摘要
目的:应用表面增强激光解析电离飞行时间质谱(surface enhanced laser desorption/ionization time of flight mass spectrometry,SELDI-TOF-MS)技术筛选出肺癌患者血清、肺泡灌洗液(bronchoalveolar lavage fluid,BALF)和肺癌组织中的差异蛋白,了解其不同病理分型和解剖分型表达的特点及影响因素,并探讨其临床意义。
方法:选用弱阳离子交换蛋白芯片(WCX-2芯片),应用SELDI-TOF-MS技术分别检测40例肺癌患者和20例肺良性病变患者血清、肺泡灌洗液(BALF)和肺组织匀浆的蛋白质质图谱,筛选出肺癌的差异蛋白,用直线相关分析研究他们之间的内在联系。进一步研究分析其中肺鳞癌患者和肺腺癌患者血清、BALF和肺组织匀浆的蛋白质质图谱,筛选出差异蛋白;比较中央型肺癌患者和周围型肺癌患者血清、BALF和肺组织匀浆的蛋白质质图谱;同时研究20例周围型肺癌患者活检前后BALF的蛋白质质图谱变化情况。用Protein Chip Biomarker System II-C型蛋白质芯片阅读仪读取芯片信息,采用Ciphergen Protein Chip 3.2版本的分析软件自动采集数据,读取峰高值,获得蛋白表达图谱,用Biomarker Pattern 5.0软件分析差异蛋白并初步建立诊断模型。
结果:一、肺癌组与肺良性病变组的差异蛋白
1.血清差异蛋白两者血清中共存在9个有统计学差异的蛋白峰(P<0.05),肺癌组血清中的差异蛋白均为上调表达,以质荷比(M/Z)为4097.08Da的差异蛋白建立分类树诊断模型,其诊断肺癌的灵敏度为77.5%(31/40),特异度为70%(14/20),正确率为75%(45/60),阳性预测值为83.78(%31/37),阴性预测值为60.87(%14/23),ROC曲线下面积为0.767。
2.BALF中的差异蛋白两者共存在20个有统计学差异的蛋白峰(P<0.05),肺癌组BALF中的差异蛋白14个上调表达, 6个下调表达,以M/Z为8133.51Da和15117.01Da的差异蛋白联合建立分类树诊断模型最好,其诊断肺癌的灵敏度为80%(32/40),特异度为80%(16/20),正确率为80%(48/60),阳性预测值为88.89%(32/36),阴性预测值为66.67%(16/24),ROC曲线下面积为0.919。
3.肺组织匀浆的差异蛋白两者共存在18个有统计学差异的蛋白峰(P<0.05),肺癌组肺组织匀浆中的差异蛋白4个上调表达,14个下调表达,以M/Z为3648.28Da的差异蛋白建立分类树诊断模型最好,其诊断肺癌的灵敏度为72.5%(29/40),特异度为85%(17/20),正确率为76.67%(46/60),阳性预测值为90.63%(29/32),阴性预测值为60.71%(17/28),ROC曲线下面积为0.867。
4.直线相关分析结果(1)肺癌患者血清中M/Z为4097.08Da的建模差异蛋白与肺组织匀浆中M/Z为3648.28Da的建模差异蛋白之间无相关性(P>0.05),但与BALF中M/Z为8133.51Da的建模差异蛋白之间呈正相关(P=0.025、r =0.355),与BALF中M/Z为15117.01Da的建模差异蛋白亦呈正相关(P=0.021、r =0.363);(2)肺癌患者BALF中M/Z为8133.51Da与M/Z为15117.01Da的建模差异蛋白之间呈正相关(P<0.0001、r =0.618);(3)肺癌患者BALF中M/Z为8133.51Da和M/Z为15117.01Da的建模差异蛋白与肺组织匀浆中M/Z为3648.28Da的建模差异蛋白之间均无相关性(P>0.05)。
二、不同病理类型肺癌的差异蛋白
(一)肺鳞癌组与肺良性病变组1.血清:两者共存在2个有统计学差异的蛋白峰(P<0.05),肺鳞癌患者的血清中差异蛋白1个上调表达、1个下调表达,以M/Z为5124.24Da的差异蛋白建立分类树诊断模型,其诊断肺鳞癌的灵敏度为85%(17/20),特异度为65%(13/20),正确率为75%(30/40),阳性预测值为62.96%(17/24),阴性预测值为81.25%(13/16),其ROC曲线下面积为0.750。2.BALF:两者共存在9个有统计学差异的蛋白峰(P<0.05),肺鳞癌患者的BALF中差异蛋白4个上调表达、5个下调表达,以M/Z为7967.29Da和10843.45Da的差异蛋白联合建模最好,其诊断肺鳞癌的灵敏度为80%(16/20),特异度为80%(16/20),正确率为80%(32/40),阳性预测值为80%(16/20),阴性预测值为80%(16/20),ROC曲线下面积为0.916。3.肺组织匀浆:两者共存在8个有统计学差异的蛋白峰(P<0.05),肺鳞癌患者的肺组织匀浆中差异蛋白2个上调表达、6个下调表达,以M/Z为7914.59Da和8709.66Da的差异蛋白联合建模最好,其诊断肺鳞癌的灵敏度为75%(15/20),特异度为90%(18/20),正确率为82.5%(33/40),阳性预测值为88.24%(15/17),阴性预测值为78.26%(18/23),ROC曲线下面积为0.930。
(二)肺腺癌组与肺良性病变组1.血清:两者共存在8个有统计学差异的蛋白峰(P<0.05),肺腺癌患者的血清中均为上调表达,以M/Z为9295.79Da的差异蛋白建立分类树诊断模型较好,其诊断肺腺癌的灵敏度为75%(15/20),特异度为65%(13/20),正确率为70%(28/40),阳性预测值为68.18%(15/22),阴性预测值为72.22%(13/18),其ROC曲线下面积为0.844。2.BALF:两者共存在9个有统计学差异的蛋白峰(P<0.05),肺腺癌患者的BALF中均为上调表达,以M/Z为7923.01Da的差异蛋白建立分类树诊断模型较好,其诊断肺腺癌的灵敏度为80%(16/20),特异度为90%(18/20),正确率为85%(34/40),阳性预测值为88.89%(16/18),阴性预测值为81.82%(18/22),其ROC曲线下面积为0.933。3.肺组织匀浆:两者共存在7个有统计学差异的蛋白峰(P<0.05),肺腺癌患者的肺组织匀浆中的差异蛋白1个为上调表达、6个为下调表达,选择M/Z为2452.49Da的差异蛋白建立分类树诊断模型,其诊断肺腺癌的灵敏度为70%(14/20),特异度为85%(17/20),正确率为77.5%(31/40),阳性预测值为82.35%(14/17),阴性预测值为70.83%(17/24),其ROC曲线下面积为0.825。
三、中央型肺癌组和周围型肺癌组的差异蛋白
(一)血清差异蛋白
1.中央型肺癌组与肺良性病变组血清的差异蛋白两者共存在7个有统计学差异的蛋白峰(P<0.05),中央型肺癌患者的血清中差异蛋白5个上调表达、2个下调表达,以M/Z为6198.95Da的差异蛋白建立分类树诊断模型诊断效率最好,其诊断中央型肺癌的灵敏度为90%(18/20),特异度为60%(12/20),正确率为75%(30/40),阳性预测值为69.23%(18/26),阴性预测值为85.71%(12/14),ROC曲线下面积为0.800。
2.周围型肺癌组与肺良性病变组血清的差异蛋白两者共存在6个有统计学差异的蛋白峰(P<0.05),周围型肺癌患者的血清中的差异蛋白5个上调表达、1个下调表达,以M/Z为6636.54Da的差异蛋白建立分类树诊断模型诊断效率最好,其诊断周围型肺癌的灵敏度为70%(14/20),特异度为90%(18/20),正确率为80%(32/40),阳性预测值为87.5%(14/16),阴性预测值为75% (18/24),ROC曲线下面积为0.725。
(二)BALF中的差异蛋白
1.中央型肺癌组与肺良性病变组BALF的差异蛋白两者共存在12个有统计学差异的蛋白峰(P<0.05),中央型肺癌患者的BALF中的差异蛋白11个上调表达、1个下调表达,以M/Z为11308.2Da的差异蛋白建立分类树诊断模型诊断效率较好,其诊断中央型肺癌的灵敏度为70%(14/20),特异度为90%(18/20),正确率为80%(32/40),阳性预测值为87.5%(14/16),阴性预测值为75%(18/24),ROC曲线下面积为0.925。
2.周围型肺癌患者与肺良性病变患者BALF的差异蛋白两者共存在11个有统计学差异的蛋白峰(P<0.05),在周围型肺癌患者BALF中均为上调表达,以M/Z为7981.70Da的差异蛋白建分类树诊断模型最好,其诊断周围型肺癌的灵敏度为85%(17/20),特异度为90%(18/20),正确率为87.5%(35/40) ,阳性预测值为89.47%(17/19) ,阴性预测值为85.71%(18/21),ROC曲线下面积为0.943。
(三)、肺组织匀浆的差异蛋白
1.中央型肺癌组与肺良性病变组肺组织匀浆的差异蛋白两者共存在10个有统计学差异的蛋白峰(P<0.05),在中央型肺癌患者的肺组织匀浆中差异蛋白1个上调表达、9个下调表达,以M/Z为3504.44Da的差异蛋白建立分类树诊断模型诊断效率最好,其诊断中央型肺癌的灵敏度为70%(14/20),特异度为90%(18/20),正确率为80%(32/40),阳性预测值为87.5%(14/16),阴性预测值为75%(18/24),ROC曲线下面积为0.929。
2.周围型肺癌组与肺良性病变组肺组织匀浆的差异蛋白两者共存在11个有统计学差异的蛋白峰(P<0.05),在周围型肺癌患者的肺组织匀浆中均为下调表达,以M/Z为5394.38Da的差异蛋白建立分类树诊断模型诊断效率最好,其诊断周围型肺癌的灵敏度为75%(15/20),特异度为90%(18/20),正确率为82.5%(33/40),阳性预测值为88.24%(15/17),阴性预测值为78.26%(18/23),ROC曲线下面积为0.908。
四、周围型肺癌活检前后BALF的差异蛋白
1.周围型肺癌患者活检前BALF与肺良性病变组患者BALF的差异蛋白(见第三部分:“周围型肺癌患者与肺良性病变组患者BALF的差异蛋白”)。
2.周围型肺癌患者活检后BALF与肺良性病变组患者BALF的差异蛋白两者共存在14个有统计学差异的蛋白峰(P<0.05),在周围型肺癌患者BALF中13个为上调表达, 1个为下调表达,选择M/Z为7670.82Da、8046.49Da、15127.61Da和16067.91Da的差异蛋白峰建立分类树诊断模型灵敏度和特异度较好,其中以M/Z为7670.82Da的差异蛋白峰建模最好,其诊断周围型肺癌灵敏度为85%(17/20),特异度为100%(20/20),正确率为92.5%(37/40),阳性预测值为100%(17/17),阴性预测值为86.96%(20/23),ROC曲线下面积为0.925。
结论:1.SELDI-TOF-MS技术能筛选出肺癌患者的血清、BALF和肺组织中有统计学差异的蛋白峰,其中血清中M/Z为4097.08Da、BALF中M/Z为8133.51Da和M/Z为15117.01Da及肺组织匀浆中M/Z为3648.28Da的差异蛋白分别建立的分类树诊断模型诊断肺癌的灵敏度、特异度及正确率均较为理想,其中BALF中的差异蛋白比血清中的检测效率更高,可作为早期诊断肺癌的更敏感且特异的标志物。血清中M/Z为4097.08Da的差异蛋白与BALF中M/Z为8133.51Da和M/Z为15117.01Da的差异蛋白均呈正相关,提示肺癌患者BALF中的差异蛋白与血清的来源或形成机制相同。
2.肺鳞癌和肺腺癌BALF及肺组织匀浆中的差异蛋白较血清多,诊断效率比血清高,肺鳞癌BALF中M/Z为7967.29Da和10843.45Da的差异蛋白峰联合、肺组织匀浆中M/Z为7914.59Da和8709.66Da的差异蛋白峰联合建模及肺腺癌BALF中M/Z为7923.01Da的差异蛋白峰建立分类树诊断模型,其诊断的灵敏度、特异度和正确率均达到75%～90%,有可能作为诊断肺鳞癌和肺腺癌的标志蛋白。
3.中央型肺癌和周围型肺癌BALF和肺组织匀浆中的差异蛋白较血清多,诊断效率比血清高,中央型肺癌BALF中M/Z为11308.2Da、肺组织匀浆中M/Z为3504.44Da的差异蛋白和周围型肺癌BALF中M/Z为7981.70Da、肺组织匀浆中M/Z为5394.38Da的差异蛋白建立分类树诊断模型诊断的灵敏度、特异度及正确率为70%～90%,均比血清高,可以作为较好的诊断中央型肺癌和周围型肺癌的标志蛋白。
4.周围型肺癌活检前后BALF中存在不同的差异蛋白,活检后BALF中出现的差异蛋白更多,诊断周围型肺癌的临床意义更大。活检前M/Z为7981.70Da和活检后M/Z为7670.82Da的差异蛋白分别建立的分类树诊断模型诊断周围型肺癌的灵敏度、特异度和正确率均较为理想,可能成为更好的早期诊断周围型肺癌的肿瘤标志物。
Objectives To scan the protein mass spectra in the sera, bronchoalveolar lavage fluid(BALF)and lung homogenate of patients with lung cancer by surface enhanced laser desorption/ionization time of flight mass spectrometry (SELDI-TOF-MS) ,screen out their differential proteins, and also study their characteristics of different pathological and anatomical classification and the influence factors of protein expression, and finally explore the clinical significance of differential proteins in patients with lung cancer.
Methods SELDI-TOF-MS and protein chips weak cation exchange (WCX-2 chip) were used to detect the protein mass spectra and screen out the differential proteins in the sera, BALF and lung homogenate of 40 patients with lung cancer and 20 patients wtith benign pulmonary diseases. The correlations of the differential proteins were analyzed by linear correlation analysis. The protein mass spectra and the differential proteins in the sera, BALF and lung homogenate of patients with squamous cell carcinoma or adenocarcinoma were further studied. The protein mass spectra in the sera, BALF and lung homogenate of patients with central lung cancer were compared with those with peripheral lung cancer. The changes of the protein mass spectra of BALF, which collected before and after lung biopsy in 20 patients with peripheral lung cancer, were studied at the same time. The information of protein chips were read by Protein Chip Biomarker System II-C. The data and the peaks’height of protein mass spectra were automatically collected and read by Ciphergen Protein Chip 3.2 Software to acquire the protein expression mass spectra. The differential proteins were analyzed and the initial diagnostic models were set up by Biomarker Pattern 5.0 Software.
Results
Part 1 Differential Proteins in Patients with Lung Cancer and Patients with Benign Pulmonary Diseases
The protein differences in the patients with lung cancer compared with benign diseases were followed as: (1) Sera:There were significantly different expression of nine protein peaks in the sera of the two groups .All of them were significantly up-regulated in the sera of the lung cancer group(P<0.05).The differential protein whose M/Z was 4097.08Da was selected to establish classification tree of diagnostic model. The sensitivity of diagnosing lung cancer was 77.5%(31/40),the specificity was 70%(14/20),the accuracy was 75%(45/60),the positive predictive value (PV+) was 83.78%(31/37),the negative predictive value(PV-) was 60.87%(14/23)and the area under the ROC curve was 0.767. (2)BALF: There were significantly different expression of twenty protein peaks in the BALF of the two groups (P<0.05). Fourteen of them were significantly up-regulated in the BALF of the lung cancer group.Six of them were significantly down-regulated. The differential protein peaks whose M/Z were 8133.51Da and 15117.01Da were selected to establish classification tree diagnosis model together. The sensitivity of diagnosing lung cancer was 80%(32/40),the specificity was 80%(16/20), the accuracy was 80%(48/60), the PV+ was 88.89%(32/36), the PV- was 66.67%(16/24) and the area under the ROC curve was 0.919.(3) lung homogenate: There were significantly different expression of eighteen protein peaks in the lung homogenate of the two groups (P<0.05). Four of them were significantly up-regulated in the lung homogenate of the lung cancer group. Fourteen of them were significantly down-regulated.The differential protein whose M/Z was 3648.28Da was selected to establish the classification tree of diagnostic model. The sensitivity of diagnosing lung cancer was 72.5%(29/40),the specificity was 85%(17/20), the accuracy was 76.67%(46/60), the PV+ was 90.63%(29/32), the PV- was 60.71%(17/28) and the area under the ROC curve was 0.867.
The results of the linear correlation analysis were followed as :(1). There were no correlation between the protein in sera(M/Z 4097.08Da) and in lung homogenate(M/Z 3648.28Da) of patients with lung cancer(P>0.05), but there were positive correlation between the protein (M/Z 4097.08Da) in sera and (M/Z 8133.51Da) in BALF(P=0.025, r =0.355), also in sera (M/Z 4097.08Da)and BALF (M/Z 15117.01Da) of patients with lung cancer (P=0.021, r =0.363).(2) There were positive correlation between the two proteins whose M/Z were 8133.51Da and 15117.01Da in BALF of patients with lung cancer(P<0.0001,r =0.618). (3)There were no correlation between the proteins (M/Z 8133.51Da and 15117.01Da) in BALF and the protein (M/Z 3648.28Da) in lung homogenate of patients with lung cancer(P>0.05).
Part 2 Differential Proteins of Different Pathological Classification of Lung cancer
1. The squamous cell carcinoma group and the benign pulmonary diseases group :(1). Sera: There were significantly different expression of two protein peaks in the sera of the two groups (P<0.05). One was significantly up-regulated in sera of the squamous cell carcinoma group .Another was significantly down-regulated. The protein whose M/Z was 5124.24Da was selected to establish classification tree of diagnostic model. The sensitivity of diagnosing squamous cell carcinoma was 85%(17/20),the specificity was 65%(13/20), the accuracy was 75%(30/40), the PV+ was 62.96%(17/24), the PV- was 81.25%(13/16) and the area under the ROC curve was 0.750.(2).BALF: There were significantly different expression of nine protein peaks in the BALF of the two groups (P<0.05). Four of them were significantly up-regulated in BALF of the squamous cell carcinoma group. Five of them were significantly down-regulated. The differential protein whose M/Z were 7967.29Da and 10843.45Da were selected to establish classification tree of diagnostic model together. The sensitivity of diagnosing lung squamous cell carcinoma was 80%(16/20),the specificity was 80%(16/20), the accuracy was 80%(32/40), the PV+ was 80%(16/20), the PV- was 80%(16/20) and the area under the ROC curve was 0.916. (3) lung homogenate: There were significantly different expression of eight protein peaks in the lung homogenate of the two groups (P<0.05). Two of them were significantly up-regulated in lung homogenate of the squamous cell carcinoma group. Six of them were significantly down-regulated. The differential proteins whose M/Z were 7914.59Da and 8709.66Da were selected to establish classification tree of diagnostic model. The sensitivity of diagnosing squamous cell carcinoma was 75%(15/20),the specificity was 90%(18/20), the accuracy was 82.5%(33/40), the PV+ was 88.24%(15/17), the PV- was 78.26%(18/23) and the area under the ROC curve was 0.930.
2. The adenocarcinoma group and the benign pulmonary diseases group :(1) sera: There were significantly different expression of eight protein peaks in the sera of the two groups (P<0.05). All of them were significantly up-regulated in sera of the adenocarcinoma group. The protein whose M/Z was 9295.79Da was selected to establish classification tree of diagnostic model. The sensitivity of diagnosing adenocarcinoma was 75%(15/20),the specificity was 65%(13/20), the accuracy was 70%(28/40), the PV+ was 68.18%(15/22), the PV- was 72.22%(13/18) and the area under the ROC curve was 0.844. (2)BALF:There were significantly different expression of nine protein peaks in the BALF of the two groups (P<0.05). All of them were significantly up-regulated in BALF of the adenocarcinoma group. The protein whose M/Z was 7923.01Da was selected to establish classification tree of diagnostic model. The sensitivity of diagnosing adenocarcinoma was 80%(16/20),the specificity was 90%(18/20), the accuracy was 85%(34/40), the PV+ was 88.89%(16/18), the PV- was 81.82%(18/22) and the area under the ROC curve was 0.933. (3) lung homogenate: There were significantly different expression of seven protein peaks in the lung homogenate of the two groups (P<0.05). One of them was significantly up-regulated in lung homogenate of the adenocarcinoma group. Six of them were significantly down-regulated. The protein whose M/Z was 2452.49Da was selected to establish classification tree of diagnostic model. The sensitivity of diagnosing adenocarcinoma was 70%(14/20),the specificity was 85%(17/20), the accuracy was 77.5%(31/40), the PV+ was 82.35%(14/17), the PV- was 70.83%(17/24) and the area under the ROC curve was 0.825. Part 3 Differential Proteins in Central Lung Cancer Group and Peripheral Lung Cancer Group
1 Differential proteins in sera
(1). Differential proteins in sera of the central lung cancer group and the benign pulmonary diseases group: There were seven protein peaks with significantly diffenent expression in the two groups (P<0.05). Five of them were significantly up-regulated in sera of the central lung cancer group. Two of them were significantly down-regulated. The differential protein whose M/Z was 6198.95Da was selected to establish classification tree of diagnostic model. The sensitivity of diagnosing the central lung cancer was 90%(18/20),the specificity was 60%(12/20), the accuracy was 75%(30/40), the PV+ was 69.23%(18/26), the PV- was 85.71%(12/14) and the area under the ROC curve was 0.800.
(2) Differential proteins of sera in patients with peripheral lung cancer group or benign pulmonary diseases group: There were significantly different expression of six protein peaks in the sera of the two groups (P<0.05). Five of them were significantly up-regulated in sera of the peripheral lung cancer group. One of them was down-regulated. The protein whose M/Z was 6636.54Da was selected to establish classification tree of diagnostic model. The sensitivity of diagnosing peripheral lung cancer was 70%(14/20),the specificity was 90%(18/20), the accuracy was 80%(32/40), the PV+ was 87.5%(14/16), the PV- was 75% (18/24) and the area under the ROC curve was 0.725.
2 Differential proteins in BALF
(1) Differential proteins in BALF of central lung cancer group and the benign pulmonary diseases group: There were significantly different expression of twelve protein peaks in the BALF of the two groups (P<0.05). Eleven of them were significantly up-regulated in BALF of the central lung cancer group. One of them was significantly down-regulated. The differential protein whose M/Z was 11308.2Da was selected to establish classification tree of diagnostic model. The sensitivity of diagnosing central lung cancer was 70%(14/20),the specificity was 90%(18/20), the accuracy was 80%(32/40), the PV+ was 87.5%(14/16), the PV- was 75%(18/24) and the area under the ROC curve was 0.925.
(2) Differential proteins in BALF of peripheral lung cancer group and benign pulmonary diseases group: There were significantly different expression of eleven protein peaks in the BALF of the two groups (P<0.05). All of them were significantly up-regulated in BALF of the peripheral lung cancer group. The protein whose M/Z was 7981.70Da was the best one of the proteins selected to establish classification tree of diagnostic model. The sensitivity of diagnosing peripheral lung cancer was 85%(17/20),the specificity was 90%(18/20), the accuracy was 87.5% (35/40), the PV+ was 89.47%(17/19), the PV- was 85.71%(18/21) and the area under the ROC curve was 0.943.
3 Differential proteins in lung homogenate
(1) Differential proteins in lung homogenate of the central lung cancer group and the benign pulmonary diseases group: There were significantly different expression of ten protein peaks in the lung homogenate of the two groups (P<0.05). One of them was significantly up-regulated in lung homogenate of the central lung cancer group. Nine of them were significantly down-regulated. The differential protein whose M/Z was 3504.44Da was selected to establish classification tree of diagnostic model. The sensitivity of diagnosing the central lung cancer was 70%(14/20),the specificity was 90%(18/20), the accuracy was 80%(32/40), the PV+ was 87.5%(14/16), the PV- was 75%(18/24) and the area under the ROC curve was 0.929.
(2) Differential proteins in lung homogenate of the peripheral lung cancer group and the benign pulmonary diseases group: There were significantly different expression of eleven protein peaks in the lung homogenate of the two groups (P<0.05). All of them were significantly up-regulated in lung homogenate of the peripheral lung cancer group. The protein whose M/Z was 5394.38Da was the best one of the proteins selected to establish classification tree of diagnostic model. The sensitivity of diagnosing peripheral lung cancer was 75%(15/20),the specificity was 90%(18/20), the accuracy was 82.5%(33/40), the PV+ was 88.24%(15/17), the PV- was 78.26%(18/23) and the area under the ROC curve was 0.908.
Part 4 Differential Proteins in BALF of Patients with Peripheral Lung Cancer before and after Lung Biopsy
1. Differential proteins in the BALF before lung biopsy of the peripheral lung cancer group and the benign pulmonary diseases group: There were the same results as the differential proteins of BALF in patients with peripheral lung cancer group and the benign pulmonary diseases group,which were described in Part 3.
2. Differential proteins of BALF in patients with peripheral lung cancer or the benign pulmonary diseases after lung biopsy: There were significantly different expression of fourteen protein peaks in the BALF of the two groups (P<0.05). Thirteen of them in BALF were significantly up-regulated after lung biopsy for patients with peripheral lung cancer. One of them was significantly down-regulated. The proteins whose M/Z were 7670.82Da、8046.49Da、15127.61Da and 16067.91Da were selected to establish classification tree of diagnostic model and 7670.82Da was the best one of them. The sensitivity of diagnosing peripheral lung cancer was 85%(17/20) , the specificity was 100%(20/20), the accuracy was 92.5%(37/40), the PV+ was 100%(17/17), the PV- was 86.96%(20/23) and the area under the ROC curve was 0.925.
Conclusions
1. SELDI-TOF-MS technology can detect the protein mass spectra and screen out the differential proteins in the sera, BALF and lung homogenate in patients with lung cancer. The sensitivity, specificity and accuracy in proteins with the M/Z of 4097.08Da in sera, with the M/Z of 8133.51Da and 15117.01Da in BALF and with the M/Z of 3648.28Da in lung homogenate are quite ideal. The differential proteins in BALF are more efficient to diagnose lung cancer than those in sera and they can be the sensitive and specific biomarkers to diagnose lung cancer at the early stage. There are positive correlations between the protein (M/Z 4097.08Da) in sera and the proteins (M/Z 8133.51Da and M/Z 15117.01Da) in BALF of patients with lung cancer, which means the proteins have the same origin or formation mechanism.
2. There are more differential proteins in BALF and lung homogenate in patients with squamous cell carcinoma or adenocarcinoma and they show higher efficiency to diagnose squamous cell carcinoma or adenocarcinoma compared with the proteins in sera when they were selected to set up the classification tree of diagnostic model. The sensitivities, specificities and accuracies of the diagnosis models set up by differential proteins in BALF (M/Z 7967.29Da and 10843.45Da) and in lung homogenate (M/Z 7914.59Da and 8709.66Da) of patients with squamous cell cacinoma and the protein (M/Z 7923.01Da) in BALF of patients with adenocarcinoma were about 75% to 90%.These differential proteins are possible to be the biomarkers to diagnose the squamous cell carcinoma and the adenocarcinoma.
3. There are more differential proteins in BALF and in lung homogenate of patients with central lung cancer or peripheral lung cancer, and they show higher efficiency to diagnose central lung cancer or peripheral lung cancer compared with the proteins in sera when they were selected to set up the classification tree of diagnostic model. The sensitivities, specificities and accuracies of the diagnostic models set up by the proteins in BALF (M/Z 11308.2Da) and in lung homogenate (M/Z 3504.44Da) of patients with central lung cancer and the proteins in BALF (M/Z 7981.70Da) and in lung homogenate (M/Z 5394.38Da) of patients with peripheral lung cancer are about 70% to 90% and show higher efficiency to diagnose lung cancer compared with the proteins in sera . These differential proteins are possible to be better biomarkers to diagnose central lung cancer and peripheral lung cancer.
4. There are many differential proteins in BALF before and after lung biopsy for pateints with peripheral lung cancer. There are more differential proteins in the BALF after lung biopsy and the proteins have more clinical significance to diagnose the peripheral lung cancer.The sensitivities, specificities and accuracies of the diagnosis models set up by the proteins in BALF before (M/Z 7981.70Da) and after (M/Z 7670.82Da) lung biopsy are quite ideal. These differential proteins are possible to be better tumor markers to diagnose peripheral lung cancer at the early stage.
引文
[1] Toll BA,Salovey PO,Malley SS,et al.Message framing for smoking cessation:The interaction of risk perceptions and gender[J].Nicotine Tob Res,2008,10(1):195-200.
[2] Jemal A, Thomas A, Murray T, et al. Cancer statistics, 2003[J].CA Cancer. J Clin 2003,53(1):5–26.
[3] Peter B, James R, Ugo Pastorino, et al. Computed Tomography Screening and Lung Cancer Outcomes [J].JAMA, 2007,297:953-961.
[4] Trkanjec JT, Peros-Golubici? T, Grozdek D,et al. The role of transbronchial lung biopsy in the diagnosis of solitary pulmonary nodule [J]. Coll Antropol, 2003, 27(2):669-675.
[5]曹锐彬,何丽华.经电子支气管镜肺活检对周围型肺癌的诊断价值[J].疑难病杂志,2005,4(4):210-212.
[6]汪斌超,涂海燕,陈正贤,等.纤维支气管镜检查技术在周围型肺癌病理诊断中的应用评价[J].承德医学院学报,2004,2l(2):114-116.
[7]王兴胜,崔社怀,林海龙,等.经纤维支气管镜肺活检及刷检在肺部周围性占位性病变诊断中的价值[J].第三军医大学学报,2005,27(15):1579-1580.
[8] Kang LQ, Song ZW, Li ZX,et al. Preliminary study on CT perfusion imaging in guiding biopsy of pulmonary lumps[J]. Chin Med J (Engl), 2009, 122(7):807-812.
[9] Chlap Z, Jedynak U, Sladek K. Mast cell: it's significance in bronchoalveolar lavage fluid cytologic diagnosis of bronchial asthma and interstitial lung disease [J] .Pneumonol Alergol Pol,1998,66 (6):321-329.
[10] Jacobs JA, de Brauwer E.BAL fluid cytology in the assessment of infectious lung disease [J].Hosp Med, 1999, 60(8):550-555.
[11] Mikuz G, Gschwendtner A. Value of bronchoalvoeolar lavage in the diagnosisof lung disease [J] .Verh Dtsch Ges Pathol, 2000, 84:129-35.
[12] Griese M.Pulmonary surfactant in health and human lung diseases: state of the art[J]. Eur Respir J, 1999, 13(6):1455-1476.
[13] Hutchens TW, Yip Tai-Yung. New desorption strategies for the mass spectrometric analysis of macromolecules [J]. RAPID COMMUN MASS SP, 1993, 7 (1):576-580.
[14] Vorderwülbecke S, Cleverley S, Weinberger SR, et al. Protein quantification by the SELDI-TOF-MS-based Protein Chip System [J].Nat Methods, 2005, 2(5):393.
[15] Seibert V, Wiesner A, Buschmann T, et al. Surface-enhanced laser desorption ionization time-of-flight mass spectrometry (SELDI TOF-MS) and Protein Chip technology in proteomics research[J]. Pathol Res Pract JT-Pathology, research and practice, 2004, 200(2):83-94.
[16] Petricoin EF,Liotta LA.Clinical applications of proteomics[J].J Nutr,2003,133(7Suppl):2476s-2484s.
[17] Figeys D.Proteomics in 2002:A year of technical development and wide-rangingapplications[J].Analytical Chem 2003,75(12):2891-2905.
[18] Petricoin EF,Ardekani AM,Hitt BA,et al.Use of proteomic patterns in sera to identify ovarian cancer[J].Lancet 2002,359(9306):572-577.
[19] Adam BL, Qu Y, Davis JW, et al. Serum protein fingerprinting coupled with a pattern-matching algorithm distinguishes prostate cancer from benign prostate hyperplasia and healthy men[J].Cancer Res 2002,62(13):3609-3614.
[20]操东杰.肺癌筛查和早期诊断进展[J].临床肺科杂志,2007,12(6):531.
[21]陶华林.肿瘤标志物的研究及临床应用进展[J].国外临床生物化学与检验学分册.2005,26(1):54-56.
[22]杨德昌,杨拴盈.肺癌诊断水平的现状与进展[J].中华结核和呼吸杂志,2001,24(8):450.
[23] Swensen SJ, Jett JR, Hartman TE, et al. Lung Cancer Screening with CT: Mayo Clinic Experience [J].Radiology, 2003, 226(3):756-761.
[24] Kulpa J,Wojcik E,Reinfuss M,et a1.Carcinoembryonic antigen,squamous cell carcinoma antigen,CYFRA21-1,and neuron-specific enolase in squamous cell lung cancer patients[J].Clin Chem,2002,48(11):1931.
[25] Zhong L,Peng X,Hidalgo GE,et a1.Identification of circulating antibodies to tumor-associated proteins for combined use as markets of non-small cell lung cancer[J].Proteomies,2004,4(4):1216.
[26] Weinbergr SR, Dalmasso EA, Fung ET. Current achievements using protein chip array technology [J]. Curt Opin Chem Biol,2005,6:86-91.
[27] Zhukov TA, Johanson RA, Cantor AB, et al. Discovery of differential protein profiles specific for lung tumors and pre-malignant lung lesions by SELDI mass spectrometry [J].Lung Cancer 2003, 40(3):267-279.
[28]赵健,黄维华,王远东,等.应用SELDI-TOF-MS技术筛选肺腺癌患者血清诊断标志物[J].肿瘤基础与临床,2008,21(1):7-9.
[29]李香兰,尤庆山,杨艳梅,等.表面增强激光解吸电离飞行时间质谱技术筛查肺癌血清特异度蛋白质及其临床意义[J].现代生物学进展,2007,7(2): 234-236.
[30]何伟伟,陈海泉,罗晓阳,等.表面增强激光解析离子化飞行时间质谱技术筛选的肺腺癌血清标志蛋白与传统肿瘤标志物的比较[J].上海医学,2008,31(7):483-486.
[31] William Jacot, Ludovic Lhermitte, Nadege Dossat, et al. Serum Proteomic Profiling of Lung Cancer in High-Risk Groups and Determination of Clinical Outcomes [J].Journal of Thoracic Oncology ,2008,3(8):840-850.
[32] No?l-Georis I, Bernard A, Falmagne P, Wattiez R. Database of bronchoalveolar lavage fluid proteins[J].J Chromatogr B Analyt Technol Biomed Life Sci.2002,771(1-2):221-236.
[33] Su L,Zhang J,Wei XQ,et al.The correlation between the differ-ential expressionion of Slp2 gene (stomatin like protein-2) and the differentiation and metastasis of the lung cancer[J]. Chin J Clin Oncol, 2005, 32(12):678-680.
[34] Su L, Zhang J, Wei XQ, et al. Construction of antisense Slp2 gene and its effects on growth and proliferation of lung cancer cell lines [J]. Chin J Lung Cancer, April 2007, 10(2):120-123.
[35]孙健.支气管肺泡灌洗液的蛋白质组研究[J].国外医学卫生学分册,2004,3l(l):51-55.
[36] Chi Shing ,Cho W. Potentially useful biomarkers for the diagnosis, treatment and p rognosis of lung cancer[J]. Biomed Pharmacother,2007, 61(9):515.
[37]戴明,罗荣城.蛋白质组学在肺癌诊治研究中的应用[J].癌症,2007,26(6):669-672.
[38]田应选,杨拴盈,南岩东.激光捕获显微切割联合SELDI蛋白质芯片筛选肺腺癌早期诊断标志蛋白[J].四川大学学报(医学版),2009,40(1):157-161.
[39]区兆基,曹志成,叶大同,等.蛋白质芯片筛选香港非吸烟女性肺腺癌组织的肿瘤标志物[J].癌症,2008,27(3):264-271.
[40]吕超,沈靖,吴楠,等.非小细胞肺癌淋巴转移血清蛋白指纹图谱分析[J].中华医学杂志,2009 ,89(35):2481-2485.
[41]简红,廖美琳,万欢英,等.女性肺癌流行病学、病因学和分子生物学研究进展[J].中国肿瘤杂志,2007,16(5): 311-313.
[42] Chen, Cole P. Air pollution and adenocarcinoma of the lung [J].Lung Cancer, 2005, 49 (Suppl 2)∶S189 (P2285).
[43] Dalai R, Behera D, B. T. Assessment of domestic cooking as a risk factor for the development of lung cancer in non-smoking Indian women. Lung Cancer ,2005,49 (Suppl 2)∶S190(P2287).
[44] Gibellino F, Nicolosi M, Grasso G. Hormonal risk factors in women lung cancer [J]. Lung Cancer ,2005,49 ( Suppl 2)∶S384 (P222).
[45]周宝森,王天爵,张群弟,等.女性肺腺癌危险因素分析[J].中国公共卫生杂志,2000,16(6):536-539.
[46]高云,方健,刘叙仪,等.305例非小细胞肺癌骨转移的诊断、治疗和预后分析[J].中国肺癌杂志,2006,9(4),357-361.
[47]廖萍,王文静,朱佩云,等. SELDI-TOF-MS技术检测肺腺癌血清生物标志物[J].中国肿瘤,2006,15(4):258-261.
[48]李鑫,王志华,马玉彦,等. SELDI-TOF-MS技术分析高低转移肺腺癌细胞株差异表达蛋白[J].实用肿瘤学杂志,2OO6,20(3):183-186.
[49]洪璇,陈公琰,王萌,等.应用SELDI蛋白质芯片技术筛选肺腺癌血清标志物的研究[J].实用肿瘤学杂志,20O7,21(2):101-103.
[50]廖晓东,毛伟敏,郑智国.非小细胞肺癌患者血清蛋白指纹图谱的构建及其临床意义[J].实用癌症杂志,2009,24(5):457-461.
[51] Funai K,Yokose T, Ishii G, et al. Clinicopathologic characteristics of peripheral squamous cell carcinoma of the lung[J]. Am Surg Pathol, 2003, 27(6):978-984.
[52]刘洋,孙雨,周乃康,等.多层面螺旋CT图像重建判断中央型肺癌可切除性研究与病理、手术相关性分析[J].中国肿瘤临床,2006,33(12):706-708.
[53] Okada M,Nishio W,Sakamoto T,et a1.Correlation between computed tomographic findings,bronchioloalveolar carcinoma component,and biolgic behavior of small-sized lung adenocarcinomas[J].J Thorac Cardiovasc Surg,2004,127(3):857-861.
[54]冯若飞,邹文辉,李明生,等.细菌内毒素诱导人肺腺癌细胞凋亡的实验研究[J].西北民族大学学报(自然科学版),2008,29(1):44-48.
[55]王洪武.周围型肺癌的早期诊断策略[J].中国肺癌杂志,2008,11(4):577-580.
[56]唐良法,江莲,吴晓,等.支气管肺泡灌洗液癌胚抗原检测对周围型肺癌的诊断价值[J].临床肺科杂志,2009,14(7):924-925.
[57]胡风劲,袁家芳,袁道国,等.支气管肺泡灌洗液中TSGF和CA-199检测的临床分析[J].重庆医学,2009,38(11):1395-1396.
[58] Issaq HJ,Conrads TP,Prieto DA,et a1.SELDI-TOF-MS for diagnostic Proteomics[J].J Anal Chem,2003;75:148-15.
[1] Parkin DM, Bray F, Ferlay J, et al. Global cancer statistics, 2002 [J]. CA Cancer J Clin, 2005, 2(55): 74-108.
[2] Jemal A, Thomas A, Murray T, et al. Cancer statistics, 2003[J].CA Cancer. J Clin 2003,53(1):5–26.
[3] Swensen SJ, Jett JR, Hartman TE, et al. Lung Cancer Screening with CT: Mayo Clinic Experience [J]. Radiology, 2003, 226(3):756-761.
[4] Hutchens TW, Yip Tai-Yung. New desorption strategies for the mass spectrometric analysis of macromolecules [J]. RAPID COMMUN MASS SP, 1993, 7(1):576-580.
[5] Vorderwülbecke S, Cleverley S, Weinberger SR, et al. Protein quantification by the SELDI-TOF-MS-based Protein Chip System [J].Nat Methods, 2005, 2(5):393.
[6] Petricoin EF , Liotta LA . Clinical applications of proteomics[J] . J Nutr,2003,133(7Suppl):2476s-2484s.
[7] Figeys D.Proteomics in 2002:A year of technical development and wide-ranging applications[J].Chem 2003;75(12):2891-2905.
[8] Petricoin EF, Ardekani AM, Hitt BA, et al. Use of proteomic patterns in sera to identify ovarian cancer[J]. Lancet 2002,359(9306):572-577.
[9] Adam BL, Qu Y, Davis JW, et al. Serum protein fingerprinting coupled with a pattern-matching algorithm distinguishes prostate cancer from benign prostate hyperplasia and healthy men[J].Cancer Res 2002,62(13):3609-3614.
[10]廖萍,王文静,朱佩云,等. SELDI-TOF-MS技术检测肺腺癌血清生物标志物[J].中国肿瘤,2006,15(4):258-261.
[11]罗晓阳,陈海泉,王文静,等.表面增强激光解析离子化飞行时间质谱技术筛选肺腺癌血清肿瘤标记物[J].中国癌症杂志,2006,l6(1):14-16.
[12]李鑫,王志华,马玉彦,等. SELDI-TOF-MS技术分析高低转移肺腺癌细胞株差异表达蛋白[J].实用肿瘤学杂志,2006,20(3):183-186.
[13]洪璇,陈公琰,王萌,等.应用SELDI蛋白质芯片技术筛选肺腺癌血清标志物的研究[J].实用肿瘤学杂志,20O7,21(2):101-103.
[14]赵健,黄维华,王远东,等.应用SELDI-TOF-MS技术筛选肺腺癌患者血清诊断标志物[J].肿瘤基础与临床,2008,21(1):7-9.
[15]何伟伟,陈海泉,罗晓阳,等.表面增强激光解析离子化飞行时间质谱技术筛选的肺腺癌血清标志蛋白与传统标志物的比较[J].上海医学,2008,31(7):483-486.
[16]廖晓东,毛伟敏,郑智国.非小细胞肺癌患者血清蛋白指纹图谱的构建及其临床意义[J].实用癌症杂志,2009,24(5):457-461.
[17]区兆基,曹志成,叶大同,等.蛋白质芯片筛选香港非吸烟女性肺腺癌组织的肿瘤标志物[J].癌症,2008,27(3):264-271.
[18]田应选,杨拴盈,南岩东.激光捕获显微切割联合SELDI蛋白质芯片筛选肺腺癌早期诊断标志蛋白[J].四川大学学报(医学版),2009,40(1):157-161.
[19] Yang SY, Xiao XY, Zhang WG, et al. Application of sera SELDI proteomic patterns in diagnosis of lung cancer[J].BMC Cancer 2005,5(1):83.
[20] Luo S, Wang M, Li Y. Clinical significance in diagnosing lung cancer with the combined determination of sera tumor markers[J].J tumor Marker Oncology 2004,19(1):90.
[21]李香兰,尤庆山,杨艳梅,等.表面增强激光解吸电离飞行时间质谱技术筛查肺癌血清特异度蛋白质及其临床意义[J].现代生物学进展,2007,7(2):234-236.
[22]戴嵩玮,王小敏,刘丽云,等.一个在肺癌血清中高表达的标志分子SAA的发现及鉴定[J].中国科学C辑:生命科学,2007,37(2):129-134.
[23]吕超,沈靖,吴楠,等.非小细胞肺癌淋巴转移血清蛋白指纹图谱分析[J].中华医学杂志,2009 ,89(35):2481-2485.
[24] Zhukov TA, Johanson RA, Cantor AB, et al. Discovery of differential proteinprofiles specific for lung tumors and premalignant lung lesions by SELDI mass spectrometry [J] . Lung Cancer, 2003,40(3):267-279.
[25] Chen G, Tarek TG, Huang CC, et al. Proteomic analysis of lung adenocarcinoma: Identification of a highly expressioned set of proteins in tumors[J].Clin Cancer Res, 2002, 8 (7) : 2298-2305.
[26] Yanagisawa K, Shyr Y, Xu BJ, et al. Proteomic patterns of tumour subsets in non-small-cell lung cancer[J].Lancet2003; 362(9382):433–439.
[27] Au JS, Cho WC; Yip TT; Deep proteome profiling of sera from never-smoked lung cancer patients [J].Biomedecine & pharmaco-therapy, 2007,61(9):570-577.
[28] William Jacot, Ludovic Lhermitte, Nadege Dossat, et al. Serum Proteomic Profiling of Lung Cancer in High-Risk Groups and Determination of Clinical Outcomes [J].Journal of Thoracic Oncology ,2008,3(8):840-850.
[29] Fung ET, Thulasiraman V, Weinberger SR, et al. Protein biochips for differential profiling[J].Curr Opin Biotechnol.2001;12(1):65-69.
[30] Paweletz CP, Trock B, Pennanen M, et al. Proteomic patterns of nipple aspirate fluids obtained by SELDI-TOF: potential for new biomarkers to aid in the diagnosis of breast cancer[J].Dis Markers.2001;17(4):301-307.
[31]李倩,廖尚英,杨桂,等.SELDI-TOF-MS技术及其在癌症早期诊断中的应用[J].生物技术通报,2004,2:42-44.
[32]罗晓阳.蛋白质芯片技术在肿瘤标志物筛选中的研究进展[J].中国癌症杂志,2004,14(4):391-395.
[33] CHOWC (曹志成) ,YUEKM (余坚文) ,LEUNGWN (梁荣能). Proteomics-leading the post-genomeera [J]. Chinese Biotechnology (中国生物工程杂志),2005, 25(1):33-38.