基于阻抗信息的乳腺组织BP网络识别方法
|
摘要
生物电阻抗测量技术(Bio-Impedance Measuring Technology,BIMT)是利用生物组织与器官的电特性及其变化,提取与人体生理、病理相关信息的一种无损伤检测技术。该方法具有无创、廉价、安全、操作简单和信息丰富等特点,因此在医学界,尤其是在临床诊断中受到专家学者的重视,有着广泛的应用前景。
本课题针对当前乳腺癌治疗手段中的保乳手术,将生物电阻抗测量技术应用到术中乳腺癌灶边缘界定中,实时评价手术切缘,帮助施术医师在手术中合理选择组织切除范围,从而在手术结束前清除残留病灶。这样可减轻患者心理和生理上的痛苦,具有重要的临床意义。
本课题重点研究乳腺癌组织与正常组织(腺体和脂肪组织)电特性差异,从组织频阻特性曲线中寻找特征参数,对乳腺组织进行识别。通过对临床数据的研究分析,发现组织的频阻斜率信息可以用来对组织进行辨识。因此利用特性曲线在特殊频段内的斜率值和多个连续频段内的斜率值分别对三种组织区分。针对参数多,模型未知等特点,建立BP网络模型,研究神经网络集成方法,应用神经网络知识对组织进行辨识。
通过研究,发现BP神经网络对于多个参数的输入有较好的分类能力,经过反复多次训练出来的神经网络具有良好的组织区分能力。集成后的神经网络对乳腺各种组织识别的敏感性和特异性更佳。
Bio-Impedance Measuring Technology (BIMT) is a noninvasive measuring technology, the purpose of which is to extract physiological and pathological information by detecting the electrical properties of tissues and organs. It has the feature of low-cost, safety, easy-operating and wealth of information. So it is been paid special attention by the researchers in clinical medicine areas.
Based on the bio-impedance measurement technology, this paper presents the intraoperative margin assessment of breast cancer for purpose of breast conserving surgery. It can help the surgeon judge the margin instantly and then choose the excising area reasonably. So that the remained focus could be cleaned in time. It can shorten the waiting time and lighten patients’pain in partial mastectomy.
This paper focuses on researching the differences of electrical properties between breast cancer and normal breast tissues (mammary gland and adipose). The characteristic parameters searched from the frequency-resistance curve were studied to distinguish different tissues. Based on the analysis of clinical data, frequency-resistance slope information was found good enough to be used in the recognition of breast tissues. This paper utilized the slopes of special frequencies and multi- frequencies to identify three different tissues. Because of the multi-parameters, uncertain model and so on, the BP network was established and the integration of several neural networks was researched to recognize breast tissues.
Results of the research indicated that BP network had good capability of classification while there were many input parameters. BP networks trained for many times have nice abilities to distinguish breast cancer from normal tissues. The integral neural network possessed high sensibility and specificity to the recognition of breast tissues.
本课题针对当前乳腺癌治疗手段中的保乳手术,将生物电阻抗测量技术应用到术中乳腺癌灶边缘界定中,实时评价手术切缘,帮助施术医师在手术中合理选择组织切除范围,从而在手术结束前清除残留病灶。这样可减轻患者心理和生理上的痛苦,具有重要的临床意义。
本课题重点研究乳腺癌组织与正常组织(腺体和脂肪组织)电特性差异,从组织频阻特性曲线中寻找特征参数,对乳腺组织进行识别。通过对临床数据的研究分析,发现组织的频阻斜率信息可以用来对组织进行辨识。因此利用特性曲线在特殊频段内的斜率值和多个连续频段内的斜率值分别对三种组织区分。针对参数多,模型未知等特点,建立BP网络模型,研究神经网络集成方法,应用神经网络知识对组织进行辨识。
通过研究,发现BP神经网络对于多个参数的输入有较好的分类能力,经过反复多次训练出来的神经网络具有良好的组织区分能力。集成后的神经网络对乳腺各种组织识别的敏感性和特异性更佳。
Bio-Impedance Measuring Technology (BIMT) is a noninvasive measuring technology, the purpose of which is to extract physiological and pathological information by detecting the electrical properties of tissues and organs. It has the feature of low-cost, safety, easy-operating and wealth of information. So it is been paid special attention by the researchers in clinical medicine areas.
Based on the bio-impedance measurement technology, this paper presents the intraoperative margin assessment of breast cancer for purpose of breast conserving surgery. It can help the surgeon judge the margin instantly and then choose the excising area reasonably. So that the remained focus could be cleaned in time. It can shorten the waiting time and lighten patients’pain in partial mastectomy.
This paper focuses on researching the differences of electrical properties between breast cancer and normal breast tissues (mammary gland and adipose). The characteristic parameters searched from the frequency-resistance curve were studied to distinguish different tissues. Based on the analysis of clinical data, frequency-resistance slope information was found good enough to be used in the recognition of breast tissues. This paper utilized the slopes of special frequencies and multi- frequencies to identify three different tissues. Because of the multi-parameters, uncertain model and so on, the BP network was established and the integration of several neural networks was researched to recognize breast tissues.
Results of the research indicated that BP network had good capability of classification while there were many input parameters. BP networks trained for many times have nice abilities to distinguish breast cancer from normal tissues. The integral neural network possessed high sensibility and specificity to the recognition of breast tissues.
引文
[1]周雪瑞,黄选东,乳腺与乳腺癌,生物学通报,2004,39(7):26~27
[2]天津日报,肿瘤医院一项调查表明本市乳腺癌发病率上升,天津日报,2001
[3]JM.Yuan, QS.wang, RK.Ross, Diet and breast cancer in Shanghai and Tianjin, China, 1995, 71(6): 1353~1358
[4]金宗浩,乳腺增生和乳腺癌,上海:上海科学技术文献出版社,2002,112~134
[5]张保宁,乳腺癌保乳手术的研究进展,中国普外基础与临床杂志,2005,12(3):204~206
[6]肖亚苏,基于阻抗测量的术中乳腺癌灶边缘界定电极结构优化:[硕士学位论文],天津:天津大学,2007
[7]刘君,方志沂,于泳等,乳腺癌保乳手术安全范围的研究,中国肿瘤临床,2005,32(15):856~860
[8]孙慎友,乳腺癌保留乳房手术治疗新进展,中国普通外科杂志,2004,13(5):363~366
[9]牛昀,乳腺肿瘤病理诊断学,天津科学技术出版社,2006
[10]廖琪梅,董秀珍,付峰,人体乳腺组织电阻抗特性的研究,国际生物医学工程杂志,2006,29(4):218~221
[11]Fricke H, Morse S, The electrical capacity of tumors of the breast, J Cancer Res, 1926, 10: 340~347
[12]Surowiec AJ,Stuchly SS,Barr JB, Dielectric properties of breast carcinoma and the surrounding tissues, IEEE Trans Biomed Eng, 1988, 35: 257~263
[13]Morimoto T, Kinouchi Y, Iritani T, Measurement of the electrical bio-impedance of breast tumors, Eur Surg Res , 1990, 22: 86~92
[14]Morimoto T, Kimura S, Konishi Y, Komaki K, Uyama T, Monden Y, Kinouchi Y, Iritani T. A study of the electrical bio-impedance of tumors. J Invest Surg, 1993, 6: 25~32
[15]Jossinet, J, Variability of impedivity in normal and pathological breast tissue, Med.&Biol.Eng.&Computing, 1986, 34: 346~350
[16]Jossinet, J, The impedivity of freshly excised human breast tissue, Physiol. Meas., 1998, 19: 61~75
[17]Jossinet, J, Schmitt M, A review of parameters for the bioelectricalcharacterization of breast tissue, Ann NY Acad Sci, 1999, 873: 30~41
[18]Chauveau N, Hamzaoui L, Rochaix P, Rigaud B, Voigt JJ, Morucci JP. Ex vivo discrimination between normal and pathological tissues in human breast surgical biopsies using bioimpedance spectroscopy. Ann NY Acad Sci, 1999, 873: 42~50
[19]刘锐岗,付峰,史学涛等,正常妇女电阻抗扫描乳腺检测数据的初步分析,第四军医大学学报,2004,25(21):1994~1999
[20]A. M. Dijkstra, B. H. Brown, A. D. Leathard, Review Clinical Application of Electrical Impedance Tomography, Journal of Medical Engineering & Technology, 1993, 3 (17): 89~98
[21]任超世,生物电阻抗技术与人体功能信息,电子科技导报,1998,11(1):17~19
[22]Geddes, L.A., Baker, L.E., Principles of Applied Biomedical Instrumentation, John Wiley & Sons, New York, 1968, 5 (2): 150~205
[23]Cole KS, Electric Impedance of Suspensions of Spheres, J.Gen.physicol., 1928, 12 (1): 29~36
[24]Cole KS, Cole RH, Dispersion and Absorption in Dielectrics, J.Chem.phys., 1941, 9 (4): 341~351
[25]Sverre Grimnes, Orjan Grottem Martinsen, Bio-impedance and Bioelectricity Basics, Great Britain: ACADEMIC PRESS, 2000
[26]Schwan HP, Determination of Biological Impedance, Physical techniques in biomedical research, 1963, 6 (4): 323~406
[27]章晓丽,生物阻抗动态信息提取与处理关键技术研究:[硕士学位论文],天津:天津大学,2007
[28]王春艳,应用于肺部检测的电阻成像系统:[硕士学位论文],天津:天津大学,2007
[29]Y. Zou.and Z. Guo, A review of electrical impedance techniques for breast cancer detection, Medical engineering & physics, 2003, 25: 79~90
[30]Blad B, Baldetop B, Impedance spectra of tumor tissue in comparison with normal tissue: a possible clinical application for electrical impedance tomography, Physiol Meas, 1996, 17: A105~A115
[31]孙高,周宝森,杨静宇等,基于人工神经网络的肺癌诊断研究,中国卫生统计,1999,16(3):142~144
[32]洪文学,李佳书,严衍玲,人工神经网络在肿瘤检测和诊断中的应用研究进展,北京生物医学工程,2007,26(4):446~448
[33]黎衍云,李锐,张胜年,人工神经网络及其在疾病筛查中应用前景,环境与职业医学习,2006,23(1):71~73
[34]华贻军,人工神经网络在临床医学中的应用,实用医学杂,2005,21(3):330~332
[35]陈新平,临床医学中的神经网络技术,中国现代医学杂志,2003,13(9):46~50
[36]吴拥军,吴逸明,屈凌波等,人工神经网络技术在肺癌诊断中的应用研究,中华微生物学和免疫学杂志,2003,23(8):646~649
[37]李晓薇,刘文凯,李军会,人工神经网络在恶性肿瘤诊断中的应用,中国农业大学学报,2001,6(3):6~10
[38]刘琼荪,何离庆,基于人工神经网络的乳腺癌诊断模型,重庆大学学报,2003,26(4):70~72
[39]吴鹭萍,人工神经网络在医学上的应用,医学信息:医学与计算机应用,2000,13(7):368~370
[40]周开利,康耀红,神经网络模型及其MATLAB仿真程序设计,北京:清华大学出版社,2005
[41]李丽霞,王彤,范逢曦,BP神经网络设计探讨,现代预防医学,2005,32(2):128~130
[42]焦李成,神经网络系统理论,西安:西安电子科技大学出版社,1990
[43]韩力群,人工神经网络教程,北京:北京邮电大学出版社,2006
[44]闻新等,Matlab神经网络仿真与应用,北京:科学出版社,2003
[45]高大文,王鹏,蔡臻超,人工神经网络中隐含层节点与训练次数的优化,哈尔滨工业大学学报,2003,35(2):207~209
[46]邓丽,吴春梅,罗芳琼,神经网络集成及研究进展,柳州师专学报,2007,22(4):4741~4746
[2]天津日报,肿瘤医院一项调查表明本市乳腺癌发病率上升,天津日报,2001
[3]JM.Yuan, QS.wang, RK.Ross, Diet and breast cancer in Shanghai and Tianjin, China, 1995, 71(6): 1353~1358
[4]金宗浩,乳腺增生和乳腺癌,上海:上海科学技术文献出版社,2002,112~134
[5]张保宁,乳腺癌保乳手术的研究进展,中国普外基础与临床杂志,2005,12(3):204~206
[6]肖亚苏,基于阻抗测量的术中乳腺癌灶边缘界定电极结构优化:[硕士学位论文],天津:天津大学,2007
[7]刘君,方志沂,于泳等,乳腺癌保乳手术安全范围的研究,中国肿瘤临床,2005,32(15):856~860
[8]孙慎友,乳腺癌保留乳房手术治疗新进展,中国普通外科杂志,2004,13(5):363~366
[9]牛昀,乳腺肿瘤病理诊断学,天津科学技术出版社,2006
[10]廖琪梅,董秀珍,付峰,人体乳腺组织电阻抗特性的研究,国际生物医学工程杂志,2006,29(4):218~221
[11]Fricke H, Morse S, The electrical capacity of tumors of the breast, J Cancer Res, 1926, 10: 340~347
[12]Surowiec AJ,Stuchly SS,Barr JB, Dielectric properties of breast carcinoma and the surrounding tissues, IEEE Trans Biomed Eng, 1988, 35: 257~263
[13]Morimoto T, Kinouchi Y, Iritani T, Measurement of the electrical bio-impedance of breast tumors, Eur Surg Res , 1990, 22: 86~92
[14]Morimoto T, Kimura S, Konishi Y, Komaki K, Uyama T, Monden Y, Kinouchi Y, Iritani T. A study of the electrical bio-impedance of tumors. J Invest Surg, 1993, 6: 25~32
[15]Jossinet, J, Variability of impedivity in normal and pathological breast tissue, Med.&Biol.Eng.&Computing, 1986, 34: 346~350
[16]Jossinet, J, The impedivity of freshly excised human breast tissue, Physiol. Meas., 1998, 19: 61~75
[17]Jossinet, J, Schmitt M, A review of parameters for the bioelectricalcharacterization of breast tissue, Ann NY Acad Sci, 1999, 873: 30~41
[18]Chauveau N, Hamzaoui L, Rochaix P, Rigaud B, Voigt JJ, Morucci JP. Ex vivo discrimination between normal and pathological tissues in human breast surgical biopsies using bioimpedance spectroscopy. Ann NY Acad Sci, 1999, 873: 42~50
[19]刘锐岗,付峰,史学涛等,正常妇女电阻抗扫描乳腺检测数据的初步分析,第四军医大学学报,2004,25(21):1994~1999
[20]A. M. Dijkstra, B. H. Brown, A. D. Leathard, Review Clinical Application of Electrical Impedance Tomography, Journal of Medical Engineering & Technology, 1993, 3 (17): 89~98
[21]任超世,生物电阻抗技术与人体功能信息,电子科技导报,1998,11(1):17~19
[22]Geddes, L.A., Baker, L.E., Principles of Applied Biomedical Instrumentation, John Wiley & Sons, New York, 1968, 5 (2): 150~205
[23]Cole KS, Electric Impedance of Suspensions of Spheres, J.Gen.physicol., 1928, 12 (1): 29~36
[24]Cole KS, Cole RH, Dispersion and Absorption in Dielectrics, J.Chem.phys., 1941, 9 (4): 341~351
[25]Sverre Grimnes, Orjan Grottem Martinsen, Bio-impedance and Bioelectricity Basics, Great Britain: ACADEMIC PRESS, 2000
[26]Schwan HP, Determination of Biological Impedance, Physical techniques in biomedical research, 1963, 6 (4): 323~406
[27]章晓丽,生物阻抗动态信息提取与处理关键技术研究:[硕士学位论文],天津:天津大学,2007
[28]王春艳,应用于肺部检测的电阻成像系统:[硕士学位论文],天津:天津大学,2007
[29]Y. Zou.and Z. Guo, A review of electrical impedance techniques for breast cancer detection, Medical engineering & physics, 2003, 25: 79~90
[30]Blad B, Baldetop B, Impedance spectra of tumor tissue in comparison with normal tissue: a possible clinical application for electrical impedance tomography, Physiol Meas, 1996, 17: A105~A115
[31]孙高,周宝森,杨静宇等,基于人工神经网络的肺癌诊断研究,中国卫生统计,1999,16(3):142~144
[32]洪文学,李佳书,严衍玲,人工神经网络在肿瘤检测和诊断中的应用研究进展,北京生物医学工程,2007,26(4):446~448
[33]黎衍云,李锐,张胜年,人工神经网络及其在疾病筛查中应用前景,环境与职业医学习,2006,23(1):71~73
[34]华贻军,人工神经网络在临床医学中的应用,实用医学杂,2005,21(3):330~332
[35]陈新平,临床医学中的神经网络技术,中国现代医学杂志,2003,13(9):46~50
[36]吴拥军,吴逸明,屈凌波等,人工神经网络技术在肺癌诊断中的应用研究,中华微生物学和免疫学杂志,2003,23(8):646~649
[37]李晓薇,刘文凯,李军会,人工神经网络在恶性肿瘤诊断中的应用,中国农业大学学报,2001,6(3):6~10
[38]刘琼荪,何离庆,基于人工神经网络的乳腺癌诊断模型,重庆大学学报,2003,26(4):70~72
[39]吴鹭萍,人工神经网络在医学上的应用,医学信息:医学与计算机应用,2000,13(7):368~370
[40]周开利,康耀红,神经网络模型及其MATLAB仿真程序设计,北京:清华大学出版社,2005
[41]李丽霞,王彤,范逢曦,BP神经网络设计探讨,现代预防医学,2005,32(2):128~130
[42]焦李成,神经网络系统理论,西安:西安电子科技大学出版社,1990
[43]韩力群,人工神经网络教程,北京:北京邮电大学出版社,2006
[44]闻新等,Matlab神经网络仿真与应用,北京:科学出版社,2003
[45]高大文,王鹏,蔡臻超,人工神经网络中隐含层节点与训练次数的优化,哈尔滨工业大学学报,2003,35(2):207~209
[46]邓丽,吴春梅,罗芳琼,神经网络集成及研究进展,柳州师专学报,2007,22(4):4741~4746