信息工程与人工智能学院知识库

已经有研究^[1]证明患者接受治疗后肿瘤较高的最大标准化摄取值（Maximum Standardized Uptake Value，SUVmax）与局部晚期非小细胞肺癌患者较差的生存率相关，并且当肿瘤SUVmax取5.0的二值截取值时，其与患者的生存预后具有显著的相关性（p<0.05）。而在临床实践中，患者在接受治疗后并不会总是再次进行PET/CT检查。因此，基于患者接受治疗前的FDG-PET图像预测出患者接受治疗后的肿瘤SUVmax是否大于5.0可能会对预测局部晚期非小细胞肺癌患者的总生存期有潜在的价值。并且由于患者治疗后的数据搜集比较困难，目前对非小细胞肺癌患者的生存分析大部分都是基于患者接受治疗前的一些信息，很少有同时基于患者治疗前和治疗后的信息进行的生存分析。

因此，本文基于患者接受治疗前的FDG-PET对患者接受治疗后的肿瘤SUVmax是否大于5.0进行了预测，并评估了其对局部晚期非小细胞肺癌患者总生存期的影响；并进一步研究了患者接受治疗前和接受治疗后的临床信息对患者生存分析的影响。

基于治疗后SUVmax的预测对局部晚期非小细胞肺癌患者的生存分析研究。研究内容包括两个部分，分别是患者接受治疗后的肿瘤SUVmax是否大于5.0的预测以及生存分析。首先是患者接受治疗后的肿瘤SUVmax是否大于5.0的预测。其次，采用Cox单因素和多因素分析对患者的临床特征进行挑选，将预测出的结果以及临床特征同时输入到Cox比例风险模型（Cox Proportional Hazards Model^[2], Cox 模型）中对患者进行生存分析的预测。最终的实验结果表明利用患者接受治疗前的FDG-PET图像可以准确预测出患者接受治疗后的肿瘤SUVmax是否大于5.0，并且预测出的结果可以提升患者生存分析的准确性。

基于临床信息对局部晚期非小细胞肺癌患者的生存分析研究。研究内容同样包括两个部分，临床特征的选择以及患者生存分析的预测。利用Cox单因素和多因素分析进行临床特征的选择，然后用Cox 模型对患者生存分析进行预测。最终的实验结果显示，利用患者治疗前和治疗后的临床信息可以准确预测出患者的生存率。

本文最终的实验结果显示，基于患者接受治疗前的FDG-PET预测出的患者接受治疗后的肿瘤SUVmax是否大于5.0的结果可以显著提升患者生存分析的准确性，是局部晚期非小细胞肺癌患者预后的一个重要因素，对医生的临床决策具有一定的指导意义。基于临床特征对患者进行生存分析的研究可以更加深入地了解患者的病情和治疗效果，为医生制定更加个性化的治疗方案提供参考。这项研究对于提高局部晚期非小细胞肺癌患者的生存率和治疗效果具有重要的临床意义。

It has been demonstrated^[1] that the higher maximum standardized uptake value (SUVmax) of tumors after patient treatment is associated with poorer survival in patients with locally advanced non-small cell lung cancer and that when the tumor SUVmax takes a binary cutoff value of 5.0, it is significantly associated with survival prognosis of patients correlation (p<0.05). In clinical practice, treated patients do not always undergo repeat PET/CT. Therefore, prediction of post-treatment tumor SUVmax greater than 5.0 based on pre-treatment FDG-PET may be potentially valuable in predicting overall survival in patients with locally advanced non-small cell lung cancer. Since it is difficult to collect post-treatment data, most of the current survival analyses of patients with non-small cell lung cancer are based on some pre-treatment information, and few studies have examined analyses based on both pre- and post-treatment information, this paper further investigates the impact of pre- and post-treatment clinical information on the survival analysis of patients.

Therefore, this article predicted whether the tumor post-treatment SUVmax was greater than 5.0 based on the patient's pre-treatment information, and evaluated its impact on the overall survival of locally advanced non-small cell lung cancer patients. Furthermore, the study investigated the impact of pre- and post-treatment clinical information on patient survival analysis.

This study focuses on the survival analysis of locally advanced non-small cell lung cancer patients based on the prediction of post-treatment SUVmax. The study consists of two parts: predicting whether the tumor SUVmax after treatment is greater than 5.0 and survival analysis. First, a 3D convolutional neural network (3D CNN) was used to extract features from the pre-treatment whole-body FDG-PET to predict whether the tumor SUVmax after treatment is greater than 5.0. Second, Cox proportional hazards model^[2] (Cox Model) was used to conduct survival analysis by selecting clinical features with Cox univariate and multivariate analysis and combining them with the predicted results. The results indicate that using pre-treatment FDG-PET images can predict the tumor SUVmax after treatment relatively accurately and improve the accuracy of survival analysis.

This study focuses on the survival analysis of locally advanced non-small cell lung cancer patients based on clinical information. The study also consists of two parts: selecting clinical features and predicting patient survival. Cox univariate and multivariate analysis were used to select clinical features, and Cox Model was used to predict patient survival. The results indicate that using both pre- and post-treatment clinical information can accurately predict patient survival time, and the selected clinical features reflect the impact of clinical information on patient survival prognosis.

This article predicted the impact of tumor SUVmax after treatment on patient survival analysis and found that the prediction results based on pre-treatment FDG-PET can significantly improve the accuracy of patient survival analysis, which is an important factor in the prognosis of locally advanced non-small cell lung cancer patients and has certain guiding significance for clinical decision-making. The study of patient survival analysis based on pre- and post-treatment clinical features can provide reference for doctors to develop more personalized treatment plans, and deepen the understanding of patients' condition and treatment effectiveness. This research has important clinical significance for improving the survival rate and treatment effectiveness of locally advanced non-small cell lung cancer patients.