摘要
目的探讨超声影像组学评分(RS)联合常规超声图像特征鉴别诊断甲状旁腺腺瘤(PA)与甲状旁腺增生(PH)的临床应用价值。方法选取在我院行手术治疗的甲状旁腺功能亢进症患者133例(181个病灶),其中PA 66例(67个病灶),PH 67例(114个病灶),所有病灶按7∶3的比例随机分为训练集(126个病灶)和验证集(55个病灶),术前均行常规超声检查,基于灰阶超声图像勾画感兴趣区并提取影像组学特征。采用最小绝对收缩和选择算子(LASSO)回归筛选特征并构建影像组学模型,计算每个病灶的RS;采用多因素Logistic回归分析筛选差异有统计学意义的常规超声图像特征并构建常规超声模型,进一步构建基于RS联合常规超声图像特征的列线图模型。绘制受试者工作特征(ROC)曲线分析各模型对PA与PH的鉴别诊断效能;绘制校准曲线评估影像组学模型和列线图模型的校准度。结果基于LASSO回归筛选出8个影像组学特征用于构建影像组学模型,训练集中PA和PH的RS分别为(0.64±0.86)分、(-0.17±0.63)分;验证集中PA和PH的RS分别为(0.59±0.74)分、(-0.08±0.65)分,差异均有统计学意义(均P<0.05)。多因素Logistic回归分析显示,病灶最大径、外周线状高回声和极性供支血管征均为鉴别PA与PH的独立影响因素(均P<0.05),基于上述3个独立影响因素构建常规超声模型。在训练集和验证集中,影像组学模型鉴别诊断PA与PH的曲线下面积(AUC)分别为0.764和0.750,常规超声模型的AUC分别为0.812和0.838,列线图模型的AUC分别为0.825和0.856;列线图模型的AUC均高于影像组学模型,差异均有统计学意义(均P<0.05)。校准曲线显示,影像组学模型和列线图模型的校准度均较高,与实际结果的一致性均较好。结论RS联合常规超声图像特征能准确鉴别诊断PA与PH,具有较好的临床应用价值。
Objective To explore the clinical value of ultrasound radiomics score(RS)combined with features of conventional ultrasound imaging in the differential diagnosis of parathyroid adenoma(PA)and parathyroid hyperplasia(PH).Methods A total 133 patients with hyperparathyroidism(181 lesions in total)who underwent surgery in our hospital were selected,including 66 cases of adenoma(67 lesions)and 67 cases of hyperplasia(114 lesions).A ratio of 7∶3 was used to divide the lesions into training set(126 lesions)and validation set(55 lesions).All lesions were underwent conventional ultrasound examination.The region of interest(ROI)was delineated based on ultrasound images and the ultrasound radiomics features were extracted.The least absolute shrinkage and selection operator(LASSO)regression was used to screen features and construct the radiomics model,the RS of each lesion was calculated.Multivariate Logistic regression was used to screen conventional ultrasound features with statistically significant differences and construct the conventional ultrasound model,and the nomogram model based on RS and features of conventional ultrasound imaging was further constructed.Receiver operating characteristic(ROC)curve was drawn to analyze the differential diagnostic efficacy of each model.The calibration curve was drawn to evaluate the calibration between the radiomics model and the nomogram model.Results A total of 8 radiomics features were selected for constructing the radiomics model based on LASSO regrtssion.The RS of PA and PH in training set were(0.64±0.86)points and(-0.17±0.63)points,while RS of PA and PH in validation set were(-0.59±0.74)points and(-0.08±0.65)points,respectively.The differences were statistically significant(both P<0.05).Multivariate Logistic regression analysis showed that lesion maximum diameter,peripheral high-eco bright line and polar vascular sign were independent influencing factors for differential diagnosis of PA and PH(all P<0.05).The conventional ultrasound model was constructed based on above 3 independent influencing factors.In the training set and validation set,the area under the curve(AUC)of the radiomics model in differentiating PA and PH were 0.764 and 0.750,respectively,the AUC of the conventional ultrasound model were 0.812 and 0.838,respectively,and the AUC of the nomogram model were 0.825 and 0.856,respectively.The AUC of the nomogram model were higher than those of the radiomics model in both the training set and validation set,and the differences were statistically significant(both P<0.05).Calibration curve showed that both the radiomics model and the nomogram model had high calibration accuracy,and good consistency with the actual results.Conclusion RS combined with features of conventional ultrasound imaging can accurately differentiate PA and PH,which has a good clinical application value.
作者
马可欣
杨雨薇
吕发金
刘丽萍
MA Kexin;YANG Yuwei;LV Fajin;LIU Liping(Department of Ultrasound,the First Affiliated Hospital of Chongqing Medical University,Chongqing 400016,China)
出处
《临床超声医学杂志》
CSCD
2024年第3期210-216,共7页
Journal of Clinical Ultrasound in Medicine
关键词
超声检查
影像组学
甲状旁腺腺瘤
甲状旁腺增生
鉴别诊断
列线图
Ultrasonography
Radiomics
Parathyroid adenoma
Parathyroid hyperplasia
Differential diagnosis
Nomogram
