分享:
分享到微信朋友圈
X
临床研究
T2WI-FS影像组学联合影像学特征在预测HIFU消融子宫肌瘤疗效中的价值
秦石泽 黄小华 王芳 唐玲玲 蒋雨

Cite this article as: Qin SZ, Huang XH, Wang F, et al. Value of T2WI-FS radiomics combined with imaging features in predicting the efficacy of HIFU ablation of hysteromyoma[J]. Chin J Magn Reson Imaging, 2022, 13(5): 59-63, 69.本文引用格式:秦石泽, 黄小华, 王芳, 等. T2WI-FS影像组学联合影像学特征在预测HIFU消融子宫肌瘤疗效中的价值[J]. 磁共振成像, 2022, 13(5): 59-63, 69. DOI:10.12015/issn.1674-8034.2022.05.011.


[摘要] 目的 探讨T2加权脂肪抑制成像(T2-weighted imaging fat suppression,T2WI-FS)影像组学联合影像学特征在预测高强度聚焦超声(high intensity focused ultrasound,HIFU)消融子宫肌瘤疗效中的价值。材料与方法 回顾性分析142例临床确诊并接受HIFU消融治疗的子宫肌瘤患者资料,包括172个肌瘤。其中低消融率组(消融率<70%) 77个、高消融率组(消融率≥70%) 95个,以7∶3的比例随机将其分为训练集和测试集。选用矢状位T2WI-FS序列利用3D Slicer软件勾画三维的子宫肌瘤实质周围并提取影像组学特征(包括形状特征、灰度共生矩阵、灰度游程长度矩阵、灰度大小区域矩阵、领域灰度差矩阵、灰度依赖矩阵和一阶特征),同时收集可能与HIFU消融疗效有关的影像学特征[包括子宫肌瘤体积、T2WI-FS信号强度、彩色多普勒血流成像信号、T1加权成像(T1-weighted imaging,T1WI)增强肌瘤强化程度、T1WI增强肌瘤信号均匀度、子宫肌瘤类型]。采用方差阈值法、单变量选择法和最小绝对收缩和选择算子实现影像组学特征筛选。采用Logistic回归和随机森林分别建立两种预测HIFU消融子宫肌瘤疗效的联合模型,通过受试者工作特征(receiver operating characteristic,ROC)曲线评估模型预测性能并使用Delong检验比较两种联合模型的预测效能。结果 Logistic回归模型显示测试集ROC曲线下面积(area under the curve,AUC)为0.855、特异度0.783、敏感度0.724。随机森林模型显示测试集AUC为0.796,特异度0.696、敏感度0.759。训练集和测试集Delong检验结果显示两种联合模型预测效能差异无统计学意义(P>0.05)。结论 基于T2WI-FS影像组学联合影像学特征建立的模型可以较准确地预测HIFU消融子宫肌瘤的疗效。
[Abstract] Objective To investigate the value of T2-weighted imaging fat suppression (T2WI-FS) radiomics combined with imaging features in predicting the efficacy of high intensity focused ultrasound (HIFU) ablation of hysteromyoma.Materials and Methods A total of 142 patients with hysteromyoma diagnosed clinically and treated with HIFU ablation were analyzed retrospectively, including 172 hysteromyomas, including 77 in the low ablation rate group (ablation rate<70%) and 95 in the high ablation rate group (ablation rate≥70%). They were randomly divided into training set and test set in the ratio of 7:3. The sagittal T2WI-FS sequence was selected, and the three-dimensional surrounding hysteromyoma parenchyma was delineated by 3D Slicer software, and the radiomics features were extracted (including shape, gray level co-occurrence matrix, gray level run-length matrix, gray level size zone matrix, neighboring gray tone difference matrix, gray level dependence matrix and first order). At the same time, the imaging features that may be related to the ablation effect of HIFU were collected (including hysteromyoma volume, T2WI-FS signal intensity, color Doppler flow imaging signal, enhancement degree of T1WI enhanced hysteromyoma, signal uniformity of T1WI enhanced hysteromyoma and type of hysteromyoma). The variance threshold method, univariate selection method and least absolute shrinkage and selection operator were used to screen the radiomics features. Logistic regression and random forest were used to establish two mixed models to predict the efficacy of HIFU ablation of hysteromyoma. The prediction performance of the model was evaluated by receiver operating characteristic (ROC) curve, and the prediction efficiency of the two mixed models was compared by Delong test.Results The Logistic regression model shows that the area under the curve (AUC) of the ROC curve of the test set is 0.855, the specificity is 0.783 and the sensitivity is 0.724. The AUC of random forest in the test set is 0.796, the specificity is 0.696 and the sensitivity is 0.759. The Delong test results of training set and test set showed that there was no significant difference in the prediction efficiency between the two mixed models (P>0.05).Conclusions The model based on T2WI-FS radiomics and imaging features can accurately predict the efficacy of HIFU in ablation of hysteromyoma.
[关键词] 影像组学;子宫肌瘤;高强度聚焦超声;磁共振成像;预测;疗效
[Keywords] radiomics;hysteromyoma;high intensity focused ultrasound;magnetic resonance imaging;prediction;efficacy

秦石泽 1   黄小华 1*   王芳 2   唐玲玲 1   蒋雨 1  

1 川北医学院附属医院放射科,南充 637000

2 上海联影智能有限公司研发部,上海 200232

黄小华,E-mail:15082797553@163.com

作者利益冲突声明:全部作者均声明无利益冲突。


基金项目: 南充市市校合作项目 19SXHZ0429 南充市市校科技战略合作项目 20SXQT0303
收稿日期:2022-01-03
接受日期:2022-04-08
中图分类号:R445.2  R737.33 
文献标识码:A
DOI: 10.12015/issn.1674-8034.2022.05.011
本文引用格式:秦石泽, 黄小华, 王芳, 等. T2WI-FS影像组学联合影像学特征在预测HIFU消融子宫肌瘤疗效中的价值[J]. 磁共振成像, 2022, 13(5): 59-63, 69. DOI:10.12015/issn.1674-8034.2022.05.011

       子宫肌瘤是子宫常见的良性肿瘤,发病率约为70%[1],部分患者会出现月经异常、贫血等症状,严重影响生活质量[2]。高强度聚焦超声(high intensity focused ultrasound,HIFU)可诱导子宫肌瘤凝固性坏死,在子宫肌瘤的治疗中得以广泛应用[3, 4, 5],而保证 HIFU长期疗效的关键是术后获得较高的消融率[6, 7, 8]。因此实现子宫肌瘤消融率的精准预测,将有助于评估HIFU疗效,及时剔除不适合HIFU治疗的病例。

       目前影像参数,例如T2WI信号强度[9, 10]、表观扩散系数[11]和超声灌注参数[12, 13]等常用于消融率的预测中,但这些参数会受到多种因素的干扰,即便是术前具有相同信号的子宫肌瘤,术后消融率仍有较大差异,缺乏精准定量测量。而影像组学基于组织异质性分析,高通量地从影像数据中挖掘定量特征,用于临床决策支持系统[14],反映子宫肌瘤本身的纹理变化。MRI因其对比度分辨率高、多平面和多参数成像等特点,成为评价子宫肌瘤较准确的成像方式[15]

       然而基于T2加权脂肪抑制成像(T2-weighted imaging fat suppression,T2WI-FS)影像组学模型预测HIFU疗效的研究较少,模型也较单一,且有研究[16]未单独进行模型效能验证。因此本研究旨在进一步完善试验,探讨基于T2WI-FS影像组学联合影像学特征建立的两种联合模型在预测HIFU消融子宫肌瘤疗效中的价值。

1 材料和方法

1.1 研究对象

       本研究经川北医学院附属医院医学伦理委员会批准,免除受试者知情同意,批准文号:2022ER017-1。回顾性分析2019年11月至2021年8月在本院临床确诊为子宫肌瘤且行高强度聚焦超声(high intensity focused ultra-sound,HIFU)消融的患者资料,于术前和术后3 d内行MRI,共收集142例,包括172个肌瘤,选用术前MR图像勾画靶区。

1.2 纳入和排除标准

       纳入标准:(1)>18岁,绝经前或围绝经期女性;(2)临床及影像学检查确诊为子宫肌瘤,并明确肌瘤数量、位置、大小及性质;(3)行HIFU治疗的子宫肌瘤个数不超过10,3 cm≤直径≤10 cm;(4)既往无相关手术或药物治疗史;(5)非月经期。排除标准:(1)妊娠期及哺乳期妇女;(2)合并其他子宫或附件疾病;(3)合并心脏、肝脏、肾脏等脏器功能衰竭;(4)子宫肌瘤提示有恶性倾向;(5)广泛的腹部瘢痕组织;(6)图像质量差无法勾画靶区。

1.3 分组

       有研究指出HIFU疗效与消融率相关,消融率≥70%的肌瘤在术后一年内体积明显缩小[17],且2年累积临床复发率<10%[18]。因此本研究中以术后消融率70%为界,分为两组:低消融率组(消融率<70%)和高消融率组(消融率≥70%)。其中低消融率组77个肌瘤,高消融率组95个肌瘤。

1.4 检查方法

       MRI采用GE Discovery MR750 3.0 T,32通道体部相控阵列线圈。扫描序列和参数见表1。增强扫描在注射对比剂后约15 s (动脉早期)、30 s (动脉中期)、45 s (动脉晚期)进行,对比剂钆贝葡胺以2.0 mL/s的流率静脉注射(0.1 mmol/kg,上海博莱科信谊药业有限责任公司),后用10 mL生理盐水冲洗。动脉晚期图像用于影像学特征的获取。

表1  MRI各序列主要参数
Tab. 1  MRI main parameters of each sequence

1.5 影像学特征观察指标

       收集可能与HIFU消融疗效有关的影像学特征,包括子宫肌瘤体积、T2WI-FS信号强度(低信号:肌瘤信号强度低于或等于骨骼肌信号强度;等信号:肌瘤信号强度高于骨骼肌,但低于子宫肌层信号强度;高信号:肌瘤信号强度等于或高于子宫肌层信号强度)、彩色多普勒血流成像(color Doppler flow imaging,CDFI)信号(少量:Adler I级,可见1~2个短棒状或点状血流信号;稍丰富:Adler Ⅱ级,可见1个较长血管或3~4个点状血管,长血管长度可超过或接近肿块半径;丰富:Adler Ⅲ级,可见2个较长血管或5个以上点状血管)、T1WI增强肌瘤强化程度(轻度:肌瘤强化程度低于子宫肌层;中等:肌瘤强化程度与子宫肌层相当;明显:肌瘤强化程度高于子宫肌层)、T1WI增强肌瘤信号均匀度、子宫肌瘤类型。

1.6 图像感兴趣区分割及影像组学特征提取

       两名具有丰富盆腔诊断经验的放射医师采用开源软件3D Slicer (V4.10.2,https://www.slicer.org)进行感兴趣区(region of interest,ROI)分割及特征提取(图1)。沿子宫肌瘤边缘勾画ROI包括病变的囊性和坏死区,舍弃边缘轮廓显示不清的图像。从ROI中提取7个常见特征组:形状特征、灰度共生矩阵(gray level co-occurrence matrix,GLCM)、灰度游程长度矩阵(gray level run-length matrix,GLRLM)、灰度大小区域矩阵(gray level size zone matrix,GLSZM)、邻域灰度差矩阵(neighboring gray tone difference matrix,NGTDM)、灰度依赖矩阵(gray level dependence matrix,GLDM)及一阶特征。

图1  勾画ROI示意图。灰色部分表示勾画的靶区,选用矢状位T2WI-FS序列,沿子宫肌瘤边缘勾画。
Fig. 1  Delineation of ROI. The gray part indicates the outlined target area, and sagittal T2WI-FS sequence is selected to outline along the edge of hysteromyoma.

1.7 评估测量者间的一致性检验

       从全部对象中随机抽取约1/3的T2WI-FS序列图像,由两名医师同时勾画感兴趣区并提取特征,进行观者间影像组学特征的一致性评价(interclass correlation coefficients,ICC),ICC>0.75认为具有较好的一致性。

1.8 影像组学特征筛选

       为筛选出可重复性好、信息量大且无冗余的特征,减少过拟合问题的产生,本研究针对T2WI-FS序列的影像组学特征,采用方差阈值法、单变量选择法和最小绝对收缩和选择算子(least absolute shrinkage and selection operator,LASSO)实现特征筛选。

1.9 预测模型的构建

       数据集以7∶3的比例随机分成训练集和测试集,训练集∶测试集=120 (高消融率∶低消融率=67∶53)∶52 (高消融率∶低消融率=28∶24)。将降维后保留的最佳影像组学特征和两组间差异有统计学意义的影像学特征采用Logistic回归和随机森林分别建立两种联合模型。绘制出模型的AUC、敏感度、特异度、准确度及精确度等参数指标来对模型的预测价值进行评估,通过Delong 检验对比两种模型AUC的差值,P<0.05为差异有统计学意义。影像组学特征分析、联合模型建立及模型内部验证,均在联影uAI Research Portal (V730)软件完成(图2)。

图2  影像组学研究流程图。注:包括感兴趣区域分割、特征提取、特征选择、模型建立及验证4个部分。
Fig. 2  Flow chart of radiomics research. Including four parts: ROI segmentation, feature extraction, feature selection, model establishment and verification.

1.10 统计学分析

       常规影像参数采用开源软件R (V4.1.1,https://www.r-project.org)语言进行统计学分析。采用Kolmogorov-Smirnov方法对计量资料进行正态性检验,符合正态分布的计量资料以x¯±s表示;不符合正态分布的计量资料以M (QR)表示。两组间比较采用Wilcoxon秩和检验。计数资料以例数或构成比表示,两组间比较采用χ2检验或连续校正χ2检验。

2 结果

2.1 一般资料

       本研究共纳入142例患者,包括172个肌瘤。患者年龄27~58岁,低消融率组年龄(44.86±5.70)岁,高消融率组年龄(43.86±5.64)岁。

2.2 影像学特征分析

       低消融率组和高消融率组的影像学特征分析见表2。两组子宫肌瘤体积、T2WI-FS信号强度、T1WI增强肌瘤强化程度差异有统计学意义(P<0.05),CDFI信号、子宫肌瘤类型、T1WI增强肌瘤信号均匀度差异无统计学意义(P>0.05)。将上述两个组间差异有统计学意义的影像学特征纳入联合模型。

表2  影像学特征分析结果
Tab. 2  Analysis results of imaging features

2.3 影像组学特征分析

       通过观察者间特征ICC一致性评价,将ICC>0.75的特征保留,剩余808个稳定特征(图3)。采用方差阈值法,将特征方差低于0.8的去除;采用selectKBest方法,将在低消融率组和高消融率组间差异无统计学意义(P>0.05)的特征去除;最后采用最小绝对收缩和选择算子(least absolute shrinkage and selection operator,LASSO)算法,筛选出预测HIFU消融疗效最为重要的特征(图4)。LASSO选择的非零系数特征名称及各特征系数(表3)。

图3  组间一致性检验示意图。ICC>0.75(红线以上)表示特征具有较好一致性,筛选后剩余808个稳定特征。
Fig. 3  Consistency test. ICC>0.75 (above the red line) indicates that the features have good consistency, and 808 stable features remain after screening.
图4  LASSO系数变化曲线图。纵轴为系数,横轴表示-Log (Alpha)值。利用LASSO算法筛选影像组学特征,当Alpha=0.301时,共12个影像组学特征系数不等于0。该图展示-Log(0.301)时,12个不等于0的影像组学特征的LASSO系数变化曲线。
Fig. 4  LASSO coefficient variation curve. LASSO algorithm is used to screen radiomics features. When alpha=0.301, a total of 12 feature coefficients are not equal to 0. This figure shows -Log (0.301), the LASSO coefficient variation curve of 12 radiomics features that are not equal to 0.
表3  筛选后12个影像组学特征及系数
Tab. 3  12 radiomics features and coefficients after screening

2.4 联合模型建立及评估

       通过Logistic回归和随机森林分别建立基于12个影像组学特征和3个影像学特征的联合模型。两种模型在测试集中AUC分别为0.855、0.796。训练集和测试集Delong检验结果显示两种联合模型预测效能差异无统计学意义(P>0.05)。详见表4

表4  两种模型预测效能及Delong检验结果
Tab. 4  Prediction efficiency and Delong test results of the two models

3 讨论

       本研究基于T2WI-FS的12个影像组学特征和3个影像学特征,采用Logistic回归和随机森林构建了两种模型,用于无创性预测子宫肌瘤患者HIFU消融疗效。发现两种模型在训练集和测试集中均取得较好效能。本研究创新地将影像学特征与影像组学联合并使用不同机器学习模型,在定性的基础上加入定量测量,更准确地预测了疗效,为术前评估HIFU消融难度提供新方法,对于指导子宫肌瘤的治疗具有一定临床价值。

3.1 相关研究比较

       目前,已有部分研究探讨了影像组学在预测HIFU疗效中的价值。如Zheng等[19]在研究中基于T2WI序列和DWI序列提取影像组学特征,通过多种机器学习分类器构建HIFU消融疗效预测模型,其结果表明支持向量机构建的模型预测效能较优。韦超等[16]基于临床特征和影像组学特征构建消融率多元线性联合回归模型,发现子宫肌瘤HIFU术后实际消融率和预测消融率相关性较高,具有一定的潜在应用价值。但是韦等未单独划分数据集以证明模型的稳定性及消除模型潜在的过拟合问题。本研究将数据集以7∶3的比例随机分成训练集和测试集,在训练集和测试集中模型的AUC相近,在一定程度上避免了过拟合问题、评估了模型的泛化能力。并进一步证实了影像组学模型的可靠性和稳定性以及其在预测疗效方面的应用前景。

3.2 影像学特征研究结果分析

       本研究中低消融率组和高消融率组子宫肌瘤体积和T2WI-FS信号强度、T1WI增强肌瘤强化程度差异有统计学意义。说明T2WI-FS高信号、T1WI增强明显强化和体积较小的子宫肌瘤不利于HIFU消融治疗,这一发现与之前的研究[20, 21]一致。T2WI高信号表明子宫肌瘤内血管生成、细胞密集[22],T1WI强化程度越高表明子宫肌瘤内血液供应越丰富[20]。超声能量难以聚焦于这两种类型的子宫肌瘤,不利于消融治疗。然而信号强度会受到组织类型、主磁场强度和磁场均匀度的影响,且测量时受主观因素影响较大,增加了预测结果的不确定性。

3.3 影像组学在预测子宫肌瘤HIFU消融疗效中的应用价值分析

       影像组学可以揭示肿瘤空间分布的异质性,纹理特征在一定程度上反映组织的病理学改变[23]。子宫肌瘤的病理变化在图像信号、形状、纹理等方面引起的微小差异可以通过影像组学的方法进行量化。同时,影像组学有可能成为基于人工智能的生物标记物,非侵入性捕捉子宫肌瘤的异质性。相较于T1WI增强序列,T2WI-FS序列不需要使用对比剂,对肝肾的损伤较小,其信号受磁场强度的影响较小,获得的图像更加稳定、可靠,且大部分组织信号的改变主要显示于T2WI-FS序列。不仅如此,T2WI-FS序列作为评价子宫肌瘤重要的基础序列,其信号变化与子宫肌瘤病理分型有一定关系。因此本文选择基于T2WI-FS序列提取影像组学特征建立联合模型。本研究提取了12个影像组学特征,包括5个CLCM、2个GLDM、2个GLRLM、1个NGTDM、1个GLSZM和1个一阶特征。其中GLCM反映了灰度图像中不同体素的排列规则及图像灰度在步长、方向等变化的综合信息,GLCM中熵越高,说明肿瘤的异质性越高[24, 25]。而肿瘤的异质性可能来自于肿瘤内部细胞变化、血管生成和细胞外基质[26],本研究中低消融率组GLCM的熵大于高消融率组,提示异质性较高的子宫肌瘤不适于HIFU消融治疗,这可能是因为低消融率的子宫肌瘤内细胞数目增多、血管生成导致异质性升高。疗效预测一直都是HIFU技术临床应用面临的挑战之一,术前预判消融疗效将有助于临床医生选择最适于HIFU治疗的患者,从而尽可能保证成功率,降低失败风险,使用最少的医疗成本却能最大程度地减轻患者痛苦。

3.4 局限性

       本研究中仍存在一些不足:(1)数据均来源于单中心,在患者选择方面可能存在偏移;(2)样本量较小;(3)仅纳入单一序列建模,获取信息可能不够全面。因此,期待未来能纳入更大样本的多中心数据,基于多参数MRI建立更加综合、完善的预测模型。

       综上所述,基于T2WI-FS序列建立的影像组学联合模型,对于预测子宫肌瘤HIFU消融疗效具有良好的效能,将有助于临床医生更好地筛选HIFU治疗中获益最大的患者,为治疗决策提供参考,制订精准诊疗方案。

[1]
Stewart EA, Cookson CL, Gandolfo RA, et al. Epidemiology of uterine fibroids: a systematic review[J]. BJOG, 2017, 124(10): 1501-1512. DOI: 10.1111/1471-0528.14640.
[2]
Giuliani E, As-Sanie S, Marsh EE. Epidemiology and management of uterine fibroids[J]. Int J Gynaecol Obstet, 2020, 149(1): 3-9. DOI: 10.1002/ijgo.13102.
[3]
Lyon PC, Rai V, Price N, et al. Ultrasound-guided high intensity focused ultrasound ablation for symptomatic uterine fibroids: preliminary clinical experience[J]. Ultraschall Med, 2020, 41(5): 550-556. DOI: 10.1055/a-0891-0729.
[4]
Lee KW. The Asian perspective on HIFU[J]. Int J Hyperthermia, 2021, 38(2): 5-8. DOI: 10.1080/02656736.2021.1889697.
[5]
Wang CJ, Lin G, Huang YT, et al. A feasibility analysis of the ArcBlate MR-guided high-intensity focused ultrasound system for the ablation of uterine fibroids[J]. Abdom Radiol (NY), 2021, 46(11): 5307-5315. DOI: 10.1007/s00261-021-03203-8.
[6]
Fennessy FM, Tempany CM, McDannold NJ, et al. Uterine leiomyomas: MR imaging-guided focused ultrasound surgery: results of different treatment protocols[J]. Radiology, 2007, 243(3): 885-893. DOI: 10.1148/radiol.2433060267.
[7]
Łoziński T, Ciebiera M, Łuczyńska E, et al. Magnetic resonance-guided high-intensity focused ultrasound ablation of uterine fibroids-efficiency assessment with the use of dynamic contrast-enhanced magnetic resonance imaging and the potential role of the administration of uterotonic drugs[J]. Diagnostics (Basel), 2021, 11(4): 715. DOI: 10.3390/diagnostics11040715.
[8]
Wei C, Fang X, Wang CB, et al. The predictive value of quantitative DCE metrics for immediate therapeutic response of high-intensity focused ultrasound ablation (HIFU) of symptomatic uterine fibroids[J]. Abdom Radiol (NY), 2018, 43(8): 2169-2175. DOI: 10.1007/s00261-017-1426-7.
[9]
Gorny KR, Woodrum DA, Brown DL, et al. Magnetic resonance-guided focused ultrasound of uterine leiomyomas: review of a 12-month outcome of 130 clinical patients[J]. J Vasc Interv Radiol, 2011, 22(6): 857-864. DOI: 10.1016/j.jvir.2011.01.458.
[10]
Zhao WP, Zhang J, Han ZY, et al. A clinical investigation treating different types of fibroids identified by MRI-T2WI imaging with ultrasound guided high intensity focused ultrasound[J]. Sci Rep, 2017, 7(1): 10812. DOI: 10.1038/s41598-017-11486-5.
[11]
Sainio T, Saunavaara J, Komar G, et al. Feasibility of apparent diffusion coefficient in predicting the technical outcome of MR-guided high-intensity focused ultrasound treatment of uterine fibroids-a comparison with the Funaki classification[J]. Int J Hyperthermia, 2021, 38(1): 85-94. DOI: 10.1080/02656736.2021.1874545.
[12]
Wang YJ, Zhang PH, Zhang R, et al. Predictive value of quantitative uterine fibroid perfusion parameters from contrast-enhanced ultrasound for the therapeutic effect of high-intensity focused ultrasound ablation[J]. J Ultrasound Med, 2019, 38(6): 1511-1517. DOI: 10.1002/jum.14838.
[13]
Luo W, Zhang PD, Yang X, et al. Role of contrast-enhanced ultrasonography in MR-guided focused ultrasound ablation on uterus fibroids: lesion selection and assessment of ablative effects[J]. Eur Radiol, 2022, 32(3): 2110-2119. DOI: 10.1007/s00330-021-08294-x.
[14]
Lambin P, Leijenaar RTH, Deist TM, et al. Radiomics: the bridge between medical imaging and personalized medicine[J]. Nat Rev Clin Oncol, 2017, 14(12): 749-762. DOI: 10.1038/nrclinonc.2017.141.
[15]
Testa AC, di Legge A, Bonatti M, et al. Imaging techniques for evaluation of uterine myomas[J]. Best Pract Res Clin Obstet Gynaecol, 2016, 34: 37-53. DOI: 10.1016/j.bpobgyn.2015.11.014.
[16]
韦超. 基于常规MRI和T2WI-影像组学对子宫肌瘤病理分型、HIFU消融难度和即刻消融率的预测研究[D]. 合肥: 安徽医科大学, 2021.
Wei C. Prediction of pathological classification, ablation difficulty and immediate ablation rate of uterine fibroids treated by HIFU based on conventional MRI and T2WI-radiomics features [D]. Hefei: Anhui Medical University, 2021.
[17]
秦娟, 陈文直, 陈锦云, 等. 超声消融子宫肌瘤后残余瘤组织的核磁共振影像学转归[J]. 中国肿瘤临床, 2013, 10(2): 111-115. DOI: 10.3969/j.issn.1000-8179.2013.02.013.
Qin J, Chen WZ, Chen JY, et al. Magnetic resonance imaging outcome of residual tumor after ultrasound ablation for uterine fibroids[J]. Cancer Biol Med, 2013, 10(2): 111-115. DOI: 10.3969/j.issn.1000-8179.2013.02.013.
[18]
Stewart EA, Gostout B, Rabinovici J, et al. Sustained relief of leiomyoma symptoms by using focused ultrasound surgery[J]. Obstet Gynecol, 2007, 110(2 Pt 1): 279-287. DOI: 10.1097/01.AOG.0000275283.39475.f6.
[19]
Zheng YN, Chen LP, Liu MQ, et al. Nonenhanced MRI-based radiomics model for preoperative prediction of nonperfused volume ratio for high-intensity focused ultrasound ablation of uterine leiomyomas[J]. Int J Hyperthermia, 2021, 38(1): 1349-1358. DOI: 10.1080/02656736.2021.1972170.
[20]
Liu ZQ, Gong CM, Liu YC, et al. Establishment of a scoring system for predicting the difficulty level of high-intensity focussed ultrasound ablation of uterine fibroids[J]. Int J Hyperthermia, 2018, 34(1): 77-86. DOI: 10.1080/02656736.2017.1325015.
[21]
Keserci B, Duc NM. The role of T1 perfusion-based classification in magnetic resonance-guided high-intensity focused ultrasound ablation of uterine fibroids[J]. Eur Radiol, 2017, 27(12): 5299-5308. DOI: 10.1007/s00330-017-4885-x.
[22]
Huang HR, Ran JB, Xiao ZB, et al. Reasons for different therapeutic effects of high-intensity focused ultrasound ablation on excised uterine fibroids with different signal intensities on T2-weighted MRI: a study of histopathological characteristics[J]. Int J Hyperthermia, 2019, 36(1): 477-484. DOI: 10.1080/02656736.2019.1592242.
[23]
赖树盛, 郑石磊. 基于磁共振T2WI-FS的影像组学对前列腺癌盆腔淋巴结转移的诊断价值[J]. 中国医科大学学报, 2021, 50(3): 230-234. DOI: 10.12007/j.issn.0258-4646.2021.03.008.
Lai SS, Zheng SL. Diagnostic value of radiomics based on fat-suppressed T2-weighted imaging magnetic resonance imaging for pelvic metastatic lymph nodes of prostate cancer[J]. J China Med Univ, 2021, 50(3): 230-234. DOI: 10.12007/j.issn.0258-4646.2021.03.008.
[24]
彭文静, 徐凯, 刘正立, 等. 乳腺X线图像纹理分析鉴别乳腺小结节良恶性的价值[J]. 中国医学影像学杂志, 2018, 26(12): 890-893, 898. DOI: 10.3969/j.issn.1005-5185.2018.12.003.
Peng WJ, Xu K, Liu ZL, et al. Mammogram texture analysis in differential diagnosis of benign and malignant breast nodules[J]. Chin J Med Imaging, 2018, 26(12): 890-893, 898. DOI: 10.3969/j.issn.1005-5185.2018.12.003.
[25]
李华秀, 李振辉, 李鹍, 等. CT影像组学预测局部进展期直肠癌新辅助治疗的效果[J]. 中国医学影像学杂志, 2020, 28(1): 44-50. DOI: 10.3969/j.issn.1005-5185.2020.01.010.
Li HX, Li ZH, Li K, et al. Efficacy of CT radiomics in predicting neoadjuvant therapy of locally advanced rectal cancer[J]. Chin J Med Imaging, 2020, 28(1): 44-50. DOI: 10.3969/j.issn.1005-5185.2020.01.010.
[26]
李志恒, 卢增新, 赵振华, 等. ADC图影像组学预测局部晚期直肠癌患者新辅助化疗的疗效[J]. 中国临床医学影像杂志, 2021, 32(11): 817-822. DOI: 10.12117/jccmi.2021.11.013.
Li ZH, Lu ZX, Zhao ZH, et al. Predicting the curative effect of neoadjuvant chemotherapy for locally advanced rectal cancer by radiomics of ADC maps[J]. J Chin Clin Med Imaging, 2021, 32(11): 817-822. DOI: 10.12117/jccmi.2021.11.013.

上一篇 术前MRI影像组学列线图对乳腺癌腋窝淋巴结转移的预测价值
下一篇 2型糖尿病患者小脑Crus Ⅰ静息态磁共振功能连接研究
  
诚聘英才 | 广告合作 | 免责声明 | 版权声明
联系电话:010-67113815
京ICP备19028836号-2