卒中相关性肺炎患者神经功能恶化的影像指标及其预测价值分析

目的 探讨卒中相关性肺炎(stroke-associated pneumonia,SAP)患者发生神经功能恶化(neurological deterioration,ND)的影像标志物,并对其预测价值进行分析.方法 回顾性收集2020年1月至12月急诊或神经内、外科收治的SAP患者,根据美国国立卫生研究院卒中量表(Na...

Prediction on neurological deterioration based on imaging features in patients with stroke-associated pneumonia

Objective To investigate the risk factors for the occurrence of neurological deterioration (ND) in patients with stroke-associated pneumonia (SAP) and to establish a predictive model using imaging features. Methods We retrospectively recruited SAP patients admitted to emergency or neurology departments during January to December 2020, and SAP patients were divided into ND and non-ND groups according to whether the National Institutes of Health Stroke Scale (NIHSS) score increased by ≥ 4 (total) during hospitalization. The clinical characteristics, laboratory tests, brain and lung imaging features of the two groups were compared. Variables with a P value<0.05 by univariate analysis were included in Logistic regression analysis. Results We retrospectively recruited 73 SAP patients. ND was diagnosed in 15 patients (20.54%). The chest computed tomography (CT) score, large infarction, thalamic infarction and occipital lobe infarction were significantly different between the two groups (P<0. 05). In particular, chest CT score (OR=1.218,95%CI:1.010-1.469), thalamic infarction (P<0.05, OR=10.016, 95%CI:1.523-65.862) and large cerebral infarction (P<0.05, OR=9.033, 95%CI:1.746-46.742) were independent risk factors for the occurrence of ND. For prediction of ND, the area under curve (AUC) of the logistic model was 0.849 (95% CI:0.752-0.947); the AUC of chest CT score was 0.744 (95% CI:0.611-0.878) with sensitivity 80.00%, specificity 63.80%, predictive positive value (PPV) 36.36%,predictive negative value (PNV) 92.50%, accuracy 67.12%; the AUC of thalamic infarction was 0.624 (95% CI :0.450-0.797), with sensitivity 33.33%, specificity 91.38%, PPV 50.00%, PNV 84.13%, accuracy 79.45%; and the AUC of combined large infarction and chest CT score was 0.799 (95%CI:0.690-0.909) with sensitivity 100.00%, specificity 48.28%, PPV 33.33%, PNV 100.00%, accuracy 58.9%. Conclusion Our study indicated that chest CT score, thalamic infarction and large cerebral infarction may be independently associated with ND in SAP. The chest CT score may predict ND in SAP patients with higher sensitivity and accuracy, what's more, the higher the chest CT score (>4.5), the more likely to develop ND in SAP patients. Thalamic infarction may predict ND in SAP patients with higher specificity and accuracy. The combined analysis of chest CT score and lesion size of infarction can help clinicians to stratify the SAP patients with higher risk to develop ND.