Coronavirus disease 2019 (COVID-19) has had a global social and economic impact. An easy assessment procedure to handily identify the mortality risk of inpatients is urgently needed in clinical practice. Therefore, the aim of this study was to develop a simple nomogram model to categorize patients who might have a poor short-term outcome. A retrospective cohort study of 189 COVID-19 patients was performed at Shanghai Ren Ji Hospital from December 12, 2022 to February 28, 2023. Chest radiography and biomarkers, including KL-6 were assessed. Risk factors of 28-day mortality were selected by a Cox regression model. A nomogram was developed based on selected variables by SMOTE strategy. The predictive performance of the derived nomogram was evaluated by calibration curve. In total, 173 patients were enrolled in this study. The 28-day mortality event occurred in 41 inpatients (23.7%). Serum KL-6 and radiological severity grade (RSG) were selected as the final risk factors. A nomogram model was developed based on KL-6 and RSG. The calibration curve suggested that the nomogram model might have potential clinical value. The AUCs for serum KL-6, RSG, and the combined score in the development group and validation group were 0.885 (95% CI: 0.804–0.952), 0.818 (95% CI: 0.711–0.899), 0.868 (95% CI: 0.776–0.942) and 0.932 (95% CI: 0.862–0.997), respectively. Our results suggested that the nomogram based on KL-6 and RSG might be a potential method for evaluating 28-day mortality in COVID-19 patients. A high combined score might indicate a poor outcome in COVID-19 patients with pneumonia.

中文翻译：

---

使用血清 KL-6 和胸部影像学严重程度预测 COVID-19 肺炎患者的 28 天死亡率：一项回顾性队列研究

2019 年冠状病毒病（COVID-19）对全球社会和经济产生了影响。临床实践中迫切需要一种简单的评估程序来轻松识别住院患者的死亡风险。因此，本研究的目的是开发一个简单的列线图模型来对短期结果可能较差的患者进行分类。 2022年12月12日至2023年2月28日，上海仁济医院对189名COVID-19患者进行了回顾性队列研究。对胸部X光检查和包括KL-6在内的生物标志物进行了评估。通过 Cox 回归模型选择 28 天死亡率的危险因素。根据 SMOTE 策略选定的变量开发列线图。通过校准曲线评估导出列线图的预测性能。总共有 173 名患者参加了这项研究。 41 名住院患者（23.7%）发生了 28 天死亡事件。选择血清KL-6和放射学严重程度等级（RSG）作为最终危险因素。基于KL-6和RSG开发了列线图模型。校准曲线表明列线图模型可能具有潜在的临床价值。开发组和验证组中血清KL-6、RSG和综合评分的AUC分别为0.885（95% CI：0.804-0.952）、0.818（95% CI：0.711-0.899）、0.868（95% CI：分别为 0.776–0.942) 和 0.932 (95% CI: 0.862–0.997)。我们的结果表明，基于 KL-6 和 RSG 的列线图可能是评估 COVID-19 患者 28 天死亡率的潜在方法。综合评分高可能表明患有肺炎的 COVID-19 患者预后不佳。