目的:构建预测头颈部腺样囊性癌(ACCHN)患者生存情况的列线图。方法:选择2000年至2017年SEER数据库的2 108例ACCHN患者作为建模队列,通过Cox回归分析筛选影响患者预后的独立因素,构建预测患者3 a、5 a总生存率(OS)和癌症特异性生存率(C...目的:构建预测头颈部腺样囊性癌(ACCHN)患者生存情况的列线图。方法:选择2000年至2017年SEER数据库的2 108例ACCHN患者作为建模队列,通过Cox回归分析筛选影响患者预后的独立因素,构建预测患者3 a、5 a总生存率(OS)和癌症特异性生存率(CSS)的列线图。选取2000年至2017年确诊的149例ACCHN患者作为外部验证队列,采用一致性指数(C指数)、ROC曲线和校准曲线评估列线图的性能。结果:年龄、性别、单原发癌、手术、化疗、原发部位、T分期、N分期、M分期是OS的影响因素。年龄、单原发癌、手术、化疗、原发部位、AJCC分期、T分期、N分期、M分期是CSS的影响因素。以上特征均整合在预测3 a和5 a OS和CSS的列线图中,内部验证C指数(95%CI)分别为0.770(0.754~0.786)和0.602(0.584~0.620),外部验证C指(95%CI)分别为0.796(0.795~0.797)和0.781(0.780~0.782)。内部验证中预测3 a和5 a OS的AUC(95%CI)分别为0.826(0.803~0.849)和0.814(0.792~0.835),预测3 a和5 a CSS的AUC(95%CI)分别为0.845(0.820~0.871)和0.834(0.811~0.857);外部验证中预测3 a和5 a OS的AUC(95%CI)分别为0.855(0.779~0.932)和0.838(0.768~0.909),预测3 a和5 a CSS的AUC(95%CI)分别为0.806(0.701~0.911)和0.806(0.726~0.886)。结论:本研究所构建的列线图可以准确预测ACCHN患者的OS和CSS,有助于个性化的预后评估和临床决策。展开更多
Four-wheel independent steering(4 WIS) system and direct yaw moment control(DYC) have an important influence on vehicle lateral stability. However, DYC has a great effect on the longitudinal velocity, and the capabili...Four-wheel independent steering(4 WIS) system and direct yaw moment control(DYC) have an important influence on vehicle lateral stability. However, DYC has a great effect on the longitudinal velocity, and the capability of 4 WIS is limited to stability.To decrease the influence on the longitudinal velocity and improve the stability of electrical vehicles, a chassis controller integrated with a 4 WIS system and a DYC system with model predictive control(MPC) is designed. The framework consists of an unscented Kalman filter(UKF) observer and an MPC that contains three blocks: supervisor blocks, upper blocks and lower blocks. First, the sideslip angle, longitudinal velocity and lateral tire forces are estimated by the UKF observer;second, a bicycle model is utilized in the supervisor to calculate the desired values;third, the upper blocks are designed with the MPC to optimize the target steering angles and longitudinal tire forces under the constraints of subsystems;to facilitate the design of the MPC, a nonlinear tire is simplified based on the Taylor expansion method;finally, the target steering angles and longitudinal tire forces are achieved by the lower blocks. The integrated controller is simulated on the co-simulation platform of MATLAB-Carsim. The results show that the proposed integrated controller has less impact on longitudinal velocity and could effectively improve vehicle stability.展开更多
文摘目的:构建预测头颈部腺样囊性癌(ACCHN)患者生存情况的列线图。方法:选择2000年至2017年SEER数据库的2 108例ACCHN患者作为建模队列,通过Cox回归分析筛选影响患者预后的独立因素,构建预测患者3 a、5 a总生存率(OS)和癌症特异性生存率(CSS)的列线图。选取2000年至2017年确诊的149例ACCHN患者作为外部验证队列,采用一致性指数(C指数)、ROC曲线和校准曲线评估列线图的性能。结果:年龄、性别、单原发癌、手术、化疗、原发部位、T分期、N分期、M分期是OS的影响因素。年龄、单原发癌、手术、化疗、原发部位、AJCC分期、T分期、N分期、M分期是CSS的影响因素。以上特征均整合在预测3 a和5 a OS和CSS的列线图中,内部验证C指数(95%CI)分别为0.770(0.754~0.786)和0.602(0.584~0.620),外部验证C指(95%CI)分别为0.796(0.795~0.797)和0.781(0.780~0.782)。内部验证中预测3 a和5 a OS的AUC(95%CI)分别为0.826(0.803~0.849)和0.814(0.792~0.835),预测3 a和5 a CSS的AUC(95%CI)分别为0.845(0.820~0.871)和0.834(0.811~0.857);外部验证中预测3 a和5 a OS的AUC(95%CI)分别为0.855(0.779~0.932)和0.838(0.768~0.909),预测3 a和5 a CSS的AUC(95%CI)分别为0.806(0.701~0.911)和0.806(0.726~0.886)。结论:本研究所构建的列线图可以准确预测ACCHN患者的OS和CSS,有助于个性化的预后评估和临床决策。
基金supported by the Natural Science Foundation Project of Chongqing(Grant No.cstc2018jcyjAX0077)the Open Fund of Key Laboratory of Advanced Manufacturing Technology for Automobile Parts,Ministry of Education(Grant No.2018KLMT06)the Graduate Research and Innovation Foundation of Chongqing,China(Grant No.CYB18059)
文摘Four-wheel independent steering(4 WIS) system and direct yaw moment control(DYC) have an important influence on vehicle lateral stability. However, DYC has a great effect on the longitudinal velocity, and the capability of 4 WIS is limited to stability.To decrease the influence on the longitudinal velocity and improve the stability of electrical vehicles, a chassis controller integrated with a 4 WIS system and a DYC system with model predictive control(MPC) is designed. The framework consists of an unscented Kalman filter(UKF) observer and an MPC that contains three blocks: supervisor blocks, upper blocks and lower blocks. First, the sideslip angle, longitudinal velocity and lateral tire forces are estimated by the UKF observer;second, a bicycle model is utilized in the supervisor to calculate the desired values;third, the upper blocks are designed with the MPC to optimize the target steering angles and longitudinal tire forces under the constraints of subsystems;to facilitate the design of the MPC, a nonlinear tire is simplified based on the Taylor expansion method;finally, the target steering angles and longitudinal tire forces are achieved by the lower blocks. The integrated controller is simulated on the co-simulation platform of MATLAB-Carsim. The results show that the proposed integrated controller has less impact on longitudinal velocity and could effectively improve vehicle stability.
