Abstract:

ObjectiveTo screen blood protein markers of colorectal cancer（CRC） in ProteomeXchange database，and to establish a screening model for evaluating its diagnostic value for CRC. Methods The differentially expressed protein of CRC was screened in PXD018304 dataset of ProteomeXchange database，and bioinformatics analysis was performed. From May 2021 to January 2022，108 newly diagnosed CRC patients（CRC group） and 100 healthy subjects（healthy control group） in Renji Hospital Affiliated to Shanghai Jiao Tong University School of Medicine were enrolled and classified into training set and validation set according to the ratio of 8∶2. The differentially expressed proteins and 5 tumor markers [carcinoembryonic antigen（CEA），carbohydrate antigen（CA） 19-9，CA242，CA50，CA72-4] were determined. Stepwise binary Logistic regression analysis（backward likelihood ratio method） was used to establish a CRC screening model. Receiver operating characteristic（ROC） curve was used to evaluate the efficacy of single and combined determinations in the diagnosis of CRC. Results A total of 350 differentially expressed proteins were screened from the PXD018304 dataset，which included 214 up-regulated proteins and 136 down-regulated proteins. According to the results of Gene Ontology（GO） enrichment analysis，Kyoto Encyclopedia of Genes and Genomes（KEGG） pathway analysis，protein-protein interaction（PPI） network and other bioinformatics analysis for differentially expressed proteins，taking into account clinical popularity，10 candidate proteins were screened out. The down-regulated proteins were apolipoprotein A1（apo A1），fibronectin（FN），glutathione reductase（GR） and transferrin（TRF），and the up-regulated proteins were apolipoprotein C3（apo C3），ceruloplasmin（CER），C-reactive protein（CRP），complement 4（C4），fibrinogen（Fib）and beta₂-microglobulin（β₂-MG）. Compared with healthy control group，CRC group had lower serum levels of apo A1，apo C3，FN，GR and TRF（P<0.001）. The levels of serum CER，CRP，C4，β₂-MG，CEA，CA19-9 and plasma Fib were increased（P<0.001）. There was no statistical significance in serum CA242，CA50 and CA72-4 levels between the 2 groups（P>0.05）. The markers with areas under curves（AUC） >0.7 for CRC diagnosis were apo A1，apo C3，CRP，FN，GR，TRF，β₂-MG and CEA. The highest single determination efficiency was apo A1（AUC=0.898），with a sensitivity of 81.48% and a specificity of 86.00%. The AUC of the screening model combined with apo A1，CRP，FN，TRF and CEA was 0.959，the sensitivity was 87.21%，the specificity was 92.50%，and the accuracy was 89.76%. Conclusions Apo A1 and other 10 proteins may be used as biomarkers for CRC screening. The screening model combined with apo A1，CRP，FN，TRF and CEA has a high diagnostic efficacy for CRC，which can provide a reference for clinical CRC screening.

Key words: Screening model, Proteomics, Biomarker, Colorectal cancer