Abstract: Objective: We systematically evaluate the risk prediction models for Chemotherapy-Induced Oral Mucositis in order to provide reference for clinical staff to select appropriate prediction model or develop a new one. Methods: The studies about risk prediction models for Chemotherapy-Induced Oral Mucositis were searched. Two researchers independently screened the literature, extracted the data, and evaluated the bias risk as well as applicability of the included study. Results: A total of 8 studies including 12 models were identified. Among them, 11 models reported the area under the ROC curve which was 0.61-0.97, and 5 models reported the calibration. Only one model was externally validated. Predictors reported repeatedly of the models were age, chemo-therapy regimens, times of chemotherapy, body mass index, radiation dose, oral hygiene and smoking history. All included studies showed good applicability, but had high risk of bias. Conclusion: The risk prediction models for Chemotherapy-Induced Oral Mucositis have good discrimination and applicability, but the risk of bias is high. And the predictive factors of each model are different. In the future, it is necessary to adjust the existing models or develop more targeted, low bias risk prediction models according to different tumor types under the guidance of methodological statements, and conduct internal and external validation.

Key words: chemotherapy; stomatitis; prediction; nursing care; systematic review