基于多参数3.0T MRI影像组学的列线图在卵巢附件肿块定性中的价值研究

A study on value of nomograms in characterization of ovarian adnexal masses based on multiparametric 3.0T MRI radiomics

摘要

目的 旨在评估临床特征、多参数3.0T磁共振成像(MRI)影像组学对卵巢附件肿块定性的预测价值,并构建预测卵巢附件肿块恶性病变的列线图模型。方法 本研究选取2021年1月至2023年1月期间在承德市中心医院经手术组织学病理确诊的147例卵巢附件肿块患者作为研究对象,根据患者组织病理结果分为良性组(n=83)和恶性组(n=64)。比较两组患者临床资料;通过Radcloud平台提取3.0T MRI影像组学特征,再采用Lasso法筛选出影响显著的影像组学特征并绘制影像组学列线图;使用患者临床危险因素特征和影像组学评分联合构建影像组学-临床组合列线图模型;并采用受试者工作特征(ROC)曲线下面积(AUC)比较两种模型对卵巢附件肿块恶性病变的预测价值。结果 两组患者年龄、绝经状态、肿瘤大小、卵巢受累、强化特点及腹腔积液等临床特征差异无统计学意义(P>0.05),而恶性组内膜增生、肿瘤病灶成分为囊实或实性、囊壁有壁结节占比比良性组高,差异具有统计学意义(χ²值分别为22.377、11.332、22.940,P<0.05);列线图和ROC结果显示,影像组学预测恶性病变的AUC为0.867(95%CI:0.621~0.905),灵敏度为83.25%,特异度为89.63%;影像组学-临床组合列线图预测模型的AUC为0.913(95%CI:0.691~0.959),灵敏度为85.13%,特异度为90.01%。结论 卵巢附件肿块患者3.0T MRI影像组学评分能够反映卵巢附件肿块恶性病变情况,且影像组学-临床组合列线图模型在预测卵巢附件肿块患者恶性病变方面具有较高的诊断价值。

Abstract

Objective To evaluate predictive value of clinical characteristics and multiparametric 3.0T magnetic resonance imaging (MRI) radiomics for characterizing ovarian adnexal masses,and to construct a nomogram model for predicting malignant lesions in ovarian adnexal lesions. Methods 147 patients with ovarian adnexal masses who underwent surgical histopathological diagnosis in Chengde Central Hospital from January 2021 to January 2023 were included in this study.According to histopathological result,the patients were categorized into benign mass group (n=83) and malignant mass group (n=64).The clinical data of the patients were compared between the two groups.3.0T MRI radiomic features of the patients were extracted via Radcloud platform,and Lasso regression analysis was used to screen those significant radiomics features and a radiomics nomogram was constructed.The patient􀆳s clinical risk factor characteristics were combined with radiomics score (Rad-score) to construct a radiomics-clinical combination nomogram model.The area under the curve (AUC) of receiver operating characteristic (ROC) curve was used to compare predictive values of the two nomogram models for malignant lesions in ovarian adnexal masses. Results In terms of clinical features as age,menopausal status,tumor size,ovarian involvement,enhancement characteristics and ascites,there were no statistically significant differences between the two groups (all P >0.05).However,the patients in the malignant mass group presented higher proportions of endometrial hyperplasia,cystic-solid or solid tumor components,and cyst wall nodules compared to the benign mass group,and the differences were significant (χ ²=22.377,11.332 and 22.940 respectively,all P <0.05).The analyses of the ROC curve and the nomogram shown that the AUC of radiomics for predicting malignant lesions was 0.867 (95%CI: 0.621-0.905),with a sensitivity of 83.25% and specificity of 89.63%,and the AUC of the radiomics-clinical combination prediction model was 0.913 (95%CI: 0.691-0.959),with a sensitivity of 85.13% and specificity of 90.01%. Conclusion The Rad-score derived from 3.0T MRI images of the patients with ovarian adnexal masses can reflect malignant character of these masses. Furthermore,the radiomics-clinical combination nomogram model exerts high diagnostic value in predicting malignant lesions in patients with ovarian adnexal masses.

关键词

3.0T 磁共振成像 / 影像组学 / 列线图模型 / 卵巢附件肿块 / 定性

Key words

3.0T MRI / radiomics / nomogram model / ovarian adnexal mass / characterization

参考文献

[1] Saheb S K, Sreenivasarao D. An optimized siamese neural network with deep linear graph attention model for gynecological abdominal pelvic masses classification[J]. Abdom Radiol (NY), 2025,50(4): 1831-1859.

[2] Shetty M. Imaging of a Pelvic Mass:Uterine[J]. Semin Ultrasound CT MR, 2023,44(6): 528-540.

[3] 王晓蕾, 郭金花, 管秀红, 等. 卵巢囊性占位性病变在多层螺旋CT中的图像表现及鉴别诊断[J]. 中国CT和MRI杂志, 2019,17(10): 96-98.

[4] 张晓宇, 庞孟春, 周强. MRI影像组学预测三阴性乳腺癌患者发生放疗抵抗的应用研究[J]. 中国妇幼健康研究, 2025,36(6): 93-102.

[5] Tian L, Peng N, Qian Z, et al. Clinical evaluation of resection of functional areagliomas guided by intraoperative 3.0 T MRI combined with functional MRI navigation[J]. BMC Surg, 2024,24(1): 216.

[6] Wadnere N, Ahuja A, Behl S. Access the diagnostic reliability and imaging evaluation of adnexal masses diagnosis ultrasonography,magnetic resonance imaging and computed tomography[J]. International Journal of Research in Medical Sciences, 2021,9(3): 729-734.

[7] Fokkinga E, Hernandez-Tamames J A, Ianus A, et al. Advanced diffusion-weighted MRI for cancer microstructure assessment in body imaging,and its relationship with histology[J]. J Magn Reson Imaging, 2024,60(4): 1278-1304.

[8] Mayerhoefer M E, Materka A, Langs G, et al. Introduction to radiomics[J]. J Nucl Med, 2020,61(4): 488-495.

[9] Jan Y T, Tsai P S, Huang W H, et al. Machine learning combined with radiomics and deep learning features extracted from CT images:a novel AI model to distinguish benign from malignant ovarian tumors[J]. Insights Imaging, 2023,14(1): 68.

[10] Shur J D, Doran S J, Kumar S, et al. Radiomics in oncology:a practical guide[J]. Radiographics, 2021,41(6): 1717-1732.

[11] 王丰, 秦思源, 周延, 等. T2WI影像组学鉴别卵巢成人型颗粒细胞瘤与DWI高信号纤维-卵泡膜细胞肿瘤[J]. 磁共振成像, 2024,15(8): 152-157, 165.

[12] 汪朗锟, 叶蕾, 张朋. 基于MRI影像组学对膀胱癌肌层浸润预测模型的构建[J]. 临床泌尿外科杂志, 2024,39(9): 781.

[13] 高文, 谭理连, 张立. 卵巢及附件巨大肿瘤MRI定性诊断分析[J]. 实用放射学杂志, 2019,35(11): 1797-1800, 1830.

[14] Stewart C, Ralyea C, Lockwood S. Ovarian cancer:an integrated review[J]. Semin Oncol Nurs, 2019,35(2): 151-156.

[15] Torre L A, Trabert B, Desantis C E, et al. Ovarian cancer statistics,2018[J]. CA Cancer J Clin, 2018,68(4): 284-296.

[16] Zhang M, Zhang Y, Fu J, et al. Serum CA125 levels are decreased in rectal cancer but increased in fibrosis-associated diseases and in most types of cancers[J]. Prog Mol Biol Transl Sci, 2019,162: 241-252.

[17] Bazot M, Spagnoli F, Guerriero S. Magnetic resonance imaging and ultrasound fusion technique in gynecology[J]. Ultrasound Obstet Gynecol, 2022,59(2): 141-145.

[18] Avesani G, Panico C, Nougaret S, et al. ESR Essentials:characterization and staging of adnexal masses with MRI and CT-practice recommendations by ESUR[J]. Eur Radiol, 2024,34(12): 7673-7689.

[19] Baadsvik EL, Weiger M, Froidevaux R, et al. High-resolution MRI of mummified tissues using advanced short-T2 methodology and hardware[J]. Magn Reson Med, 2021,85(3): 1481-1492.

[20] Yu X, Zheng Q, Ai M, et al. MR Imaging of Typical Ovarian Hemangioma:A Case Report[J]. Curr Med Imaging, 2024,20: e15734056293540.

[21] Liu L, Zhang W, Wang Y, et al. Radiomics combined with clinical and MRI features may provide preoperative evaluation of suboptimal debulking surgery for serous ovarian carcinoma[J]. Abdom Radiol (NY), 2025,50(1): 496-512.

[22] Wu S, Zheng J, Li Y, et al. A radiomics nomogram for the preoperative prediction of lymph node metastasis in bladder cancer[J]. Clin Cancer Res, 2017,23(22): 6904-6911.

[23] 李惠玲, 麦神忠, 黎洪芳. 基于MRI的影像组学对中央腺体前列腺癌和前列腺增生结节鉴别诊断价值分析[J]. 影像研究与医学应用, 2023,7(2): 170-172.

[24] 刘雪芬, 张国福, 金俊, 等. 多序列MRI的影像组学在卵巢上皮性肿瘤鉴别及疗效预测分析[J]. 中国全科医学, 2023,26(S1): 89-91.

[25] 夏旭东, 李铭, 王海彬, 等. 基于MRI影像组学鉴别卵巢卵泡膜细胞瘤与阔韧带肌瘤[J]. 中国医学影像学杂志, 2022,30(2): 145-152.

[26] Zeng Y, Wang X, Wu J, et al. Survival analysis of patients with extrahepatic cholangiocarcinoma:a nomogram for clinical and MRI features[J]. BMC Med Imaging, 2024,24(1): 7.

[27] Chen J, Gu H, Zhang Y, et al. MRI-based nomogram for differentiation of ovarian fibrothecoma and broad ligament myoma[J]. Sci Rep, 2022,12(1): 8122.

[28] Yu X Y, Ren J, Jia Y, et al. Multiparameter MRI radiomics model predicts preoperative peritoneal carcinomatosis in ovarian cancer[J]. Front Oncol, 2021,11: 765652.

[29] 杨钱, 王海宝. 基于术前MRI影像组学及临床特征的卵巢癌淋巴结转移预测模型构建及验证[J]. 医学研究与战创伤救治, 2024,37(1): 63-68.

基金资助

河北省医学科学研究重点课题计划项目(20211799)

承德市科技计划项目(202402A007)