Serial assessment of pulmonary lesion volume by computed tomography allows survival prediction in invasive pulmonary aspergillosis.

ABSTRACT

BACKGROUND: Serial chest CT is the standard of care to establish treatment success in invasive pulmonary aspergillosis (IPA). Data are lacking how response should be defined. METHODS: Digital CT images from a clinical trial on treatment of IPA were re-evaluated and compared with available biomarkers. Total volume of pneumonia was added up after manual measurement of each lesion, followed by statistical analysis. RESULTS: One-hundred and ninety CT scans and 309 follow-up datasets from 40 patients were available for analysis. Thirty-one were neutropenic. Baseline galactomannan (OR 4.06, 95%CI 1.08-15.31) and lesion volume (OR 3.14, 95%CI 0.73-13.52) were predictive of death. Lesion volume at d7 and trend between d7 and d14 were strong predictors of death (OR 20.01, 95%CI 1.42-282.00 and OR 15.97, 95%CI 1.62-157.32) and treatment being rated as unsuccessful (OR 4.75, 95%CI 0.94-24.05 and OR 40.69, 95%CI 2.55-649.03), which was confirmed by a Cox proportional hazards model using time-dependent covariates. CONCLUSION: Any increase in CT lesion volume between day 7 and day 14 was a sensitive marker of a lethal outcome (>50%), supporting a CT rescan each one and 2 weeks after initial detection of IPA. The predictive value exceeded all other biomarkers. Further CT follow-up after response at day 14 was of low additional value. KEY POINTS • CT evaluation offers good prediction of outcome for invasive pulmonary aspergillosis. • Predictive capability exceeds galactomannan, blood counts, and lesion count. • Any progression between day 7 and day 14 constitutes a high-risk scenario.