首页|基于同步辐射光源微CT技术及分水岭图像分割方法量化乳腺癌胫骨转移小鼠模型局部骨组织及血管

基于同步辐射光源微CT技术及分水岭图像分割方法量化乳腺癌胫骨转移小鼠模型局部骨组织及血管

扫码查看
目的 观察基于同步辐射光源微CT(SR-μCT)技术及分水岭图像分割方法量化乳腺癌胫骨转移小鼠模型胫骨局部三维骨组织和血管的价值.方法 建立乳腺癌胫骨转移小鼠模型,分别予口服Vatalanib、注射Bevacizumab、联合Vatalanib与Bevacizumab治疗及注射安慰剂.于注射肿瘤细胞后第8天(T1)及第22天(T2)结合注射血管对比剂和SR-μCT对注射部位骨组织和血管网络行高分辨力同步成像,以自动化标记的分水岭图像分割方法、全局阈值方法及区域生长方法分割局部骨组织和血管,利用戴斯相似系数(DSC)评估图像分割质量;获取局部胫骨解剖学量化参数,包括骨组织体积分数(BV/TV)、血管体积分数(VV/TV)、骨组织平均厚度(B.Th)及血管平均厚度(V.Th),比较相同时间点、不同治疗方案之间,以及相同治疗方案、不同时间点之间解剖学量化参数.结果 基于标记分水岭算法分割骨组织和血管结构的DSC达0.95以上,高于其他方法(P均<0.05).不同治疗方案之间T1小鼠胫骨解剖学量化参数差异均无统计学意义(P均>0.05);联合药物治疗方案下V.Th T2结果低于其他治疗方案(P均<0.05).联合药物治疗方案下,相比T1,T2乳腺癌骨转移小鼠胫骨BV/TV、B.Th数值增大而VV/TV数值减小(P均<0.05);Vatalanib和Bevacizumab治疗方案下T2时间点小鼠胫骨BV/TV、B.Th及V.Th数值均增大(P均<0.05);安慰剂方案下BV/TV、B.Th数值增大(P均<0.05).结论 结合SR-μCT技术和自动化标记分水岭算法可初步量化评估乳腺癌胫骨转移小鼠胫骨局部骨组织和血管三维结构.
Synchrotron radiation micro-CT and automatic marker watershed image segmentation for quantifying local bone tissue and blood vessels of tibia in mouse models of breast cancer tibia metastasis
Objective To explore the value of synchrotron radiation micro-CT(SR-pCT)and automatic marker watershed image segmentation for quantifying three-dimensional local bone tissue and blood vessels in mouse models of breast cancer tibia metastasis.Methods Mouse models with breast cancer bone metastasis in tibia were established and underwent oral Vatalanib,injection of Bevacizumab,combination of Vatalanib or Bevacizumab,or injection of placebo,respectively.On the 8th day(time point T1)and the 22nd day(time point T2)after injection of tumor cells,the tibia bone was scanned using combined vascular contrast injection and SR-pCT,and high-resolution simultaneous imaging of local bone tissue and vascular network were achieved.Then bone tissue and vessels on images were segmented using automatic marker watershed image-segmentation method,global threshold method and region growing method,respectively,and the quality of segmentations were assessed with Dice similarity coefficient(DSC).The anatomical quantitative parameters of tibia with breast cancer metastasis were obtained,including the fraction of bone volume in total volume(BV/TV),fraction of vessel volume in total volume(VV/TV),the mean thickness of bone(B.Th)and the mean thickness of vessel(V.Th).The anatomical quantitative parameters were compared among different treatments at the same time point,also between different time points under the same treatment.Results DSC for segmentation of local tibia bone tissue and vascular structures based on automatic marker watershed image segmentation method were all greater than 0.95,higher than that of other methods(all P<0.05).No significant difference of anatomical quantitative parameters was found under different treatments at T1(all P>0.05),and V.Th under the combined treatment were lower than that under other treatments at T2(all P<0.05).Compared with T1,BV/TV and B.Th at T2 increased while VV/TV at T2 decreased in tibia metastasis under combination treatment(all P<0.05),BV/TV,B.Th and V.Th at T2 increased under Vatalanib or Bevacizumab treatment(all P<0.05),and BV/TV and B.Th at T2 increased under placebo(both P<0.05).Conclusion The combination of SR-pCT and automatic marker watershed method could be used to initially quantify three-dimensional structures of local bone tissue and blood vessels in mouse models of breast cancer tibia metastasis.

X-ray micro tomographyneoplasm metastasismiceblood vesselsimage processing,computer-assisted

徐豪、赵乙頔、于露、陈丁钰、刘继琛、赵涵、孟竹、Philippe Clezardin

展开 >

沈阳大学科技创新研究院,辽宁沈阳 110044

沈阳理工大学理学院,辽宁 沈阳 110158

里昂大学LYOS实验室,罗讷 里昂 69000

X线显微体层摄影术 肿瘤转移 小鼠 血管 图像处理,计算机辅助

2024

中国医学影像技术
中国科学院声学研究所

中国医学影像技术

CSTPCD北大核心
影响因子:0.763
ISSN:1003-3289
年,卷(期):2024.40(11)