放射性皮炎风险预测模型的研究进展

Research Progress on Risk Prediction Models for Radiodermatitis

摘要: 放射性皮炎是放疗患者常见的并发症之一，不仅会引起皮肤瘙痒、疼痛、脱屑甚至溃疡等不适，影响患者生活质量，严重者还会导致放疗中断，从而影响肿瘤的治疗效果和长期预后。目前放射性皮炎尚无特效治疗药物，临床管理以预防为主，因此早期识别高危人群并实施个体化预防具有重要意义。近年来，关于放射性皮炎风险预测模型的研究逐年增加，但各模型在模型构建方法、预测因子选择、结局定义、验证方式等方面存在较大异质性，限制了其在临床中的广泛推广和应用。鉴于此，本文从皮炎评估工具、模型构建方法、预测因子类型、模型性能比较四个方面，对现有放射性皮炎风险预测模型的研究进行梳理，不仅总结了各模型的技术特征，还对其在数据依赖性、过拟合风险、可解释性及临床部署成本等方面的优势与局限进行了批判性分析，以期为临床选择合适的预测工具及该领域的后续研究提供参考依据。

Abstract: Radiodermatitis is a common complication in patients undergoing radiotherapy, which not only affects their quality of life but can also lead to treatment interruptions in severe cases, thereby impacting the therapeutic efficacy for tumors. Currently, there is no specific drug treatment for radiodermatitis, and clinical management primarily focuses on prevention. Early identification of high-risk individuals and implementation of personalized preventive measures are of great significance. In recent years, research on predicting the risk of radiodermatitis has increased annually. However, various models are limited in their clinical promotion and application due to factors such as model construction methods, selection of predictors, definition of outcomes, and validation methods. This article not only summarizes the technical characteristics of existing models but also critically analyzes their advantages and limitations regarding data dependency, overfitting risk, interpretability, and clinical deployment costs. This article reviews existing studies on radiodermatitis risk prediction models from four aspects: assessment tools, model construction methods, types of predictors, and model performance comparison, aiming to provide references for clinical practice and future research.