分子影像学杂志2024，Vol.47Issue(7)：689-694,6.
医院获得性肺炎和呼吸机相关性肺炎的病原菌与影像学特征的关联性
Correlation between pathogens and imaging characteristics of hospital-acquired pneumonia and ventilator-associated pneumonia
摘要
Abstract
Objective To investigate the correlation between the pathogenic distribution of hospital-acquired pneumonia(HAP)and ventilator-associated pneumonia(VAP)with their corresponding chest CT imaging characteristics,aiming to promptly identify the distribution of HAP/VAP pathogens and accurately assess the extent of pulmonary lesions.Methods A retrospective analysis was performed for the culture of airway secretion pathogens and chest CT images of 184 patients with HAP/VAP(154 cases of HAP and 30 cases of VAP)who were diagnosed and treated at the Affiliated Hospital of Shaanxi University of Chinese Medicine from January 2020 to January 2022.Artificial intelligence software was used to segment chest CT images and quantitatively analyze pathogenic bacteria image characteristics.The quantitative indicators included whole lung volume,emphysema quantification,lesion volume,lesion volume percentage,lesion mass,and the percentage of lesion density less than-300 HU in total lesion volume.Results Gram-negative bacilli were the dominant pathogens(92.72%),with Klebsiella pneumoniae,Pseudomonas aeruginosa,Acinetobacter baumannii being the most prevalent.Surgical history/invasive procedures emerged as major risk factors in 37%of the patients.On CT images,lung consolidation was a common funding,while Klebsiella pneumoniae pneumonia showed pleural effusion and ground-glass changes;Pseudomonas aeruginosa pneumonia also shows bronchiectasis and cavitation;Acinetobacter baumannii pneumonia may also see pleural effusion and interstitial changes.Quantitative CT analysis revealed that LAA-950%in patients with Klebsiella pneumoniae pneumonia was low,whereas it was higher in patients with mixed bacterial infection.Patients with Klebsiella pneumoniae pneumonia have the biggest lung volume,whereas those with mixed infections caused by bacteria have the least.The highest proportion of LOA%in pneumonia,combined with a fungal infection(15%),was observed in patients.And LOA%was 10.5%,which could indicate the presence of fungal infection or multiple bacterial infections,with an area under the curve of 0.713(P=0.041),79%sensitivity,and 70%specificity.Conclusion By analyzing the CT imaging features and quantitative parameters of HAP/VAP,especially LOA%,it is helpful to speculate on potential pathogenic bacterial types,which has important value in the diagnosis of combined fungal or multi bacterial infections.关键词
肺炎/感染/CT/定量影像Key words
pneumonia/infection/CT/quantitative imaging引用本文复制引用
辛晓坤,蒲豆豆,江叶海,马光明,于楠..医院获得性肺炎和呼吸机相关性肺炎的病原菌与影像学特征的关联性[J].分子影像学杂志,2024,47(7):689-694,6.基金项目
陕西省自然科学基础研究计划项目(2022JM-453) （2022JM-453）
