| ObjectivePrecise preoperative staging is necessary to the individualized therapy of colorectal cancer(CRC). As tumor-staging and lymph node-staging are mainly parts of clinical TNM, we aimed to discover the distinct proteins of CRC stageâ… and lymph node staging, and establish the pattern-matching algorithms of stageâ… and local lymphatic metastasis of CRC respectively by serum protein profiling, which are potential to develop new technique of stageâ… diagnosis and lymph node-staging of CRC by surface enhanced laser desorption/ionization- time of flight- mass spectrometry(SELDI-TOF-MS).Methods245 cases of serum samples were adopted in our study, which included sporadic moderately differentiated colorectal adenocarcinoma stageâ… ,â…¡andâ…¢(31, 74 and 105 cases respectively), 16 cases of healthy people and 19 cases of colorectal nonmalignant patients age- and gender-matched to stageâ… . All the serum samples were applied to immobilized metal affinity capture(IMAC30-Cu2+) protein chips to generate mass pectra by SELDI-TOF-MS. Protein peaks identification, dectection and clustering were performed using the Biomarker Wizard software. Firstly, Proteomic spectra of serum samples from 31 sporadic moderately differentiated colorectal adenocarcinoma patients of stageâ… , and age- and gender-matched 35 controls were regarded as a learning mode and CRC stageâ… diagnosis classification tree consisting of 9 distinct proteins was generated by Biomarker Pattern software. Then the classsification tree was validated in test mode. We additionally determined carcinoembryonic antigen(CEA) in all the serumal samples included these 66 samples using an euzymelinked immunosorbent. Secondly, Proteomic spectra of serum samples from 70 CRC patients with local lymphatic metastasis and age- and gender-matched 75 CRC patients without local lymphatic metastasis were regarded as a training set and a node staging classification tree consisting of 5 distinct proteins was generated by Biomarker Pattern software. Then the classsification tree was validated with a test set double blindly including another 35 CRC patients with local lymphatic metastasis and age- and sex-matched 30 CRC patients without local lymphatic metastasis. ResultsFirstly, the Biomarker Wizard software identified an average of 61 peaks/spectrum, 9 of which were distinct(P<0.05)and 8 of the identified peaks at 2,787, 3,777, 3,816, 3,852, 5,065, 5,828, 5,855, and 4,172 were used to construct the CRC stageâ… diagnosis classification tree. The stageâ… diagnosis classification tree separated effectively CRC stageâ… from health and nonmalignant individuals, achieving a sensitivity and specificity of 93.55% and 91.43% respectively in the learning mode, and the corresponding sensitivity and specificity were 70.97% and 82.86%, and a positive predictive value was 78.57% in test mode, which surpassed CEA significantly(P<0.05).Secondly, the BioMarker Wizard software identified an average of 46 mass peaks/spectrum, 22 of which were distinct(P<0.05)and 5 of the identified peaks at 3,104, 3,781, 5,867, 7,970, 9,290 were used to construct the CRC node staging classification tree. The node staging classification tree separated effectively CRC with local lymphatic metastasis from those without local lymphatic metastasis, achieving a sensitivity of 94.29% and a specificity of 100.00%. The double blinded test challenged the model with a sensitivity of 96.67% and a specificity of 91.43%, and a positive predictive value of 90.63%.Conclusions9 and 22 proteins were detected as the distinct proteins of CRC stageâ… and lymph node stageing respectively, which were related to the development of CRC.Via comparative proteomics research, the results suggested that the CRC stageâ… diagnosis classification tree correctly distinguished CRC stageâ… from healthy and nonmalignant individuals, and showed great potential for the development of a screening test for the detection of CRC stageâ… , and the node staging classification tree correctly distinguished CRC with local lymphatic metastasis from those without local lymphatic metastasis and showed great potential for the development of a screening test for the identification of CRC patients with or without local lymphatic metastasis.
|
PDF Full Text Request