基于胸部CT体成分分析联合酰胺质子转移加权成像预测急性缺血性卒中患者功能预后的价值OA
The value of chest CT-derived body composition analysis combined with amide proton transfer-weighted imaging in predicting functional outcomes in patients with acute ischemic stroke
目的 探索基于胸部CT体成分分析联合酰胺质子转移加权(amide proton transfer-weighted,APTw)成像在评估急性缺血性脑卒中(acute ischemic stroke,AIS)患者功能预后的价值.材料与方法 前瞻性纳入108例AIS患者(年龄62.96±9.55岁),所有患者行头部APTw成像和胸部CT扫描.依据90天改良Rankin量表(modified Rankin scale,mRS)评分评估AIS患者功能结局,分为预后良好组(mRS≤2)58例和预后不良组(mRS>2)50例.基于胸部平扫CT图像进行体质成分分析:包括心外膜脂肪组织(epicardial adipose tissue,EAT)、第 12 胸椎层面内脏脂肪(visceral adipose tissue at the 12th thoracic vertebra level,VAT_T12)、第12胸椎层面皮下脂肪(subcutaneous adipose tissue at the 12th thoracic vertebra level,SAT_T12)及第12胸椎层面骨骼肌(skeletal muscle at the 12th thoracic vertebra level,SM_T12).基于APTw图像的梗死区和对侧正常白质区域定量参数[APT值的最大值(APTwmax_病灶和APTwmax_健侧)、最小值(APTwmin_病灶和APTwmin_健侧)、平均值(APTwmean_病灶和APTwmean_健侧)及APTwmax-min值]评估.通过独立样本t检验或Mann-Whitney U评估两组定量参数差异性.经单因素与多因素logistic回归分析筛选与AIS患者功能预后相关的独立预测因子,并构建列线图模型.模型的预测性能通过受试者工作特征(receiver operating characteristic,ROC)曲线进行评估,计算曲线下面积(area under the curve,AUC)值、准确度、敏感度和特异度.采用DeLong检验比较不同模型间AUC值的差异.校准曲线用于评估预测概率与实际概率之间的偏差,Hosmer-Lemeshow检验用于评估模型拟合优度.通过决策曲线分析(decision curve analysis,DCA)评估模型的临床实用性.采用沙普利加法解释(SHapley additive explanations,SHAP)列线图模型中各特征的重要性.结果 本研究最终按90天随访mRS评分分为预后良好组(58例,mRS:0~2)与不良组(50例,mRS:3~6).除入院时mRS(t=3.782,P<0.001)及美国国立卫生研究院卒中量表(National Institutes of Health Stroke Scale,NIHSS)评分(t=2.743,P=0.006)外,两组其他临床特征无显著差异.体质分析发现预后不良组较预后良好组患者EAT体积(epicardial adipose tissue volume,EATV)和第12胸椎层面内脏脂肪面积(visceral adipose tissue area at the 12th thoracic vertebra level,VATA_T12)升高(t=4.948、3.765;P<0.001),而EAT衰减值、VAT_T12衰减值及SM_T12衰减值降低(t=-3.025,P=0.002;t=-2.434,P=0.015;t=-2.350,P=0.021).基于APTw序列的定量参数分析显示预后不良组患者APTwmin_病灶值(t=-3.036,P=0.002)降低,而APTwmax-min_病灶侧值(t=2.365,P=0.018)升高.经多重共线性分析[方差膨胀因子(variance inflation factor,VIF)<5]筛选后,多因素logistic回归显示入院时mRS评分[OR=2.726(95%CI:1.347~5.517),P=0.005]、EATV[OR=1.019(95%CI:1.002~1.036),P=0.026]及APTwmin_病灶侧[OR=0.236(95%CI:0.084~0.662),P=0.006]为不良预后独立预测因子,据此构建列线图模型.各独立预测因子及列线图模型评估AIS患者功能预后的AUC值分别为0.704、0.777、0.670及0.873.校准曲线(Hosmer-Lemeshow检验X-squared=10.533,P=0.230)及DeLong检验(P<0.05)证实列线图模型预测概率与实际高度一致且效能优于单因素.DCA曲线显示在所有的阈值范围内列线图模型净临床获益最优,SHAP分析显示独立预测因子对列线图模型的贡献行排序依次为EATV、APTwmin_病灶及入院时mRS评分.结论 基于胸部CT体质成分分析联合APTw序列,可从脑局部损伤特征和全身储备状态两个维度综合评估AIS患者功能预后,为早期、精准识别高危不良预后患者提供更为可靠的多维度影像学依据,从而实现AIS患者风险分层,为AIS患者的健康管理提供依据.
Objective:To explore the value of chest CT-derived body composition analysis combined with amide proton transfer-weighted(APTw)imaging in assessing the functional prognosis of patients with acute ischemic stroke(AIS).Materials and Methods:This prospective study enrolled 108 AIS patients(age 62.96±9.55 years).All patients underwent both head APTw imaging and chest CT scans.Functional outcomes were assessed using the 90-day modified Rankin scale(mRS)and patients were dichotomized into a favorable prognosis group(mRS≤2,n=58)and an unfavorable prognosis group(mRS>2,n=50).Body composition analysis based on non-contrast chest CT images included epicardial adipose tissue(EAT),visceral adipose tissue at the 12th thoracic vertebra level(VAT_T12),subcutaneous adipose tissue at the 12th thoracic vertebra level(SAT_T12),and skeletal muscle at the 12th thoracic vertebra level(SM_T12).Quantitative parameters derived from the APTw images[including the maximum values of APT(APTwmax_lesion and APTwmax_contralateral side),minimum values(APTwmin_lesion and APTwmin_contralateral side),average values(APTwmean_lesion and APTwmean_contralateral side),and the APTwmax-min value in the infarct core and contralateral normal white matter]were assessed.Differences in quantitative parameters between groups were evaluated using independent samples t-tests or the Mann-Whitney U test.Univariate and multivariate logistic regression analyses were performed to screen for independent predictors associated with AIS functional prognosis,and a nomogram model was constructed accordingly.The predictive performance of the model was assessed using the receiver operating characteristic(ROC)curve,calculating the area under the curve(AUC),accuracy,sensitivity,and specificity.Differences in AUC values between models were compared using DeLong's test.Calibration curves were used to evaluate the deviation between predicted and actual probabilities,and the Hosmer-Lemeshow test was employed to assess model goodness-of-fit(P>0.05 indicating good fit).The clinical utility of the model was evaluated using decision curve analysis(DCA).The importance of each feature within the nomogram model was interpreted using SHapley additive explanations(SHAP).Results:Based on 90-day mRS scores,patients were categorized into the favorable prognosis group(n=58,mRS:0-2)and the unfavorable prognosis group(n=50,mRS:3-6).Apart from admission mRS(t=3.782,P<0.001)and National Institutes of Health Stroke Scale(NIHSS)(t=2.743,P=0.006),other clinical characteristics showed no significant differences between groups.Body composition analysis revealed that patients in the unfavorable prognosis group had significantly higher epicardial adipose tissue volume(EATV)and visceral adipose tissue area at the 12th thoracic vertebra level(VAT_T12)(t=4.948,3.765,P<0.001),but lower attenuation values for EAT,VAT_T12,and SM_T12(t=-3.025,P=0.002;t=-2.434,P=0.015;t=-2.350,P=0.021,respectively)compared to the favorable prognosis group.Analysis of APTw-derived quantitative parameters showed significantly lower APTwmin_lesion(t=-3.036,P=0.002)and higher APTwmax-min_lesion(t=2.365,P=0.018)values in the unfavorable prognosis group.After screening for multicollinearity[variance inflation factor(VIF)<5],multivariate logistic regression identified admission mRS score[OR=2.726(95%CI:1.347 to 5.517),P=0.005],EATV[OR=1.019(95%CI:1.002 to 1.036),P=0.026],and APTwmin_lesion[OR=0.236(95%CI:0.084 to 0.662),P=0.006]as independent predictors of unfavorable prognosis.A nomogram model was subsequently constructed.The AUC values for predicting AIS functional prognosis were 0.704,0.777,0.670,and 0.873 for the individual independent predictors and the combined nomogram model,respectively.The calibration curve(Hosmer-Lemeshow test X-squared=10.533,P=0.230)and DeLong's test(P<0.05)confirmed that the nomogram model's predicted probabilities were highly consistent with the actual outcomes and that its performance was superior to that of individual predictors.The DCA curve demonstrated that the nomogram model provided the optimal net clinical benefit across all threshold probabilities.SHAP analysis revealed that the order of feature contribution to the nomogram model was EATV,followed by APTwmin_lesion,and then admission mRS score.Conclusions:The combination of chest CT-derived body composition analysis and APTw imaging provides a comprehensive assessment of AIS functional prognosis from dual perspectives:local brain injury characteristics and systemic metabolic reserve status.This approach offers a more reliable multi-dimensional imaging basis for the early and precise identification of patients at high risk for unfavorable outcomes,thereby providing a basis for risk stratification and health management in AIS patients.
马长军;王家正;张钦和;黄依萍;刘嘉慧;王秀林;李颖;丛丰裕;刘爱连;刘晶
大连理工大学医学部,大连 116024||大连医科大学附属第一医院干细胞临床研究中心,大连 116011||大连干细胞与精准医学创新研究院,大连 116023飞利浦(中国)投资有限公司北京分公司,北京 100016大连医科大学附属第一医院干细胞临床研究中心,大连 116011||大连干细胞与精准医学创新研究院,大连 116023||大连医科大学附属第一医院放射科,大连 116011||大连市医学影像人工智能工程技术研究中心,大连 116011大连医科大学附属第一医院干细胞临床研究中心,大连 116011||大连干细胞与精准医学创新研究院,大连 116023大连医科大学附属第一医院干细胞临床研究中心,大连 116011||大连干细胞与精准医学创新研究院,大连 116023大连医科大学附属第一医院干细胞临床研究中心,大连 116011||大连干细胞与精准医学创新研究院,大连 116023大连医科大学附属第一医院干细胞临床研究中心,大连 116011||大连干细胞与精准医学创新研究院,大连 116023大连理工大学医学部,大连 116024大连医科大学附属第一医院放射科,大连 116011||大连市医学影像人工智能工程技术研究中心,大连 116011大连医科大学附属第一医院干细胞临床研究中心,大连 116011||大连干细胞与精准医学创新研究院,大连 116023
医药卫生
急性缺血性脑卒中体质成分分析定量计算机断层扫描磁共振成像酰胺质子转移加权成像功能预后
acute ischemic strokebody composition analysisquantitative computed tomographymagnetic resonance imagingamide proton transfer-weighted imagingfunctional outcome
《磁共振成像》 2026 (9)
5-15,11
辽宁省科技计划联合计划项目(编号:2025-MSLH-199)大连市科技人才创新支持项目(编号:2023RY034)Joint Research Program of Liaoning Provincial Science and Technology Plan(No.2025-MSLH-199)Dalian Science and Technology Talent Innovation Support Project(No.2023RY034).
评论