<p>Radiomic features dataset.</p>

Background
Accurate prediction of radiation-induced oral mucositis is crucial for personalized treatment in head and neck cancer. However, developing robust predictive models utilizing high-dimensional multimodal data (CT imaging, dose distribution, and clinical features) remains challenging, particularly in cohorts with limited sample sizes.
Objective
This study aimed to rigorously evaluate and compare the multi-class predictive performance of traditional machine learning algorithms and deep learning architectures under a small-cohort setting.
Methods
Multimodal data from 108 patients were collected. A comprehensive evaluation framework incorporating nine traditional machine learning algorithms and two deep learning models (a dimensionality-reduced 1D-CNN and a multimodal 3D-CNN) was established. To ensure robust evaluation, a stratified 5-fold cross-validation was employed. Model performance was comprehensively quantified using mean ± standard deviation (SD) across multiple metrics, including the Area Under the Curve (AUC), accuracy, and Matthews Correlation Coefficient (MCC).
Results
Inter-rater reliability for RIOM grading was excellent (Cohen’s kappa = 0.82, 95% CI: 0.73–0.91). Among traditional machine learning approaches, the Extra Trees (ET) algorithm achieved the highest discriminative capacity (AUC: 0.956 ± 0.046), while Logistic Regression (LR) demonstrated optimal overall accuracy (0.832 ± 0.155) and stability. Regarding deep learning, the lightweight 1D-CNN utilizing fused low-dimensional features exhibited highly competitive and robust performance (AUC: 0.900 ± 0.072; Accuracy: 0.732 ± 0.140). In stark contrast, the high-dimensional multimodal 3D-CNN suffered from severe overfitting and mode collapse phenomenon, yielding significantly inferior results (AUC: 0.568 ± 0.090; MCC: −0.025 ± 0.031).
Conclusions
For small-cohort radiomics and dosimetric analyses, ensemble learning models (e.g., ET) and appropriately regularized linear models (e.g., LR) remain highly effective. While deep learning holds promise, high-dimensional architectures like 3D-CNNs are highly susceptible to mode collapse without massive datasets. Instead, employing feature dimensionality reduction combined with lightweight networks (1D-CNN) is a vastly superior strategy to extract reliable predictive patterns from limited clinical data.

Background
准确预测放射性口腔黏膜炎（radiation-induced oral mucositis, RIOM）对头颈部癌的个性化治疗至关重要。然而，利用高维多模态数据（CT成像、剂量分布及临床特征）构建稳健的预测模型仍面临诸多挑战，尤其在样本量有限的队列研究中。
Objective
本研究旨在严格评估并对比小样本队列场景下，传统机器学习算法与深度学习架构的多分类预测性能。
Methods
本研究收集了108例患者的多模态数据。搭建了包含9种传统机器学习算法与2种深度学习模型（降维一维卷积神经网络（1D-CNN）及多模态三维卷积神经网络（3D-CNN））的全面评估框架。为确保评估结果稳健可靠，采用分层5折交叉验证策略。通过曲线下面积（Area Under the Curve, AUC）、准确率及马修斯相关系数（Matthews Correlation Coefficient, MCC）等多项指标，以均值±标准差（standard deviation, SD）的形式全面量化模型性能。
Results
RIOM分级的组间信度极佳（科恩kappa值=0.82，95%置信区间：0.73–0.91）。在传统机器学习方法中，极端随机树（Extra Trees, ET）算法的判别能力最优（AUC：0.956±0.046），而逻辑回归（Logistic Regression, LR）则展现出最佳的整体准确率（0.832±0.155）与稳定性。在深度学习模型方面，融合低维特征的轻量化1D-CNN表现出极具竞争力且稳健的性能（AUC：0.900±0.072；准确率：0.732±0.140）。与之形成鲜明对比的是，高维多模态3D-CNN出现了严重的过拟合与模式崩溃现象，所得结果显著劣于其余模型（AUC：0.568±0.090；MCC：-0.025±0.031）。
Conclusions
针对小样本队列的放射组学与剂量学分析，集成学习模型（如ET）与经过适当正则化的线性模型（如LR）依然极具有效性。尽管深度学习颇具应用前景，但高维架构（如3D-CNN）在缺乏大规模数据集的情况下极易发生模式崩溃。相较而言，将特征降维与轻量化网络（1D-CNN）相结合的策略，从有限的临床数据中提取可靠预测模式的效果更为优异。