| Pancreatic cancer is highly malignant, develops obscurely and fast, andprognosis is very poor. Therefore, effective diagnostic marker is in great needfor the clinical work. SELDI-TOF-MS (Surface enhanced laserdesorption/ionization-time of flight-mass spectrometry) is one the importanttechnologies in proteomics study, harboring the virtue of simplicity,high-throughput and high sensitivity. $ELDI-TOF-MS has been widely used inscreening tumor markers and made great progress in clinical proteomics. Thepurpose of the study was to use SELDI-TOF-M$ to analyze the serumprotein/peptide profile's difference between the pancreatic cancer patients andnormal individuals, to establish pancreatic cancer diagnostic pattern, to testifythe effectiveness of the diagnostic pattern and to combinatorially diagnosetogether with CA19-9.CM10 protein chip was selected to analyze serum protein fingerprint mapfrom 132 pancreatic cancer patients and 81 healthy volunteers. To analyze thedata, the specimens were grouped into training cluster (specimens ofestablishing pattern) and testing cluster (specimens of blind screening).Trainingcluster included serum specimens from 30 pancreatic cancer patients and 39normal controls. Protein profile's difference between pancreatic cancer andnormal controls were analyzed by SVM and to establish diagnostic pattern.Testing cluster is composed of 102 pancreatic cancer patients and 42 normalcontrols' serum which are excluded by the training cluster. Diagnostic patternwere used to blind screen testing cluster, testify the effectiveness andcombinatorially diagnose were adopted with such pattern and serum markerCA19-9.The following studies were completed:Acquirement of differential peaks from the specimens: ProteinChipSoftware 3.2.0 was used to collect MS signal, total 200 effective differentialpeaks were detected between MW 2000 Da and 20000 Da by using CM10 chip, 128 protein/peptide peaks have significantly differentiation(P＜0.05).Establishment of pancreatic cancer diagnostic pattern: Training clusterincluded serum specimens from 30 pancreatic cancer patients and 39 normalcontrols. SVM was used to establish diagnostic pattern and the patternconsists of 4 peaks derived from peak1, peak2, peak3 and peak4. MW rangedfrom 5300 to 8600Da. The specificity of the diagnostic pattern to thespecimens of establishing pattern was 100％(39/39), and the sensitivity was96.67％(29/30).Confirmation of the pancreatic cancer diagnostic pattern: The specificityof the diagnostic pattern to the specimens of healthy controls excluded bytraining cluster was 85.71％(36/42). The sensitivity of the diagnostic pattern tothe specimens of pancreatic cancer patients excluded by training cluster was93.14％(95/102). For the total 132 pancreatic cancer patients and 81 healthyvolunteers, the sensitivity is 93.94％(124/132)and the specificity was92.59％(75/81). Moreover, the sensitivity of the diagnostic pattern to the stagela (early stage) pancreatic cancer is 100％(11/11), showing the great accuracyof the pattern in diagnosing early stage pancreatic cancer.Combinatorial diagnosis by the pattern and CA19-9: The patients withCA19-9＞37U/ml were judged positive. The sensitivity of CA19-9 in diagnosing132 pancreatic cancer patients from testing cluster is 78.79％(104/132). Thenthe pattern and CA19-9 were combinatorially used to diagnose, the resultingdata showed that the sensitivity is 97.73％(129/132), dramaticallyoutperformed the serum marker CA19-9.In summary, pancreatic cancer diagnostic pattern could be established byusing SELDI-TOF-MS to analyze serum proteins profile of the pancreaticcancer patients and normal individuals and combining SVM. The patternconsisting of 4 differential peaks hold great pancreatic cancer diagnosticaccuracy. And it might well enhance our capability of detecting early pancreaticcancer by serum assays. Furthermore, combining the established pattern andserum CA19-9 improved the pancreatic cancer diagnostic accuracy. |
PDF Full Text Request