Objective: To explore the diagnostic yield of bronchoscopic rapid on-site evaluation (B-ROSE) in patients with severe invasive bronchopulmonary aspergillosis (IBPA) and provide evidence for starting antifungal treatment before microbiological results were available.
Methods: A prospective cohort study was conducted to select patients with severe pneumonia suspected of IBPA admitted to the respiratory intensive care unit (RICU) in the First Affiliated Hospital of Xinjiang Medical University from June 2014 to June 2022, and those who were primarily infected with other pathogens (such as bacteria, Mycobacterium tuberculosis) at admission were excluded. Whether the antifungal treatment was initiated or not on the basis of the bedside B-ROSE, the B-ROSE was administered as soon as possible within 24 hours after admission to RICU. The current international definition of invasive aspergillosis was used as the gold diagnostic standard, the diagnostic accordance rate, the sensitivity and specificity of B-ROSE were calculated respectively, and the receiver operator characteristic curve (ROC curve) was also plotted, to evaluate the predictive value in diagnosing IBPA.
Results: A total of 176 patients with severe pneumonia suspected of IBPA were included in the study. According to international diagnostic standards, there were 81 cases of IBPA and 95 cases of non-IBPA. According to the early diagnosis of B-ROSE, there were 89 cases of IBPA and 87 cases of non-IBPA. The diagnostic accordance rate of B-ROSE was 84.09% (148/176), the area under the ROC curve for B-ROSE in diagnosing severe IBPA was 0.844, the 95% confidence interval (95%CI) was 0.782-0.905, the sensitivity was 87.65%, the specificity was 81.05%, the positive predictive value was 79.78%, the negative predictive value was 88.51%, the rate of underdiagnosis was 12.35% (10/81), and the rate of misdiagnosis was 18.95% (18/95). Compared with the true negative group, the proportion of long-term (≥ 14 days) use of glucocorticoid [70.0% (7/10) vs. 9.1% (7/77), P < 0.01] and the proportion of cases with diabetes [40.0% (4/10) vs. 10.4% (8/77), P < 0.05] were significantly higher in the false negative group (underdiagnosis group). However, B-ROSE of both groups showed mucosal bleeding, congestion and edema [100.0% (10/10) vs. 94.8% (73/77), P > 0.05], indicating that acute mucosal inflammation was non-characteristic. Compared with the true positive group, the proportion of long-term (≥ 14 days) use of glucocorticoid in the false positive group (misdiagnosis group) was significantly reduced [33.3% (6/18) vs. 60.6% (43/71), P < 0.05]. The B-ROSE results showed the proportion of cases with mucosal white spots, black plaques and pseudomembrane was significantly reduced [16.7% (3/18) vs. 52.1% (37/71), P < 0.01] in the misdiagnosed group, which suggest that cases of long-term use of glucocorticoid and cases with B-ROSE showing mucosal white spots, black plaques and pseudomembrane were less likely to be misdiagnosed. The main diseases that were easily misdiagnosed as IBPA included pulmonary tuberculosis (38.9%, 7/18), inflammatory lung adenocarcinoma (27.8%, 5/18) and pulmonary vasculitis (16.7%, 3/18).
Conclusions: Before obtaining microbiological evidence, B-ROSE can assist in decision-making of early anti-aspergillus treatment for severe IBPA. This method is prompt, simple, and has high accuracy and reliability. If B-ROSE lacks characteristic manifestations, especially for severe pneumonia in patients with long-term use of glucocorticoid or diabetes, attention should be paid to the underdiagnosis of IBPA. Diseases such as lung tuberculosis, inflammatory lung adenocarcinoma and lung vasculitis should be vigilant against misdiagnosis as IBPA.