TNF-αmonoclonal antibody induction therapy regularly and the induced remission treatment was effective.The maintenance therapy was followed up for at least 1 year.All patients underwent computed tomography enterography(CTE)before treatment.According to whether therapeutic optimization occurred,the patients were divided into optimization group and maintenance group.Univariate analysis was used to compare the differences of clinical and CTE features between the two groups,and multivariate Logistic regression analysis was performed to select the independent factors for medication optimization.A nomogram based on the established predictive model was drawn and further validated internally.Results During the follow-up,40 patients underwent treatment optimization and 58 maintained the original treatment.Univariate analysis showed that compared with the maintenance group,the ages of onset[(35.7±14.3)years old vs.(29.6±12.3)years old,P=0.027)]and diagnosis[(37.7±17.8)years old vs.(30.6±11.1)years old,P=0.006)]were older,the level of hemoglobin[(112.9±23.2)g/L vs.(126.9±26.5)g/L,P=0.008]and serum albumin[(38.1±5.0)g/L vs.(42.5±4.9)g/L,P<0.001]was lower in the optimization group,meanwhile the degree of intestinal wall enhancement(mild:40.0%vs.74.1%,moderate and severe:60.0%vs.25.9%,P=0.001)and intestinal stenosis(no or suspicious:47.5%vs.87.9%,mild:17.5%vs.5.2%,moderate and severe:35.0%vs.6.9%,P<0.001)were significantly different between two groups.Multivariate analysis revealed that age at diagnosis(OR=1.051,95%CI:1.009-1.096,P=0.018),degree of intestinal wall enhancement(OR=3.807,95%CI:1.268-11.428,P=0.017)and moderate-severe intestinal stenosis(OR=6.550,95%CI:1.640-26.165,P=0.008)were the independent risk factors for treatment optimization,and high serum albumin level(OR=0.841,95%CI:0.747-0.946,P=0.004)was the protective factor.The area under ROC of the established predictive model was 0.856(95%CI:0.779-0.933,P<0.001)with sensitivity of 82.5%,specificity of 81.0%,and accuracy of 95.9%.Conclusion A model is established based on the CTE features including the degree of intestinal wall enhancement and intestinal stenosis combined with age at diagnosis and serum albumin level,it can predict the efficacy of anti-TNF-αmonoclonal antibody in maintenance therapy among CD patients with small intestinal involvement.