Tivity increased to 79 with specificity of 78 . doi:10.1371/journal.pone.0055171.117793 biological activity gDiagnosis Efficacy of NGAL, MIC-1 and CA19-Table 3. Diagnostic potential? of NGAL, MIC-1 and CA19-9 at pre-defined cut-off.Groups PC vs. HC CA19-9 MIC-1 NGAL PC vs. CP CA19-9 MIC-1 NGAL Stage 1/2 PC vs. HC CA19-9 MIC-1 NGAL Stage 3/4 PC vs. HC CA19-9 MIC-1 NGAL Stage 1/2 PC vs. CP CA19-9 MIC-1 NGAL Stage 3/4 PC vs. CP CA19-9 MIC-1 NGALPre-defined cut-offSensitivitySpecificity37 U/ml 1.07 ng/ml 106 ng/ml83 90 4267 46 9237 U/ml 1.07 ng/ml 106 ng/ml83 90 4261 30 5237 U/ml 1.07 ng/ml 106 ng/ml71 94 4667 46 9237 U/ml 1.07 ng/ml 106 ng/ml88 90 4467 46 92levels for PC (r = 0.179 p = 0.080) and CP cases (r = 0.459 p = 0.042). For the present study, due to the skewed distribution of biomarker levels, each biomarker measurement was log transformed (into its natural logarithm, to the base e = 2.7183) 
prior to comparison of mean levels between the three groups of patients. For the purposes of presentation, data has been reverselog transformed to allow the inclusion of units. The intra and interassay percent coefficient of variation ( CV) for NGAL and MIC1 were 4.1 , 14.3 , 5.9 and 16.1 respectively. Due to the presence of high and low standards built into the commercial kit, these coefficients were not determined for CA19-9. The mean plasma concentration (after log transformation) of NGAL, MIC-1 and CA19-9 were all significantly Avasimibe supplier higher in PC patients (111.1 ng/mL, 4.5 ng/mL, and 219.2 U/mL) than in the healthy controls (67.4 ng/mL (p = 0.01), 1.5 ng/mL (p = 0.003), and 31.5 U/mL (p = 0.001)). Additionally, serum concentration of MIC-1 and CA19-9, but not NGAL, were found to be higher in the PC patient group than in CP patients (1.6 ng/mL (p = 0.003), 31.8 U/mL (p,0.001), and 111.1 ng/mL (P.0.05) respectively) (Table 2). NGAL levels were significantly higher in patients aged 60 years or more (p = 0.045). MIC-1 levels were significantly lower in ever smokers compared to never smokers (p = 0.021). CA19-9 levels on the other hand were significantly elevated in female PC patients and in those with unresectable disease (Stage 3/4, p = 0.045 and 0.0047 respectively) (data not shown).37 U/ml 1.07 ng/ml 106 ng/ml71 94 4661 30 52Diagnostic Accuracy of NGAL, CA19-9 and MIC-We next sought to investigate the sensitivity and specificity of the three biomarkers for diagnosing PC. PC patients were divided either based on disease stage or treatment status. As post-treatment ?samples are not diagnostically relevant, only treatment naive samples were included in these analyses. In order to check diagnostic efficacy of CA19-9, MIC-1 and NGAL, these markers were evaluated at predefined cut-off of 37 U/ml, 1.07 ng/ml, 106 ng/ml as observed in earlier studies [3,6]. During this validation, NGAL was found to be 92 sensitive while MIC-1 was most specific 15755315 (94 ) in distinguishing early stage 1/2 patients from healthy controls (Table 3). However, overall performance of all the markers was quite poor. Further, we evaluated their diagnostic efficacy at optimal cut-off. For CA19-9, apart from the commonly employed cut-off value of 37 U/ml, we also used optimal cut-off (55.1 U/ml) 
as determined by ROC curve analysis. In comparison of both PC to HC and PC to CP patients, use of an higher cut-off of CA19-9 resulted in higher specificity with similar sensitivity in distinguishing PC from either CP or HCs (Figure 1) (Table 4). For all the further analysis, we used.Tivity increased to 79 with specificity of 78 . doi:10.1371/journal.pone.0055171.gDiagnosis Efficacy of NGAL, MIC-1 and CA19-Table 3. Diagnostic potential? of NGAL, MIC-1 and CA19-9 at pre-defined cut-off.Groups PC vs. HC CA19-9 MIC-1 NGAL PC vs. CP CA19-9 MIC-1 NGAL Stage 1/2 PC vs. HC CA19-9 MIC-1 NGAL Stage 3/4 PC vs. HC CA19-9 MIC-1 NGAL Stage 1/2 PC vs. CP CA19-9 MIC-1 NGAL Stage 3/4 PC vs. CP CA19-9 MIC-1 NGALPre-defined cut-offSensitivitySpecificity37 U/ml 1.07 ng/ml 106 ng/ml83 90 4267 46 9237 U/ml 1.07 ng/ml 106 ng/ml83 90 4261 30 5237 U/ml 1.07 ng/ml 106 ng/ml71 94 4667 46 9237 U/ml 1.07 ng/ml 106 ng/ml88 90 4467 46 92levels for PC (r = 0.179 p = 0.080) and CP cases (r = 0.459 p = 0.042). For the present study, due to the skewed distribution of biomarker levels, each biomarker measurement was log transformed (into its natural logarithm, to the base e = 2.7183) prior to comparison of mean levels between the three groups of patients. For the purposes of presentation, data has been reverselog transformed to allow the inclusion of units. The intra and interassay percent coefficient of variation ( CV) for NGAL and MIC1 were 4.1 , 14.3 , 5.9 and 16.1 respectively. Due to the presence of high and low standards built into the commercial kit, these coefficients were not determined for CA19-9. The mean plasma concentration (after log transformation) of NGAL, MIC-1 and CA19-9 were all significantly higher in PC patients (111.1 ng/mL, 4.5 ng/mL, and 219.2 U/mL) than in the healthy controls (67.4 ng/mL (p = 0.01), 1.5 ng/mL (p = 0.003), and 31.5 U/mL (p = 0.001)). Additionally, serum concentration of MIC-1 and CA19-9, but not NGAL, were found to be higher in the PC patient group than in CP patients (1.6 ng/mL (p = 0.003), 31.8 U/mL (p,0.001), and 111.1 ng/mL (P.0.05) respectively) (Table 2). NGAL levels were significantly higher in patients aged 60 years or more (p = 0.045). MIC-1 levels were significantly lower in ever smokers compared to never smokers (p = 0.021). CA19-9 levels on the other hand were significantly elevated in female PC patients and in those with unresectable disease (Stage 3/4, p = 0.045 and 0.0047 respectively) (data not shown).37 U/ml 1.07 ng/ml 106 ng/ml71 94 4661 30 52Diagnostic Accuracy of NGAL, CA19-9 and MIC-We next sought to investigate the sensitivity and specificity of the three biomarkers for diagnosing PC. PC patients were divided either based on disease stage or treatment status. As post-treatment ?samples are not diagnostically relevant, only treatment naive samples were included in these analyses. In order to check diagnostic efficacy of CA19-9, MIC-1 and NGAL, these markers were evaluated at predefined cut-off of 37 U/ml, 1.07 ng/ml, 106 ng/ml as observed in earlier studies [3,6]. During this validation, NGAL was found to be 92 sensitive while MIC-1 was most specific 15755315 (94 ) in distinguishing early stage 1/2 patients from healthy controls (Table 3). However, overall performance of all the markers was quite poor. Further, we evaluated their diagnostic efficacy at optimal cut-off. For CA19-9, apart from the commonly employed cut-off value of 37 U/ml, we also used optimal cut-off (55.1 U/ml) as determined by ROC curve analysis. In comparison of both PC to HC and PC to CP patients, use of an higher cut-off of CA19-9 resulted in higher specificity with similar sensitivity in distinguishing PC from either CP or HCs (Figure 1) (Table 4). For all the further analysis, we used.