Construction and validation of a predictive model for the risk of myelosuppression after first-line chemotherapy in patients with advanced colorectal cancer

WANG Xuexing; ZHANG Rong; CHU Jie; LIU Zhijin

doi:10.12173/j.issn.1004-5511.202408071

Abstract：Objective To investigate the risk factors influencing the incidence of chemotherapy-induced myelosuppression(CIM)following first-line chemotherapy in patients with moderately advanced colorectal cancer(CRC),and to develop and validate a nomogram to assess the risk of CIM in the patients.Methods The clinical data of patients with stage Ⅱ-Ⅳ CRC who received first-line chemotherapy at Anning First People's Hospital affiliated with Kunming University of Science and Technology between July 2021 and January 2024,were retrospectively analyzed.The World Health Organization classification standard for acute and subacute toxicity of anticancer drugs were used as the standards for diagnosing CIM,and the patients were subsequently categorized into CIM and non-CIM groups.The risk factors for CIM were analyzed and a multivariate Logistic regression model was employed to construct the predictive model.The model's discrimination and accuracy were assessed using the receiver operating characteristic curve(ROC)and area under curve(AUC)and the Hosmer-Lemeshow goodness of fit test.Additionally,the model's clinical utility was evaluated using calibration and clinical decision curves.Results A total of 257 patients were included,112 individuals exhibited bone CIM,corresponding to an incidence rate of 43.58％.The most prevalent severity was classified as grade Ⅰ-Ⅱ CIM.Multivariate Logistic regression analysis identified the chemotherapy cycle,reductions in white blood cell count,hemoglobin levels,and platelet count prior to chemotherapy as independent risk factors for CIM in CRC patients(P＜0.05).The AUC for the final nomogram was 0.828[95％CI(0.779,0.878)].The model demonstrated a sensitivity of 66.1％and a specificity of 85.5％.The calibration curve and clinical decision curve suggested that the nomogram model prediction had good clinical practical value.Conclusion The nomogram prediction model has predictive value for assessing the risk of CIM in CRC patients.This model can serve as a valuable reference for clinicians in making informed decisions regarding the prevention of CIM.