基于术前CT影像学数据构建肝细胞癌（肝癌）患者生存期的Nomogram预测模型。

回顾性分析2018年1月至2020年12月在青岛大学附属医院接受手术治疗的243例肝癌患者临床影像学资料。其中男206例，女37例；年龄49~70岁，中位年龄59岁。收集患者术前增强CT多模态数据，利用Pyradiomics技术提取多维度定量放射组学特征，包括肿瘤形态学特征、信号强度分布、纹理特征及信号强度-体积直方图等。将患者划分为训练集（158例）和验证集（85例）。采用Cox单因素和多因素回归分析筛选出患者生存预后的独立影响因素。基于这些因素构建Nomogram预测模型。

基于标准化后的特征数据，最终确定了30个关键的影像学特征。Cox回归分析确定3个独立的关键影像组学特征，包括glcm_MCC、glszm_Zone Percentage、shape_Sphericity。该Cox回归模型训练集的ROC AUC为0.82。基于Cox多因素回归的风险评分将患者分为高风险组和低风险组，Kaplan-Meier生存分析显示训练集低风险组和高风险组生存差异有统计学意义（χ²=7.353，P<0.05）。构建肝癌患者生存预后的Nomogram模型，验证集该模型的AUC为0.78。验证集Nomogram模型高、低风险组生存比较，差异亦有统计学意义（χ²=2.38，P<0.05），进一步验证了Nomogram模型的有效性与可靠性。

本研究基于术前CT影像数据开发的Nomogram模型能有效预测肝癌患者生存期，可以辅助临床医师进行早期预判，提供辅助决策支持。

To construct a Nomogram model for predicting the survival of patients with hepatocellular carcinoma (HCC) based on preoperative CT imaging data.

Clinical imaging data of 243 patients with HCC who underwent surgical treatment in the Affiliated Hospital of Qingdao University from January 2018 to December 2020 were retrospectively analyzed. Among them, 206 patients were male and 37 female, aged from 49 to 70 years, with a median age of 59 years. Multi-modal data of preoperative contrast-enhanced CT were collected. Multi-dimensional quantitative radiomic features were extracted by Pyradiomics, including tumor morphological features, signal intensity distribution, texture features and signal intensity-volume histogram, etc. All patient datasets were divided into the training (n=158) and validation sets (n=85). Univariate and multivariate Cox regression analyses were used to screen the independent factors affecting the survival and prognosis of HCC patients. Nomogram model was constructed based on these factors.

Based on the standardized feature data, 30 key imaging features were eventually determined. Cox regression analysis identified three independent key radiomic features including glcm_MCC, glszm_Zone Percentage and shape_Sphericity. The area under the ROC curve (AUC) of the training set in Cox regression model was 0.82. All patients were divided into the high-risk and low-risk groups based on the risk score of multivariate Cox regression analysis. Kaplan-Meier survival analysis showed that the difference in the survival between the low-risk and high-risk groups in the training set was statistically significant (χ²=7.353, P<0.05). A Nomogram model of survival and prognosis of patients with liver cancer was constructed. The AUC of the Nomogram model in the validation set was 0.78. The difference in the survival between the low-risk and high-risk groups in the validation set was statistically significant (χ²=2.38, P<0.05), which further validated the effectiveness and reliability of the Nomogram model.

Nomogram model constructed based on preoperative CT imaging data can effectively predict the survival of HCC patients, which can effectively assist early prediction and provide decision-making support for clinicians.