目的 基于CT影像组学与中医舌象特征构建肝细胞癌（HCC）根治术后早期复发的Nomogram风险预测模型。 方法 选取2018年9月至2022年9月在医院进行HCC根治术的216例患者，对其临床资料进行回顾性分析。按7：3比例随机将患者分为建模集（n = 152）和验证集（n = 64），记录两组术后1年复发情况。以建模集患者数据构建复发预测模型，患者术后均行平扫联合增强CT检查，提取影像组学特征指标，记录舌象特点，采用Cox多因素模型分析建模集患者术后复发的相关因素，构建风险预测模型。以验证集数据对模型进行验证。 结果 建模集患者1年复发52例，复发率34.21%；验证集患者1年复发23例，复发率35.94%。Cox多因素模型分析显示，Rad-score、舌质及舌形是HCC术后复发的独立影响因素（P < 0.05），基于此建立Nomogram列线图，ROC分析显示，列线图模型判断建模集与验证集患者术后复发的AUC分别为0.811和0.824，敏感度分别为0.875和0.833，特异度分别为0.617和0.750。 结论 HCC根治术后早期复发与Rad-score、舌质及舌形相关，基于此构建的Nomogram预测模型，对判断术后早期复发具有较高准确性。

Objective To construct a Nomogram risk prediction model for early recurrence of hepatocellular carcinoma （HCC） radical hepatectomy based on CT imaging omics and traditional Chinese medicine tongue imaging features. Methods 216 patients who underwent HCC radical hepatectomy in the First Affiliated Hospital of Henan University of Traditional Chinese Medicine from September 2018 to September 2022 were selected，the clinical dataes were retrospective analyzed.The patients were randomly divided into modeling set （n = 152） and validation set （n = 64） by 7：3 ratio，the recurrence situation one year after surgery of the two groups were recorded.A recurrence prediction model was constructed by the modeling set of patient dataes， all patients underwent postoperative plain scan combined with enhanced CT examination， and the Imaging omics feature indicators was extracted， the tongue imaging characteristics was recorded，the Cox multivariate model was used to analyze the factors related to postoperative recurrence in the modeling set of patients， and a risk prediction model was constructed，the model was validated by validation set dataes. Results There were 52 patients in the modeling set experienced recurrence within 1 year， the recurrence rate was 34.21%；there were 23 cases of recurrence in the validation set patients within 1 year， the recurrence rate was 35.94%.The Cox multivariate model analysis showed that the Rad-score， Tongue texture and tongue shape were independent influence factors for postoperative recurrence of HCC （P < 0.05）. Based on this， the Nomogram column chart was established.The ROC analysis showed that the AUC of the column chart model for predicting postoperative recurrence in the modeling set and validation set patients were 0.811 and 0.824， the sensitivities were 0.875 and 0.833 respectively， the specificities were 0.617 and 0.750 respectively. Conclusions The early postoperative recurrence of HCC is related to Rad score， tongue texture， and tongue shape. Based on this， the Nomogram prediction model constructed has high accuracy in predicting early postoperative recurrence.