AI-ROSE is an innovative immunofluorescence staining combined with artificial intelligence image analysis technology that uses a fully automated fluorescence microscope to image pathogens in real time. AI algorithms automatically identify pathogen types (such as bacteria, fungi, etc.) and cellular backgrounds, quickly interpret results, and automatically issue color graphic reports for clinical doctors to take earlier and more accurate targeted treatment for critically ill patients. This study used bronchoalveolar lavage fluid as a biological sample and compared it with traditional examination methods to analyze the diagnostic accuracy and clinical practicality of AI-ROSE in patients with severe pneumonia.

Inclusion Criteria:

• Age 18 years and above
• The preliminary clinical diagnosis is a patient with lower respiratory tract infection
• Meets the criteria for severe pneumonia, which means one of the following conditions exists: ① Main criteria: septic shock requiring vasoactive drug support; Respiratory failure requiring mechanical ventilation; ② Secondary criteria: Meet at least 3 or more criteria, including respiratory rate>30 times/minute; Oxygenation index (PaO ₂/FiO ₂)<250; Multiple leaf segment infiltration; Consciousness disorders/orientation disorders; Urea nitrogen level>20mg/dL; White blood cell count<4 * 10 ^ 9/L; platelet count<100 * 10 ^ 9/L; core body temperature<36.8 ° C; hypotension requiring active fluid resuscitation
• The patient agrees to undergo bronchoscopy and bronchoalveolar lavage, and agrees to perform AI-ROSE testing on the bronchoalveolar lavage fluid

Exclusion Criteria:

• Patients with poor basic condition, severe illness, and inability to tolerate bronchoscopy examination
• The bronchoalveolar lavage fluid sample is not qualified

Infected severe pneumonia patients