To analyze the status and risk factors of nosocomial infection in patients with malignant tumors during chemotherapy, and to construct a nomogram model for predicting the risk of infection.

Total of 402 patients with malignant tumors who were hospitalized for chemotherapy in Suzhou Ninth Hospital Affiliated to Soochow University from January 2024 to October 2024 were selected. The patients showed no signs of infection upon admission. The patients were divided into infection group (47 cases) and non-infection group (355 cases) according to whether nosocomial infection occurred during hospitalization. The clinical data of the two groups were compared. The risk factors of nosocomial infection in cancer patients undergoing chemotherapy were analyzed by multivariate Logistic regression analysis, and nomogram model was constructed to predict the risk of infection. The predictive efficacy of Logistic regression model and nomogram model were analyzed by receiver operating characteristic (ROC) curve. The two models were compared by Delong test. The stability and clinical value of the dominant model were evaluated by K-fold cross-validation and decision curve.

The hospital infection rate for the first hospitalization chemotherapy of 402 patients with malignant tumors was 11.69% (47/402), the main infection site was respiratory system (27 cases, 57.45%). Among the 47 infected specimens, 30 strains (51.72%) of Gram-negative bacteria, 24 strains (41.38%) of Gram-positive bacteria, and 4 strains (6.90%) of fungi were isolated and cultured. The ages of patients in the infection group and the non-infection group [(63.96 ± 6.85) years old vs. (60.22 ± 5.94) years old: t=3.982, P < 0.001], complicated with diabetes [17 (36.17%) vs. 69 (19.44%): χ²=6.911, P=0.009], neutrophil count before chemotherapy [(2.39 ± 0.47) vs. (2.59 ± 0.54) × 10⁹/L]: t=2.038, P=0.042), neutrophil count to lymphocyte count ratio (NLR) before chemotherapy [(1.07 ± 0.26) vs. (0.79 ± 0.24): t=7.442, P < 0.001], nutritional risks before chemotherapy [15 (31.91%) vs. 62 (17.46%): χ²=5.597, P=0.018)] and invasive operations [39 (82.98%) vs. 225 (63.38%): χ²=7.072, P=0.008] were all with significant differences. Multivariate Logistic regression analysis showed that age (OR=2.775, 95%CI: 1.415-5.447, P=0.003), complicated with diabetes (OR=2.106, 95%CI: 1.157-3.834, P=0.015), NLR before chemotherapy (OR=3.557, 95%CI: 1.763-7.178, P < 0.001), nutritional risk before chemotherapy (OR=1.679, 95%CI: 1.059-2.662, P=0.028), invasive procedures (OR=2.391, 95%CI: 1.224-4.673, P=0.011) were risk factors for nosocomial infection of tumor patients undergoing chemotherapy in hospital. ROC curve analysis and Delong test showed that the area under the curve (AUC) of nomogram in predicting the risk of nosocomial infection in patients with tumor chemotherapy was significantly higher than that of Logistic regression model (0.884 vs. 0.798: Z=4.137, P=0.018). The 10-fold cross-validation of the nomogram model for 100 times showed that the model had good stability. The decision curve showed that the net benefit curve of the model is located above the two extreme curves of all measures and no measure, indicating that the model had clinical practical value in this range.

The risk of nosocomial infection in cancer patients undergoing chemotherapy is high. This study based on factors such as age, whether diabetes was present, NLR before chemotherapy, whether there was nutritional risk before chemotherapy and whether invasive procedures were performed, constructed a nomogram model which has a good predictive ability for the risk of hospital-acquired infections in cancer patients undergoing inpatient chemotherapy, can be used as risk assessment tool for medical staff to identify patients with nosocomial infection.