基于超声造影联合卵巢癌风险预测模型指数的列线图模型对卵巢上皮性癌的诊断价值

Diagnostic value of Nomogram model based on contrast-enhanced ultrasound combined with risk of ovarian malignancy algorithm index in epithelial ovarian cancer

目的 基于超声造影联合卵巢癌风险预测模型(ROMA)指数构建列线图模型,探讨其对卵巢上皮性癌(EOC)的诊断价值。方法 选取我院经术后病理确诊的卵巢肿瘤患者214例,将2019年4月至2020年12月确诊的130例患者设为建模集,其中良性肿瘤组85例,EOC组45例;将2021年1月至2022年4月确诊的84例患者设为验证集,其中良性肿瘤组55例,EOC组29例。比较建模集中良性肿瘤组与EOC组患者ROMA指数及超声造影定量参数,包括始增时间(IT)、达峰时间(TTP)、峰值强度(PI)、增强速率(V1)、曲线下面积(TIC-AUC);采用多因素Logistic回归分析筛选EOC影响因子并建立列线图模型;通过建模集及验证集的校准曲线及受试者工作特征(ROC)曲线验证列线图模型对EOC的诊断效能。结果 建模集中,EOC组ROMA指数阳性占比高(77.8%),PI、V1、TIC-AUC均升高,IT、TTP均减少,与良性肿瘤组比较差异均有统计学意义(均P<0.05)。多因素Logistic回归分析显示,ROMA指数、IT、TTP、PI、TIC-AUC均是EOC的影响因子(OR=44.346、0.047、0.063、11.472、37.950,均P<0.05)。建模集和验证集的ROC曲线下面积分别为0.919(95%可信区间:0.873~0.965,P<0.001)、0.835(95%可信区间:0.737~0.933,P<0.001)。当诊断概率截断值为0.244时,列线图模型诊断EOC的敏感性、特异性、准确率分别为86.67%、80.00%、82.31%。结论 基于超声造影联合ROMA指数的列线图模型区分度和校准度均较好,其对EOC具有较高的诊断价值。

Objective To explore the diagnostic value of Nomogram model based on contrast-enhanced ultrasound combined with risk of ovarian malignancy algorithm(ROMA)index in epithelial ovarian cancer(EOC).Methods A total of 214patients with ovarian tumors diagnosed by surgery in our hospital were selected,and 130 patients admitted from April 2019 to December 2020 were set as the modeling set,including 85 cases in the benign tumor group and 45 cases in the EOC group.84 patients admitted from January 2021 to April 2022 were set as the validation set,including 55 cases in the benign tumor group and 29 cases in the EOC group.The ROMA index and quantitative parameters of contrast-enhanced ultrasound,including initial time(IT),time to peak(TTP),peak intensity(PI),enhancement rate(V1),area under the curve(TIC-AUC)were compared between the benign tumor group and the EOC group in the modeling set.Multivariate Logistic regression analysis was used to screen influencing factors of EOC to establish a Nomogram model,the calibration curve and receiver operating characteristic(ROC)curve of the modeling set and validation set were used to verify the diagnostic value of the Nomogram model for EOC.Results In the modeling set,the positive proportion of ROMA index(77.8%),PI,V1 and TIC-AUC in the EOC group were significantly increased,while the IT and TTP were significantly decreased,and the differences were statistically significant compared with those in the benign tumor group(all P<0.05).Multivariate Logistic regression analysis showed that ROMA index,IT,TTP,PI,TIC-AUC were the influencing factors of EOC(OR=44.346,0.047,0.063,11.472,37.950,all P<0.05).The areas under the ROC curves of the modeling set and validation set were 0.919(95%CI:0.873~0.965,P<0.001)and 0.835(95%CI:0.737~0.933,P<0.001),respectively.When the cut-off value of diagnostic probability was 0.244,the sensitivity,specificity and accuracy of the Nomogram model in the diagnosis of EOC were 86.67%,80.00%and 82.31%,respectively.Conclusion The Nomogram model based on contrast-enhanced ultrasound combined with ROMA index has better discrimination and calibration,it also has higher diagnostic value for EOC.