| 刘畅,胡春梅,马丹丹,等.基于卵巢-附件超声影像和临床数据构建卵巢肿瘤良恶性诊断预测模型的研究[J].中国临床保健杂志,2025,28(6):792-796. |
| 基于卵巢-附件超声影像和临床数据构建卵巢肿瘤良恶性诊断预测模型的研究 |
| Research on the construction of a benign and malignant diagnostic prediction model for ovarian tumors based on ovarian adnexal ultrasound imaging and clinical data |
| 投稿时间：2025-07-02 |
| DOI：10.3969/J.issn.1672-6790.2025.06.012 |
| 中文关键词: 卵巢肿瘤 诊断技术,妇产科 癌症早期检测 危险因素 |
| 英文关键词: Ovarian neoplasms Diagnostic techniques,obstetrical and gynecological Early detection of cancer Risk factors 〖FL |
| 基金项目:安徽省卫生健康科研项目(AHWJ2024Aa20489) |
|
| 摘要点击次数: 33 |
| 全文下载次数: 22 |
| 中文摘要: |
| 目的 探讨卵巢-附件超声影像报告和数据系统(O-RADS)分类联合糖类抗原125(CA125)水平、人附睾蛋白4(HE4)水平、绝经状态及体重指数(BMI)构建预测模型在O-RADS 3类及4类卵巢肿瘤患者中的诊断价值及预测效能。方法 回顾性分析2022年7月至2025年4月在中国科学技术大学附属第一医院(安徽省立医院)就诊的151例卵巢肿瘤患者病历资料,术前均接受超声检查并对获取的超声图像进行O-RADS分类,且临床资料完整、接受手术治疗并获取病理结果。通过单因素、多因素分析卵巢恶性肿瘤的独立危险因素。基于多因素分析结果构建预测模型,绘制列线图。采用校准曲线、临床决策曲线和受试者工作特征(ROC)曲线对模型进行验证。结果 超声O-RADS分类、CA125水平、HE4水平、绝经状态及BMI在卵巢恶性肿瘤组阳性率明显高于良性肿瘤组,差异均有统计学意义(P<0.05)；多因素logistic回归分析结果显示,O-RADS分类4类(OR=8.621,95%CI:2.684~27.696)、HE4水平升高(OR=4.222,95%CI:1.214~14.679)、CA125水平升高(OR=3.762,95%CI:1.248~11.336)、绝经后(OR=3.639,95%CI:1.263~10.481)及BMI升高(OR=2.975,95%CI:1.011~8.753)均是卵巢恶性肿瘤的独立危险因素。由此建立的卵巢恶性肿瘤预测模型校准曲线趋近于标准曲线,预测值与实测值高度吻合,决策曲线显示概率阈值为0.07~0.94时,预测卵巢恶性肿瘤模型的ROC曲线下面积、特异度、灵敏度及准确率分别为0.917(95%CI:0.874~0.961)、82.9%、91.2%、84.9%。结论 超声O-RADS分类联合CA125水平、HE4水平、绝经状态及BMI构建的预测模型判断O-RADS 3类及4类肿瘤患者良恶性的准确率较高。 |
| 英文摘要: |
| Objective To explore the diagnostic value and predictive efficacy of the prediction model constructed by combining O-RADS with clinical indicators [including CA125 level,HE4 level,menopausal status and body mass index (BMI)] in ovarian tumors of patients classified as O-RADS 3 and 4.Methods A retrospective analysis was conducted among 151 patients who visited the First Affiliated Hospital of University of Science and Technology of China from July 2022 to April 2025.All patients underwent ultrasound examinations to obtain ultrasound images for O-RADS classification.The clinical data were complete,and they all received surgical treatment and obtained pathological results.Univariate and multivariate analyses were conducted to identify the independent risk factors for ovarian malignant tumors.Based on the results of multivariate analysis,a prediction model was constructed and nomogram was drawn.The model was validated using calibration curves,decision curves,and receiver operating characteristic (ROC) curves.Results The positive rates of ultrasound O-RADS,CA125 level,HE4 level,menopausal status and BMI were significantly higher in the ovarian malignant tumor group,and the differences were statistically significant (P<0.05).The results of the multivariate logistic regression analysis showed that,O-RADS 4 (OR=8.621,95%CI:2.684-27.696),elevated HE4 level (OR=4.222,95%CI:1.214-14.679),elevated CA125 level (OR=3.762,95%CI:1.248-11.336),post-menopause (OR=3.639,95%CI:1.263-10.481),and elevated BMI index (OR=2.975,95%CI:1.011-8.753) are all independent risk factors for ovarian malignant tumors.The calibration curve approaches the standard curve,and the predicted values are highly consistent with the measured values;The decision curve shows that when the probability threshold ranges from 0.07 to 0.94,the clinical net benefit for predicting the risk of ovarian malignancy is relatively high,the area under the ROC curve (AUC),specificity,sensitivity and accuracy of the model were 0.917 (95%CI:0.874-0.961),82.9%,91.2% and 84.9%,respectively.Conclusions The prediction model constructed by combining the O-RADS with clinical indicators (including CA125 level,HE4 level,menopausal status and BMI) can better diagnose the benign and malignant nature of patients with O-RADS 3 and 4. |
|
查看全文
|
| 关闭 |
|
|
|