The significance of pro-gastrin-releasing peptide for small cell lung cancer diagnosis / 中华检验医学杂志

ABSTRACT

Objective To evaluate the clinical value of pro-gastrin-releasing peptide (ProGRP) for small cell lung cancer ( SCLC ).Methods Serum levels of ProGRP and neuron-specific enolase (NSE) were measured by both chemiluminescent immunoassay and electrochemiluminescent immunoassay in 46 patients with SCLC (26 patients with limited disease,20 patients with extensive disease ),51 patients with non-small cell lung cancer (NSCLC),45 patients with benign pulmonary diseases and 56 healthy subjects.Patients were recruited by the Affiliated Hospital of Medical College,Qingdao University,from September 2010 to April 2011.The receiver operating characteristic curves (ROC) was used to set the cutoff value of ProGRP and NSE and the areas under ROC ( ROC-AUC).The sensitivity and specificity of ProGRP and NSE were analyzed for diagnosing SCLC.Results Serum levels of ProGRP in healthy subjects,benign pulmonary diseases,NSCLC and SCLC groups were 22.9 ( 19.5 - 28.7 ),23.7 ( 20.0 - 27.8 ),28.9 (23.8-34.7) and 370.9( 129.4- 1951.6) ng/L respectively; the serum levels of NSE were 14.1 (12.5- 15.7),13.3(10.3- 15.3),16.8(11.7-22.1) and 39.9(16.1-93.9) μg/L,respectively.The Kruskal-Wallis H test showed significantly difference amoun groups of ProGRP and NSE (H =92.116 and 55.481,P ＜0.001 ).The serum levels of ProGRP in limited disease SCLC (LD-SCLC) group[ 156.2(65.4-547.5 ) ng/L]were also significantly higher than those in the healthy group,benign pulmonary diseases group and NSCLC group ( U =57,70 and 144,P ＜ 0.001 ).In extensive disease SCLC (ED-SCLC) group,the ProGRP and NSE results[ 1933.1 (325.9 -4512.1) ng/L and 61.0(35.4- 115.5 ) μg/L ]were higher than those in the LD-SCLC group ProGRP,NSE [ 24.3 ( 15.1 - 16.3 ) μg/L,U =119 and 153,P ＜ 0.05 ].Using healthy subjects group as control,the largest Youden index point of ROC was used to set the cut-off value of ProGRP and NSE (34.0 ng/L and 20.2 μg/L).The ROC-AUC of ProGRP (0.96 ) was statistically higher than that of NSE ( 0.86 ) in the SCLC group ( Z =2.57,P ＜0.05).The ROC-AUC results between combining detection of ProGRP and NSE (0.96 ) and ProGRP itself (0.96) were not significant difference ( Z =0.21,P ＞ 0.05 ).The sensitivity of ProGRP ( 89.1％ ) was statistically higher than that of NSE in the SCLC group (71.7％,x2 =4.90,P ＜0.05 ) ; the specificity of ProGRP (98.2％) compared with NSE did not have statistical significance (96.4％,x2 =0.00,P ＞0.05 ).The combining detection of ProGRP and NSE had no influence on the sensitivity and specificity compared with ProGRP itself (91.3％ vs 89.1％,94.6％ vs 98.2％,x2 were all 0.00,P ＞ 0.05 ).Using benign pulmonary diseases group as control,the largest Youden index point of ROC was used to set the cutoff value of ProGRP and NSE (49.5 ng/L and 23.1 μg/L).The ROC-AUC of ProGRP (0.95) was statistically higher than that of NSE (0.87) in the SCLC group (Z =1.99,P ＜0.05 ).The ROC-AUC of combining detection of ProGRP and NSE ( 0.95 ) and ProGRP itself ( 0.95 ) were not difference significantly ( Z =0.02,P ＞ 0.05 ).The sensitivity of ProGRP (84.8％ ) was statistically higher than that of NSE in the SCLC group (69.6％,x2 =4.00,P ＜0.05);the specificity of it (97.8％) was equal to that of NSE (97.8％,x2 =0.50,P ＞0.05 ).The combining detection of ProGRP and NSE had no obviously influence on the sensitivity and specificity compared with ProGRP itself ( 87.0％ vs 84.8％,95.6％ vs 97.8％,x2 were all 0.00,P ＞0.05 ).Using NSCLC group as control,the largest Youden index point of ROC was to set the cut-off value of ProGRP and NSE (49.1 ng/L and 23.0 μg/L).The ROC-AUC of ProGRP ( 0.90) was statistically higher than that of NSE (0.76) in the SCLC group (Z=2.90,P＜0.05).The ROC-AUC of combining detection of ProGRP and NSE (0.90 ) and ProGRP itself (0.90 ) were not difference significantly ( Z =0.00,P ＞ 0.05 ).The sensitivity of ProGRP ( 84.8％ ) was higher than that of NSE in the SCLC group ( 69.6％,x2 =4.00,P ＜ 0.05 ) ; the specificity of it ( 96.1％ ) was also higher than that of NSE (80.4％,x2 =6.13,P ＜ 0.05 ).The combining detection of ProGRP and NSE had no obviously influence on the sensitivity and specificity compared with ProGRP itself ( 87.0％ vs 84.8％,95.6％ vs 96.1％,x2 were all 0.00,P ＞ 0.05 ).Conclusion ProGRP has a higher diagnostic value than NSE in SCLC.