OBJECTIVES: The objectives of this study were to evaluate the diagnostic value for ovarian cancer using proteomic pattern established by surface-enhanced laser desorption/ionization (SELDI-TOF-MS) profiling of plasma proteins coupled with support vector machine (SVM) data analysis, and to investigate whether the proteomic pattern established by advanced ovarian cancer could be used for diagnosis of early-stage ovarian cancer patients. METHODS: The study included 44 ovarian cancer patients (11 early-stage and 33 advanced ovarian cancer patients) and 31 age-matched non-cancer controls. SELDI-TOF-MS coupled with SVM analysis was performed to establish a proteomic pattern to discriminate 33 advanced ovarian cancer patients from 31 non-cancer controls. A blind test, including 11 early-stage ovarian cancer cases, was performed to investigate whether proteomic pattern established by advanced ovarian cancer could be used for diagnosis of early-stage ovarian cancer patients. RESULTS: A seven-peak proteomic pattern was established which discriminated 33 advanced ovarian cancer patients from 31 non-cancer controls effectively. A sensitivity of 93.94% (31/33) and a specificity of 93.55% (29/31) were yielded from the proteomic pattern. Among the 7 protein peaks, 5 with mass charge ratio (m/z) 4099 Da, 5488 Da, 4144 Da, 4479 Da and 3940 Da were up-regulated, while 2 peaks, with m/z 13 783 Da and 8588 Da were down-regulated in the advanced ovarian cancer group compared with non-cancer control group. After blind test, 9 of 11 early-stage ovarian cancer patients were successfully diagnosed with the accuracy of 81.82% (9/11). CONCLUSIONS: This study demonstrated that SELDI-TOF-MS coupled with SVM is effective in distinguishing protein expression between ovarian cancer and non-cancer plasma and it may be feasible to diagnose early-stage ovarian cancer using proteomic pattern established by advanced ovarian cancer. The gained and lost protein peaks in plasma may exist in both early-stage and advanced ovarian cancer plasma. Further studies should be performed using larger sample numbers.
OBJECTIVES: The objectives of this study were to evaluate the diagnostic value for ovarian cancer using proteomic pattern established by surface-enhanced laser desorption/ionization (SELDI-TOF-MS) profiling of plasma proteins coupled with support vector machine (SVM) data analysis, and to investigate whether the proteomic pattern established by advanced ovarian cancer could be used for diagnosis of early-stage ovarian cancerpatients. METHODS: The study included 44 ovarian cancerpatients (11 early-stage and 33 advanced ovarian cancerpatients) and 31 age-matched non-cancer controls. SELDI-TOF-MS coupled with SVM analysis was performed to establish a proteomic pattern to discriminate 33 advanced ovarian cancerpatients from 31 non-cancer controls. A blind test, including 11 early-stage ovarian cancer cases, was performed to investigate whether proteomic pattern established by advanced ovarian cancer could be used for diagnosis of early-stage ovarian cancerpatients. RESULTS: A seven-peak proteomic pattern was established which discriminated 33 advanced ovarian cancerpatients from 31 non-cancer controls effectively. A sensitivity of 93.94% (31/33) and a specificity of 93.55% (29/31) were yielded from the proteomic pattern. Among the 7 protein peaks, 5 with mass charge ratio (m/z) 4099 Da, 5488 Da, 4144 Da, 4479 Da and 3940 Da were up-regulated, while 2 peaks, with m/z 13 783 Da and 8588 Da were down-regulated in the advanced ovarian cancer group compared with non-cancer control group. After blind test, 9 of 11 early-stage ovarian cancerpatients were successfully diagnosed with the accuracy of 81.82% (9/11). CONCLUSIONS: This study demonstrated that SELDI-TOF-MS coupled with SVM is effective in distinguishing protein expression between ovarian cancer and non-cancer plasma and it may be feasible to diagnose early-stage ovarian cancer using proteomic pattern established by advanced ovarian cancer. The gained and lost protein peaks in plasma may exist in both early-stage and advanced ovarian cancer plasma. Further studies should be performed using larger sample numbers.