The diagnostic value of serum cysteine protease inhibitor (CST4) in colorectal cancer: a preliminary study.

CST4 is associated with various cancers but its diagnostic value in colorectal cancer (CRC) has not been clearly established. This study aims to further validate the diagnostic value of CST4 in colorectal cancer using random forest models in machine learning based on existing research. First, the expression of CST4 in CRC patients is analyzed using data from The Cancer Genome Atlas (TCGA) and the results show that CST4 is significantly high expression in CRC patients. To further investigate the diagnostic value of CST4, this retrospective study include 59 colorectal cancer patients, 51 patients with colorectal precancerous lesions and 40 patients with non-adenomatous polyps who visited Chaohu Hospital Affiliated to Anhui Medical University from August 2022 to September 2023. Their serum was collected and the CST4 level was detected by the double-antibody sandwich enzyme-linked immunosorbent assay (ELISA) method. At the same time, CEA and CA199 were also tested. Analysis of the TCGA database show that CST4 is significantly over expressed in colorectal cancer tissues, including colon adenocarcinoma (COAD) and rectum adenocarcinoma (READ). Serum tests confirm that CST4 levels in CRC patients are significantly higher than those in the precancerous lesion and the polyp group (P < 0.001). However, no significant difference is observed between the precancerous lesion and polyp group. In the ROC curve, the area under the curve (AUC) of CST4 is 0.768 in the CRC group versus the precancerous lesion group and 0.769 in the CRC group versus the polyp group, which is significantly higher than that of CEA (AUC = 0.651/0.659) and CA199 (AUC = 0.486/0.601). In the diagnostic model constructed based on CST4, CEA and CA199, CST4 has the highest weight in differentiating CRC, further confirming its diagnostic value. The study indicates that CST4 has a higher diagnostic value in colorectal cancer (CRC) and CST4 provides a new approach for early clinical diagnosis.