Application of serum protein fingerprint in diagnosis of pancreatic cancer / 浙江大学学报·医学版

<p><b>OBJECTIVE</b>To establish serum protein fingerprint model for early diagnosis of pancreatic cancer with surface enhanced laser desorption/ionization time of flight-mass spectrometry (SELDI-TOF-MS) and bioinformatics techniques.</p><p><b>METHODS</b>A total of 73 samples were analyzed in this study, including 31 cases of pancreatic cancers, 22 cases of pancreatitis and 20 healthy individuals. Samples were first analyzed by SELDI-TOF-MS and two patterns of differentiation model were constructed with support vector machine arithmetic method.</p><p><b>RESULTS</b>The pattern 1 model differentiating pancreatic cancer patients from healthy individuals had a specificity and a sensitivity of both 100.0%. The pattern 2 model differentiating pancreatic cancer from pancreatitis had a specificity of 95.5% and a sensitivity of 93.5%.</p><p><b>CONCLUSION</b>SELDI-TOF-MS technique combined with bioinformatics can facilitate to identify biomarkers for pancreatic cancer.</p>

Apoptosis of non-tumor cells contributes to increased serum cytochrome c level in a neuroblastoma xenograft model / 中华医学杂志(英文版)

<p><b>BACKGROUND</b>Neuroblastoma (NB) is one of the most common malignant solid tumors of childhood. It is still not clear whether the apoptosis of tumor cells or the non-tumor cells contributes to the increase of concentration of cytochrome c (Cyt c) in the serum of the cancer patients. The aim of this research was to identify the source of the Cyt c in the serum when the tumor grows up by subcutaneous inoculation of human NB cells into nude mice.</p><p><b>METHODS</b>We subcutaneously inoculated human NB cells (KP-N-NS) into nude mice and collected the sera of tumor-bearing mice (n = 14) and control mice (n = 25) 4 weeks later in order to screen for and identify differentially expressed proteins in the serum. Differentially expressed proteins in the serum were screened by surface-enhanced laser desorption/ionization-time-of-flight (SELDI-TOF) mass spectrometry.</p><p><b>RESULTS</b>The relative intensity of a protein having a mass-to-charge ratio (m/z) of 11 609 was 3338.37 ± 3410.85 in the tumor group and 59.84 ± 40.74 in the control group, indicating that the expression level of this protein in the tumor group was 55.8 times higher than that in the control group. Serum proteins were separated and purified by high-performance liquid chromatography (HPLC). Liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) was performed to produce peptide mass fingerprints (PMFs). Spectrum analysis and a database search revealed that the highly expressed protein (m/z = 11 605.4) from the serum of tumor-bearing mice was the mouse Cyt c.</p><p><b>CONCLUSIONS</b>Increased concentration of Cyt c in the serum of tumor-bearing nude mice might be partially attributed to the secretion of this protein by non-tumor cells.</p>

Detection of nasopharyngeal carcinoma using surface-enhanced laser desorption and ionization mass spectrometry profiles of the serum proteome / 癌症

<p><b>BACKGROUND AND OBJECTIVE</b>Early diagnosis of nasopharyngeal carcinoma (NPC) is difficult due to the insufficient specificity of the conventional examination method. This study was to investigate potential and consistent biomarkers for NPC, particularly for early detection of NPC.</p><p><b>METHODS</b>A proteomic pattern was identified in a training set (134 NPC patients and 73 control individuals) using the surface-enhanced laser desorption and ionization-mass spectrometry (SELDI-MS), and used to screen the test set (44 NPC patients and 25 control individuals) to determine the screening accuracy. To confirm the accuracy, it was used to test another group of 52 NPC patients and 32 healthy individuals at 6 months later.</p><p><b>RESULTS</b>Eight proteomic biomarkers with top-scored peak mass/charge ratios (m/z) of 8605 Da, 5320 Da, 5355 Da, 5380 Da, 5336 Da, 2791 Da, 7154 Da, and 9366 Da were selected as the potential biomarkers of NPC with a sensitivity of 90.9% (40/44) and a specificity of 92.0% (23/25). The performance was better than the current diagnostic method by using the Epstein-Barr virus (EBV) capsid antigen IgA antibodies (VCA/IgA). Similar sensitivity (88.5%) and specificity (90.6%) were achieved in another group of 84 samples.</p><p><b>CONCLUSION</b>SELDI-MS profiling might be a potential tool to identify patients with NPC, particularly at early clinical stages.</p>

Detection and identification of specific serum biomarkers in papillary thyroid cancer / 中华肿瘤杂志

<p><b>OBJECTIVE</b>To detect and identify the potential specific serum biomarkers for diagnosis of papillary thyroid cancer.</p><p><b>METHODS</b>Samples of 35 patients with papillary thyroid carcinoma, 40 patients with benign thyroid nodule and 34 healthy individuals were analyzed using the SELDI-TOF ProteinChip System and bioinfomation technology to find the differential peaks which were separated by HPLC and then further analyzed by LC-MS/MS. The protein sequences were analyzed by SEQUEST software and searched in Bioworks database.</p><p><b>RESULTS</b>The top six mass-to-charge ratio (M/Z) peaks with the smallest P value were 6651, 6452, 7653, 7932, 15 106 and 15 848 Da, respectively. The 6651 and 6452 Da proteins were weakly expressed in papillary thyroid carcinoma but highly expressed in benign thyroid nodules and healthy individuals. The differences had statistical significance (P < 0.01). The 7653, 7932, 15 106, 15 848 Da proteins were highly expressed in papillary thyroid carcinoma but weakly expressed in benign thyroid nodules and healthy individuals. The differences were statistically significant (P < 0.01). Combination of these six proteins, using the method of leave-one-out to make crossing detection, the specificity of discriminating papillary thyroid carcinoma and non-cancer was 88.0%, and its sensitivity was 92.5%. The 6651 and 6452 Da proteins were identified as apolipoprotein C-I and apolipoprotein C-III, respectively. The 7653 and 15 106 Da proteins were identified as the same protein-alpha-globin, and the 7932 and 15,848 Da proteins were identified as the same protein-beta-globin.</p><p><b>CONCLUSION</b>The detection of differentially expressed apolipoprotein C-I, apolipoprotein C-III, alpha-globin, and beta-globin may have utility for diagnosis of papillary thyroid carcinoma and are worthy of further investigation.</p>

Application of serum protein markers to distinguish familial adenomatous polyposis (FAP) and sporadic colorectal adenomas / 中华肿瘤杂志

<p><b>OBJECTIVE</b>To screen out specifically-expressed serum protein markers in familial adenomatous polyposis (FAP) and to establish a serum protein fingerprint diagnostic model for distinguishing FAP from sporadic colorectal adenomas.</p><p><b>METHODS</b>Serum samples were collected from 19 FAP cases and 16 sporadic colorectal adenomas with informed consent. Serum protein fingerprint profiles were detected by SELDI-TOF-MS with CM 10 protein chip to screen out FAP adenoma-related serum protein markers, and support vector machine (SVG) technique was used to establish the diagnostic model to distinguish FAP from sporadic colorectal adenomas.</p><p><b>RESULTS</b>Six differently-expressed protein peaks (P < 0.01) were detected. Among them proteins of 5640, 3160, 4180 and 4290 m/z were highly expressed in FAP adenomas, and proteins of 3940 and 3400 m/z were highly expressed in sporadic colorectal adenomas. The accuracy of diagnostic model established with SVG to distinguish FAP adenomas and sporadic colorectal adenomas was 94.7% and 93.7%, respectively.</p><p><b>CONCLUSION</b>SELDI-TOF-MS can be effectively used to screen out the differentially expressed serum protein markers in FAP adenomas and sporadic colorectal adenomas, and a diagnostic model build by SVG to distinguish them has been successfully established. Therefore, a useful breakthrough point for research on molecular mechanisms of FAP pathogenesis is provided.</p>

Protein profiles of serum in gastric cancer at a high-risk area by SELDI-TOF mass spectrometry / 中华预防医学杂志

<p><b>OBJECTIVE</b>To explore the specific and sensitive biomarkers for gastric cancer detection, a surface-enhanced laser desorption and ionization protein chip mass spectrometry (SELDI-TOF-MS) was used to generate protein profiles of serum in gastric cancer at a high-risk area.</p><p><b>METHODS</b>A total of 36 gastric cancer cases and 46 subjects with superficial gastritis were selected from Linqu county, Shandong province, a high-risk area of gastric cancer. Serum samples were collected and Q10 protein chips were used to detect the serum proteomic patterns, and the sensitivity and specificity were assessed.</p><p><b>RESULTS</b>For the comparison of the gastric cancer group (26 out of 36 gastric cancer cases) versus superficial gastritis group (37 out of 46 subjects), the 6 most discriminating peaks (m/z 8587, 6945, 8243, 3899, 7035, and 9943) were identified by the ProteinChip Data Analysis System (ZUCIPDAS). The sensitivity and specificity of this pattern were 88.5% and 97.3%, respectively. A total of 19 subjects (10 gastric cancer cases and 9 superficial gastritis subjects) was selected to test the accuracy of this pattern by using blind method, and the sensitivity and specificity were 80.0% and 88.9% ,respectively.</p><p><b>CONCLUSION</b>Our findings suggest that SELDI profiling of serum might be a potential for gastric cancer detection and screening in high-risk population.</p>

Establishment of a diagnostic model of serum protein fingerprint pattern for esophageal cancer screening in high incidence area and its clinical value / 中华肿瘤杂志

<p><b>OBJECTIVE</b>To analyze the alterations of serum proteomic pattern in esophageal squamous cell carcinoma (ESCC) by SELDI-TOF-MS, to establish a diagnostic model of ESCC screening in high incidence area and investigate its clinical value.</p><p><b>METHODS</b>SELDI-TOF-MS and CM10 proteinChip were used to detect the serum proteomic patterns of 36 cases of ESCC and 38 healthy control subjects in high incidence area. The data were analyzed and a diagnostic model was established by using support vector machine (SVM). The diagnostic model was evaluated by leave-one-out cross validation.</p><p><b>RESULTS</b>At the molecular weight range of 2000 to 20,000, 31 protein peaks were significantly different between ESCC and controls (P < 0.01). A diagnostic model consisting of 4 protein peaks could do the best in diagnosis of ESCC and controls. The accuracy was 85.1%, sensitivity was 86.1%, specificity was 84.2%, and positive value was 83.8%.</p><p><b>CONCLUSION</b>The diagnostic model formed by 4 protein peaks, established in this study, can well distinguish ESCC from healthy subjects. It provides a new approach for ESCC screening in high incidence area.</p>

Preoperative molecular staging of colorectal cancers by CM10 ProteinChip and SELDI-TOF-MS analysis / 中华肿瘤杂志

<p><b>OBJECTIVE</b>To detect the serum proteomic patterns by using SELDI-TOF-MS and CM10 ProteinChip techniques in colorectal cancer (CRC) patients, and to evaluate the significance of the proteomic patterns in colorectal cancer staging.</p><p><b>METHODS</b>A total of 76 serum samples were obtained from CRC patients at different clinical stages, including Dukes A (n = 10), Dukes B (n = 19), Dukes C (n = 16) and Dukes D (n = 31). Different stage models were developed and validated by bioinformatics methods of support vector machines, discriminant analysis and time-sequence analysis.</p><p><b>RESULTS</b>The model I formed by six proteins of peaks at m/z 2759.6, 2964.7, 2048.0, 4795.9, 4139.8 and 37 761.6 could do the best as potential biomarkers to distinguish local CRC patients (Dukes A and Dukes B) from regional CRC patients (Dukes C ) with an accuracy of 86.7%. The model II formed by 3 proteins of peaks at m/z 6885.3, 2058.3 and 8567.8 could do the best to distinguish locoregional CRC patients (Dukes A, B and C) from systematic CRC patients (Dukes D) with an accuracy of 75.0%. The mode III could distinguish Dukes A from Dukes B with an accuracy of 86.2% (25/29). The model IV could distinguish Dukes A from Dukes C with an accuracy of 84.6% (22/26). The model V could distinguish Dukes B from Dukes C with an accuracy of 85.7% (30/35). The model VI could distinguish Dukes B from Dukes D with an accuracy of 80.0% (40/50). The model VII could distinguish Dukes C from Dukes D with an accuracy of 78.7% (37/47). Different stage groups could be distinguished by the two-dimensional scattered spots figure obviously.</p><p><b>CONCLUSION</b>Our findings indicate that this method can well be used in preoperative staging of colorectal cancers and the screened tumor markers may serve for guidance of integrating treatment of colorectal cancers.</p>

Preoperatively molecular staging with CM10 ProteinChip and SELDI-TOF-MS for colorectal cancer patients / 浙江大学学报（英文版）（B辑：生物医学和生物技术）

<p><b>OBJECTIVES</b>To detect the serum proteomic patterns by using SELDI-TOF-MS (surface enhanced laser desorption/ ionization-time of flight-mass spectrometry) technology and CM10 ProteinChip in colorectal cancer (CRC) patients, and to evaluate the significance of the proteomic patterns in the tumour staging of colorectal cancer.</p><p><b>METHODS</b>SELDI-TOF-MS and CM10 ProteinChip were used to detect the serum proteomic patterns of 76 patients with colorectal cancer, among them, 10 Stage I, 19 Stage II, 16 Stage III and 31 Stage IV samples. Different stage models were developed and validated by support vector machines, discriminant analysis and time-sequence analysis.</p><p><b>RESULTS</b>The Model I formed by 6 protein peaks (m/z: 2759.58, 2964.66, 2048.01, 4795.90, 4139.77 and 37761.60) could be used to distinguish local CRC patients (Stage I and Stage II) from regional CRC patients (Stage III) with an accuracy of 86.67% (39/45). The Model II formed by 3 protein peaks (m/z: 6885.30, 2058.32 and 8567.75) could be used to distinguish locoregional CRC patients (Stage I, Stage II and Stage III) from systematic CRC patients (Stage IV) with an accuracy of 75.00% (57/76). The Model III could distinguish Stage I from Stage II with an accuracy of 86.21% (25/29). The Model IV could distinguish Stage I from Stage III with accuracy of 84.62% (22/26). The Model V could distinguish Stage II from Stage III with accuracy of 85.71% (30/35). The Model VI could distinguish Stage II from Stage IV with accuracy of 80.00% (40/50). The Model VII could distinguish Stage III from Stage IV with accuracy of 78.72% (37/47). Different stage groups could be distinguished by the two-dimensional scattered spots figure obviously.</p><p><b>CONCLUSION</b>This method showed great success in preoperatively determining the colorectal cancer stage of patients.</p>

Establishment of diagnostic model of cerebrospinal protein fingerprint pattern for glioma and its clinical application / 浙江大学学报·医学版

<p><b>OBJECTIVE</b>To establish the diagnostic model of cerebrospinal protein profile for gliomas by surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF MS) and bioinformatics.</p><p><b>METHODS</b>Seventy-five samples of cerebrospinal fluid from patients with gliomas, benign brain tumors and mild brain traumas were collected. A total of 50 samples from gliomas and non-brain-tumors were divided into training sets (33 cases including 17 gliomas and 16 non-brain-tumors) and testing sets (17 cases including 5 gliomas and 12 non-brain-tumors). The cerebrospinal proteins bound to H4 chip were detected by SELDI-TOF MS, the profiles of cerebrospinal protein were gained and then analyzed with artificial neural network algorithm (ANN); and the diagnostic model of cerebrospinal protein profiles for differentiating gliomas from non-brain-tumors was established. Forty-seven of cerebrospinal samples of gliomas and benign brain tumors were divided into training sets (31 cases including 13 gliomas and 18 benign brain tumors) and testing sets (16 cases including 9 gliomas and 7 benign brain tumors), the diagnostic model of cerebrospinal protein profiles for differentiating gliomas from benign brain tumors was established based on the same method. The support vector machine (SVM) algorithm was also used for evaluation, both results were very similar, but the result derived from ANN was more stable than that from SVM.</p><p><b>RESULT</b>The diagnostic model of cerebrospinal protein profiles for differentiating gliomas from non-brain-tumors was established and was challenged with the test set randomly, the sensitivity and specificity were 100% and 91.7%, respectively. The cerebrospinal protein profiling model for differentiating gliomas from benign brain tumors was also developed and was challenged with the test set randomly, the sensitivity and specificity were 88.9%, and 100%, respectively.</p><p><b>CONCLUSION</b>The technology of SELDI-TOF MS which combined with analysis tools of bioinformatics is a novel effective method for screening and identifying tumor biomarkers of gliomas and it may provide a new approach for the clinical diagnosis of glioma.</p>

Application of serum protein fingerprinting coupled with artificial neural network model in diagnosis of hepatocellular carcinoma / 中华医学杂志(英文版)

<p><b>BACKGROUND</b>Hepatocellular carcinoma tends to present at a late clinical stage with poor prognosis. Therefore, it is urgent to explore and develop a simple, rapid diagnostic method, which has high sensitivity and specificity for hepatocellular carcinoma at an early stage. In this study, the serum proteins in patients with hepatocellular carcinoma or liver cirrhosis and in normal controls were analysed. Surface enhanced laser desorption/ionization time-of-flight mass (SELDI-TOF-MS) spectrometry was used to fingerprint serum protein using the protein chip technique and explore the value of the fingerprint, coupled with artificial neural network, to diagnose hepatocellular carcinoma.</p><p><b>METHODS</b>Of the 106 serum samples obtained, 52 were from patients with hepatocellular carcinoma, 22 from patients with liver cirrhosis and 32 from healthy volunteers. The samples were randomly assigned into a training group (n = 70, 35 patients with hepatocellular carcinoma, 14 with liver cirrhosis, and 21 normal controls) and a testing group (n = 36, 17 patients with hepatocellular carcinoma, 8 with liver cirrhosis, and 11 normal controls). An artificial neural network was trained on data from 70 individuals in the training group to develop an artificial neural network diagnostic model and this model was tested. The 36 sera in the testing group were analysed with blind prediction by using the same flowchart and procedure of data collection. The 36 serum protein spectra were clustered with the preset clustering method and the same mass/charge (M/Z) peak values as those in the training group. Matrix transfer was performed after data were output. Then the data were input into the previously built artificial neural network model to get the prediction value. The M/Z peaks of the samples with more than 2000 M/Z were normalized with biomarker wizard of ProteinChip Software version 3.1 for noise filtering. The first threshold for noise filtering was set at 5, and the second was set at 2. The 10% was the minimum threshold for clustering. The statistical analysis of the data of serum protein mass spectrum was performed in the groups (normal vs. hepatocellular carcinoma, and liver cirrhosis vs. hepatocellular carcinoma) with the t test.</p><p><b>RESULTS</b>Comparison between the groups of hepatocellular carcinoma and normal control: The mass spectra from 56 samples (hepatocellular carcinoma and normal controls) in the training group were analysed and 241 peaks were obtained. In addition, 21 peaks from them were used for comparison between the groups of hepatocellular carcinoma and normal controls (P < 0.01). Only 2 peaks at 3015 M/Z and 5900 M/Z were selected with significant difference (P < 10 (-9)). A model was developed based on these two proteins with different M/Z. It was confirmed that this artificial neural network model can be used for comparison between the groups of hepatocellular carcinoma and normal controls. The sensitivity was 100% (17/17), and the specificity was 100% (11/11). Comparison between the groups of hepatocellular carcinoma and liver cirrhosis: The mass spectra from 49 samples in the training group (including patients with hepatocellular carcinoma and liver cirrhosis) were analysed and 208 peaks were obtained. In addition, 21 peaks from them were used for comparison between the groups of hepatocellular carcinoma and liver cirrhosis (P < 0.01). Only 2 peaks at 7759 M/Z, 13134 M/Z were selected with significant difference (P < 10 (-9)). A model was developed based on these two proteins with different M/Z. It was confirmed that this artificial neural network model can be used for comparison between the groups of hepatocellular carcinoma and liver cirrhosis. The sensitivity was 88.2% (15/17), and the specificity was 100% (8/8).</p><p><b>CONCLUSIONS</b>The specific biomarkers selected with the SELDI technology could be used for early diagnosis of hepatocellular carcinoma.</p>

Serum protein fingerprinting coupled with artificial neural network distinguishes glioma from healthy population or brain benign tumor / 浙江大学学报（英文版）（B辑：生物医学和生物技术）

To screen and evaluate protein biomarkers for the detection of gliomas (Astrocytoma grade I-IV) from healthy individuals and gliomas from brain benign tumors by using surface enhanced laser desorption/ionization time of flight mass spectrometry (SELDI-TOF-MS) coupled with an artificial neural network (ANN) algorithm. SELDI-TOF-MS protein fingerprinting of serum from 105 brain tumor patients and healthy individuals, included 28 patients with glioma (Astrocytoma I-IV), 37 patients with brain benign tumor, and 40 age-matched healthy individuals. Two thirds of the total samples of every compared pair as training set were used to set up discriminating patterns, and one third of total samples of every compared pair as test set were used to cross-validate; simultaneously, discriminate-cluster analysis derived SPSS 10.0 software was used to compare Astrocytoma grade I-II with grade III-IV ones. An accuracy of 95.7%, sensitivity of 88.9%, specificity of 100%, positive predictive value of 90% and negative predictive value of 100% were obtained in a blinded test set comparing gliomas patients with healthy individuals; an accuracy of 86.4%, sensitivity of 88.9%, specificity of 84.6%, positive predictive value of 90% and negative predictive value of 85.7% were obtained when patient's gliomas was compared with benign brain tumor. Total accuracy of 85.7%, accuracy of grade I-II Astrocytoma was 86.7%, accuracy of III-IV Astrocytoma was 84.6% were obtained when grade I-II Astrocytoma was compared with grade III-IV ones (discriminant analysis). SELDI-TOF-MS combined with bioinformatics tools, could greatly facilitate the discovery of better biomarkers. The high sensitivity and specificity achieved by the use of selected biomarkers showed great potential application for the discrimination of gliomas patients from healthy individuals and gliomas from brain benign tumors.

An integrated approach utilizing proteomics and bioinformatics to detect ovarian cancer / 浙江大学学报（英文版）（B辑：生物医学和生物技术）

<p><b>OBJECTIVE</b>To find new potential biomarkers and establish the patterns for the detection of ovarian cancer.</p><p><b>METHODS</b>Sixty one serum samples including 32 ovarian cancer patients and 29 healthy people were detected by surface-enhanced laser desorption/ionization mass spectrometry (SELDI-MS). The protein fingerprint data were analyzed by bioinformatics tools. Ten folds cross-validation support vector machine (SVM) was used to establish the diagnostic pattern.</p><p><b>RESULTS</b>Five potential biomarkers were found (2085 Da, 5881 Da, 7564 Da, 9422 Da, 6044 Da), combined with which the diagnostic pattern separated the ovarian cancer from the healthy samples with a sensitivity of 96.7%, a specificity of 96.7% and a positive predictive value of 96.7%.</p><p><b>CONCLUSIONS</b>The combination of SELDI with bioinformatics tools could find new biomarkers and establish patterns with high sensitivity and specificity for the detection of ovarian cancer.</p>

The optimal combination of serum tumor markers with bioinformatics in diagnosis of colorectal carcinoma / 浙江大学学报·医学版

<p><b>OBJECTIVE</b>To identify the optimal combination of serum tumor markers with bioinformatics in diagnosis of colorectal cancer.</p><p><b>METHODS</b>The serum levels of CEA, AFP, NSE, CA199, CA242, CA724, CA211 and TPA were detected in 128 patients with colorectal carcinoma and 113 health subjects. The serum tumor markers were evaluated with the area under curves. The optimal combination of serum tumor markers was selected and the diagnostic model with artificial neural network was established.</p><p><b>RESULTS</b>CEA, CA199, CA242, CA211, CA724 were selected for the optimal combination and the artificial neural network was built. The model was evaluated by a 5-cross validation approach. The model had a specificity of 95%, sensitivity of 83% and positive predictive value of 95% in diagnosis of colorectal carcinoma.</p><p><b>CONCLUSION</b>The combination of optimal serum tumor markers has a high sensitivity and specificity in diagnosis of colorectal carcinoma.</p>

Application of serum protein pattern model in diagnosis of colorectal cancer / 中华肿瘤杂志

<p><b>OBJECTIVE</b>To explore the application of serum protein pattern models in diagnosis of colorectal cancer (CRC) by proteinchip technology.</p><p><b>METHODS</b>One hundred and forty-seven serum samples (55 CRC patients and 92 healthy individuals) randomly divided into training set (n = 87, 32 CRC patients and 55 healthy individuals) and test set (n = 60), were subjected for analysis by surface enhanced laser desorption/ionization-time of flight-mass spectrometry (SELDI-TOF-MS). Four top-scored peaks in 5910, 8930, 4476 and 8817 were detected by proteinchip software version 3.0. and were trained by a multi-layer artificial neural network (ANN) with a back propagation algorithm. An artificial neural network classifier had developed for separating CRC from the healthy group. The classifier was then challenged with the test set (60 samples including 23 CRC patients and 37 healthy individuals) to determine the validity and accuracy of the classification system.</p><p><b>RESULTS</b>The artificial neural network classifier separated the CRC from the healthy samples, with sensitivity of 82.6% and specificity of 91.9%.</p><p><b>CONCLUSION</b>Combination of SELDI-TOF-MS with the artificial neural network yields significant higher sensitivity and specificity than CEA in the diagnosis of CRC, which should be further studied.</p>