[Peripheral blood cell count composite score as a prognostic factor in patients with colorectal cancer]
P Y Guo,X H Hu,B K Li,T Lu,J M Liu,C Y Wang,W B Niu,G Y Wang,B Yu
DOI: https://doi.org/10.3760/cma.j.cn441530-20231029-00151
2024-09-25
Abstract:Objective: To develop a prognostic prediction model for patients with colorectal cancer based on a peripheral blood cell composite score (PBCS) system. Methods: This retrospective observational study included patients who had primary colorectal cancer without distant metastasis, who did not undergo radiotherapy or chemotherapy before surgery, who did not receive leukocyte or platelet-raising therapy within 1 month before surgery, and whose postoperative pathology confirmed colorectal adenocarcinoma with complete tumor resection. Patients with severe anemia, infection, or hematologic diseases before surgery, as well as those with severe heart, lung, or other important organ diseases or concurrent malignant tumors, were excluded. In total, 1021 patients with colorectal cancer who underwent surgical treatment in the Department of Gastrointestinal Surgery of the Fourth Hospital of Hebei Medical University from April 2018 to April 2020 were retrospectively included as the training set (766 patients) and the internal validation set (255 patients). Additionally, using the same criteria, 215 patients with colorectal cancer who underwent surgical treatment in another treatment group from March 2015 to December 2020 were selected as the external validation set. The "surv_cutpoint" function in R software was used to analyze the optimal cut-off values of neutrophils, lymphocytes, and platelets, and a PBCS system was established based on the optimal cut-off values. The scoring rules of the PBCS system were as follows: Neutrophils and platelets below the optimal cut-off value = 1 point, otherwise 0 points; Lymphocytes above the optimal cut-off value = 1 point, otherwise 0 points. The scores of the three cell types were added together to obtain the PBCS. Univariate and multivariate Cox regression analyses were performed to explore the correlation between patients' clinicopathological features and prognosis, and a nomogram was constructed based on the Cox regression analysis to predict patients' prognosis. The accuracy of the nomogram prediction model was validated using the C-index, calibration curve, and decision curve analysis. Results: The optimal cut-off values for neutrophils, lymphocytes, and platelets were 4.40×109/L, 1.41×109/L, and 355×109/L, respectively. The patients were divided into high and low groups according to the optimal cut-off values of these cells. Survival curve analysis showed that a high lymphocyte count (training set: P=0.042, internal validation: P=0.010, external validation: P=0.029), low neutrophil count (training set: P=0.035, internal validation: P=0.001, external validation: P=0.024), and low platelet count (training set: P=0.041, internal validation: P=0.030, external validation: P=0.024) were associated with prolonged overall survival (OS), with statistically significant differences in all cases. Survival analysis of different PBCS groups showed that patients with a high PBCS had longer OS than those with a low PBCS (P<0.05). Univariate and multivariate Cox regression analysis results showed that aspirin use history, vascular thrombus, neural invasion, CA19-9, N stage, operation time, M stage, and PBCS were independent factors affecting OS (all P<0.05). The PBCS was also an independent factor affecting disease-specific survival (P<0.05), but not progression-free survival (P>0.05). The above independent risk or protective factors were included in R software to construct a nomogram for predicting OS. The C-index (0.873), calibration curve, and decision curve analysis (threshold probability: 0.0%-75.2%) all indicated that the nomogram prediction model had good predictive performance for OS. Conclusion: This study demonstrates that the PBCS constructed based on preoperative peripheral blood levels of neutrophils, lymphocytes, and platelets is an independent factor associated with the prognosis of patients with colorectal cancer. The nomogram model constructed based on this score system exhibits good predictive efficacy for the prognosis of these patients.