本文已被：浏览 1387次 下载 371次
Received:April 29, 2022 Published Online:January 20, 2023
Received:April 29, 2022 Published Online:January 20, 2023
中文摘要: 目的 探讨从T2WI及动态增强磁共振图像中获得的影像组学特征在区分肾细胞癌（RCC）三种亚型中的价值。
方法回顾性收集安徽医科大学第三附属医院2014年3月至2020年4月经术后病理证实的84例RCC且接受术前磁共振成像（MRI）检查患者的临床影像资料,84例中，透明细胞肾细胞癌（ccRCC）46例、乳头状肾细胞癌（pRCC）20例和嫌色细胞肾细胞癌（cRCC）18例。利用3D-Slicer软件在三个序列（T2WI、EN-T1WI皮质期和EN-T1WI髓质期）上对肿瘤三维全层勾画感兴趣区（ROI），利用Python软件从肿瘤体积中提取影像组学特征。使用组内组间相关分析计算每个特征组内组间相关系数（ICC），选取ICC＞0.75的特征作为可重复提取的稳定特征。将肿块随机分为训练集和验证集（约6∶4），使用Kruskal-wallis检验筛选出每个MRI序列鉴别RCC亚型的最佳纹理特征，使用Countif函数对特征子集进行筛选，取三个序列的最佳特征，利用所筛选的基于影像组学的最佳特征分别建立T2WI、EN-T1WI皮质期和EN-T1WI髓质期三个序列的logistic回归模型。报告测试集三种亚型在三个序列的曲线下面积（AUC）、敏感度和特异度。
结果三种亚型在三个序列有显著差异的影像组学特征共16个，T2WI、EN-T1WI皮质期、EN-T1WI髓质期序列在区分ccRCC和pRCC时的AUC分别为0.833、0.895和0.885;区分ccRCC和cRCC时的AUC分别为0.822、0.856和0.766;区分pRCC和cRCC时的AUC分别为0.857、0.881和0.857。
结论基于MRI中所获得的影像组学资料，T2WI、EN-T1WI皮质期和EN-T1WI髓质期影像组学模型都可以很好的区分ccRCC、pRCC和cRCC，且以EN-T1WI皮质期诊断效能最佳。
Abstract:ObjectiveToinvestigatethevalueofradiomicsfeaturesobtainedfromT2-weightedimaging（T2WI）anddynamiccontrast-enhancedmagneticresonanceimaging（DCE-MRI）indifferentiatingtherenalcellcarcinoma（RCC）subtypes.
MethodsTheclinicalandimagingdatawereretrospectivelycollectedfrom84RCCpatientsconfirmedbypostoperativepathologyandpreoperativeMRIfromMarch2014toApril2020intheThirdAffiliatedHospitalofAnhuiMedicalUniversity.Therewere46casesofclearcellrenalcellcarcinoma（ccRCC）,20casesofpapillaryrenalcellcarcinoma（pRCC）and18casesofchromophoberenalcellcarcinoma(cRCC).Thethree-dimensionalfull-layerregionofinterest（ROI）ofwholetumorwasdelineatedonthreesequences（T2WI,EN-T1WIincorticalphaseandEN-T1WIinmedullaryphase）using3DSlicersoftware,andtheradiomicsfeatureswereextractedfromthetumorvolumeusingPythonsoftware.Thecorrelationanalysiswasusedtocalculateintra-groupandinter-groupcorrelationcoefficient（ICC）foreachfeaturegroup,andfeatureswithICCvaluesgreaterthan0.75wereconsideredreproducibleandstablefeaturesthatcouldbeextractedrepeatedly.Thelesionswererandomlydividedintotrainingsetandtestsetaccordingtotheproportionof6∶4,andthebesttexturefeaturesforeachMRIsequencetoidentifyRCCsubtypeswerescreenedusingKruskal-wallistest.CountiffunctionwasusedtoscreenfeaturesubsetsforthebestfeaturesselectionofthreesequencestoestablishthelogisticregressionmodelsofT2WIandEN-T1WIcorticalphaseandEN-T1WImedullaryphase.TheAUC,sensitivityandspecificityofthethreesubtypesofthetestsetinthreesequenceswerecalculatedandreported.
ResultsTherewere16radiomicsfeaturesofthethreesubtypeswithsignificantdifferencesinthethreesequences.AUCsofT2WIandEN-T1WIcorticalphaseandEN-T1WImedullaryphasesequencewere0.833,0.895and0.885indistinguishingccRCCandpRCC,0.822,0.856and0.766forccRCCandcRCC,and0.857,0.881and0.857forpRCCandcRCC.
ConclusionBasedonradiomicsfeaturesobtainedfromMRI,T2WIandEN-T1WIcorticalphaseandEN-T1WImedullaryphaseradiomicsmodelscanwelldistinguishccRCC,pRCCandcRCC,andEN-T1WIcorticalphasehasthebestdiagnosticperformance.
文章编号： 中图分类号：R737.11 R445.2 文献标志码：A
基金项目:合肥市科技局科研合作项目（YW201608080003）
| Author Name | Affiliation |
| LIZeng-hua,XIAChun-hua,HUDa-tao,HUANGDan-dan,FENGQian-ru，WANGYa-qi | DepartmentofRadiology，TheThirdAffiliatedHospitalofAnhuiMedicalUniversity,Hefei,Anhui230000,China |
| Author Name | Affiliation |
| LIZeng-hua,XIAChun-hua,HUDa-tao,HUANGDan-dan,FENGQian-ru，WANGYa-qi | DepartmentofRadiology，TheThirdAffiliatedHospitalofAnhuiMedicalUniversity,Hefei,Anhui230000,China |
引用文本：