Supplementary Material for: Dynamic Prognostic Score to Predict Survival in East Asian Elderly Populations with Dysphagia

Introduction: Dysphagia is a common disorder in older adults leading to severe complications, including aspiration pneumonia, dehydration, malnutrition, weight loss, and even death. However, no prognostic model has been developed to predict the prognosis of older adults with dysphagia. Methods: Data from patients with dysphagia at a single center were retrospectively reviewed between 2014 and 2017. All data were obtained from the Dryad Digital Repository. The least absolute shrinkage and selection operator regression model was used to select potentially relevant features. Multiple Cox proportional hazard analysis was used to develop a model based on the training set. Results: The nomogram comprised age, sex, percutaneous endoscopic gastrostomy, chronic heart failure, total lymphocyte count, daily calorie intake, and severe pneumonia, which provided favorable calibration and discrimination in the training dataset, with AUCs for the 1-year, 2-year, and 3-year survival predictions of 0.833, 0.871, and 0.886, respectively. Furthermore, it showed acceptable discrimination in the validation cohort, with AUCs for the 1-year, 2-year, and 3-year survival predictions of 0.884, 0.834, and 0.782, respectively. Moreover, the decision curve analysis results revealed that the nomogram was clinically beneficial. Conclusion: A nomogram, combining seven demographic and clinical factors, provided an excellent preoperative prediction of survival probability in older individuals with dysphagia. This predictive model can be used as a reference to assist clinicians in making clinical decisions.

引言：吞咽困难（Dysphagia）是老年人群中常见的功能障碍，可引发吸入性肺炎、脱水、营养不良、体重下降甚至死亡等严重并发症。然而目前尚无针对老年吞咽困难患者预后的专用预测模型。
方法：本研究回顾性分析了2014—2017年某单中心收治的吞咽困难患者临床数据，所有数据集均获取自Dryad数字仓储。采用最小绝对收缩和选择算子（Least Absolute Shrinkage and Selection Operator，LASSO）回归模型筛选潜在关联特征，并基于训练集通过多因素Cox比例风险回归分析构建预测模型。
结果：该列线图纳入年龄、性别、经皮内镜下胃造口术（Percutaneous Endoscopic Gastrostomy，PEG）、慢性心力衰竭、总淋巴细胞计数、每日热量摄入及重症肺炎共7项指标，在训练集内展现出优异的校准性能与区分能力；其1年、2年、3年生存预测的受试者工作特征曲线下面积（Area Under the Curve，AUC）分别为0.833、0.871和0.886。在验证队列中，该模型同样表现出可接受的区分度，1年、2年、3年生存预测的AUC分别为0.884、0.834和0.782。此外，决策曲线分析结果证实该列线图具备临床实用价值。
结论：本研究构建的列线图整合7项人口学与临床指标，可精准预测老年吞咽困难患者的术前生存概率，可为临床医师制定诊疗决策提供参考依据。