基于PET-CT和超声造影构建甲状腺癌术后复发转移的预测模型

Prediction model of postoperative recurrence and metastasis of thyroid cancer constructed based on PET-CT and contrast-enhanced ultrasound

目的基于正电子发射计算机断层显像(PET-CT)和超声造影参数构建甲状腺癌(TC)患者术后复发转移的预测模型,为甲状腺癌的治疗提供依据。方法回顾性选取保定市第一中心医院2018年1月—2021年3月收治的行手术治疗的116例TC患者,患者术前均进行PET-CT和超声造影检查,并收集患者临床相关资料。术后对患者进行为期3年的随访,根据随访期间是否发生复发转移将其分为发生组(32例)和未发生组(84例)。患者术后复发转移的影响因素分析采用多因素逐步Logistic回归模型;并基于PET-CT和超声造影参数构建列线图预测模型,分析患者术后复发转移的预测效能用受试者工作特征(ROC)曲线下面积(AUC)。结果发生组患者最大标准摄取值(SUVmax)高于未发生组,达峰时间(TTP)、峰值强度(PI)低于未发生组(P<0.05)。发生组患者临床分期为Ⅲ/Ⅳ期、有淋巴结转移占比均高于未发生组,术后131碘治疗占比低于未发生组(P<0.05)。多因素逐步Logistic回归分析显示,SUVmax水平(OR=4.631,95%CI:2.077~10.323)、临床分期(OR=5.427,95%CI:2.653~11.098)是患者发生术后复发转移的独立危险因素,PI水平(OR=0.254,95%CI:0.136~0.471)、术后131碘治疗(OR=0.306,95%CI:0.110~0.849)是独立保护因素(P<0.05)。基于上述影响因素构建的列线图预测模型经Bootstrap法内部验证,C-index指数为0.836(95%CI:0.734~0.938),预测患者术后复发转移的校正曲线与理想曲线趋近(P>0.05)。ROC曲线显示,列线图模型预测患者术后复发转移的敏感度为87.50%、特异性为88.10%,AUC为0.881(95%CI:0.794~0.968)(P<0.05)。结论基于PET-CT参数SUVmax和超声造影参数PI构建的列线图预测模型可较好的预测TC患者术后复发转移发生风险。

Objective To construct a predictive model for postoperative recurrence and metastasis of thyroid cancer(TC)based on positron emission computed tomography(PET-CT)and contrast-enhanced ultrasound parameters.Methods A total of 116 patients with TC who underwent surgical treatment in our hospital from January 2018 to March 2021 were retrospectively selected.All patients underwent PET-CT and contrast-enhanced ultrasound before operation,and the clinical data of the patients were collected.The patients were followed up for 3 years after operation,and they were divided into the occurrence group(32 cases)and the non-occurrence group(84 cases)according to whether recurrence and metastasis occurred during the follow-up period.The influencing factors of postoperative recurrence and metastasis were analyzed by multivariate stepwise Logistic regression model.A nomogram prediction model was constructed based on PET-CT and CEUS parameters.The area under the receiver operating characteristic(ROC)curve(AUC)was used to analyze the predictive efficacy of the model for postoperative recurrence and metastasis.Results The maximum standardized uptake value(SUVmax)in the occurrence group were higher than those in the non-occurrence group,and the time to peak(TTP),peak intensity(PI)was lower than that in the non-occurrence group(P<0.05).The proportion of patients with clinical stageⅢ/Ⅳand lymph node metastasis in the occurrence group was higher than that in the non-occurrence group,and the proportion of postoperative 131 iodine treatment was lower than that in the non-occurrence group(P<0.05).Multivariate stepwise Logistic regression analysis showed that SUVmax level(OR=4.631,95%CI:2.077~10.323),clinical stage(OR=5.427,95%CI:2.653~11.098)were independent risk factors for postoperative recurrence and metastasis,and PI level(OR=0.254,95%CI:0.136~0.471),postoperative 131I treatment(OR=0.306,95%CI:0.110~0.849)were independent protective factors(P<0.05).The nomogram prediction model based on the above influencing factors was internally verified by the Bootstrap method,and the C-index was 0.836(95%CI:0.734~0.938).The calibration curve for predicting postoperative recurrence and metastasis was close to the ideal curve(P>0.05).The ROC curve showed that the sensitivity of the nomogram model for predicting postoperative recurrence and metastasis was 87.50%,the specificity was 88.10%,and the AUC was 0.881(95%CI:0.794~0.968)(P<0.05).Conclusion The nomogram prediction model based on PET-CT parameter SUVmax and CEUS parameter PI can better predict the risk of postoperative recurrence and metastasis in TC patients.