Comparison of Sensitivity of Low-Dose CT Scan and Serum Galactomannan in Patients with Hematologic Malignancies and Positive Bronchoalveolar Lavage for Invasive Pulmonary Aspergillosis

Invasive pulmonary aspergillosis (IPA) is associated with considerable morbidity and mortality in patients with acute leukemia and those undergoing hematopoietic stem cell transplantation (HSCT). Timely diagnosis of IPA is crucial, but difficult. Strategies for early detection of IPA include serum galactomannan (GM), bronchoalveolar lavage (BAL) GM and characteristic radiologic abnormalities on CT scan; however, the relative sensitivity of these diagnostic tests is unclear. We sought to determine the utility of serum GM and characteristic CT radiologic abnormalities in diagnosing IPA in patients with hematologic malignancies who had a positive BAL for Aspergillus.

We performed a single center retrospective cohort study from 2010 to 2012 to determine the sensitivity serum GM (OD > 0.5) and low dose CT scan in patients with hematologic malignancies. All positive BAL GM samples and Aspergillus isolates from BAL fluid cultures from patients with acute leukemia and those undergoing HSCT were reviewed; all patients had abnormal chest CT scans consistent with pulmonary infection. IPA was classified as proven, probable or possible according to EORTC/MSG criteria. Recent low dose CT scans correlating to the date of the positive BAL GM or Aspergillus isolation were reviewed and graded: halo signs, cavities, crescents and nodules were deemed to be consistent with IPA, while other changes (e.g. ground-glass changes, consolidation) were considered non-specific.

A total of 49 BAL samples were included; of these, 31 were considered probable IPA and 18 possible IPA (on the basis of non-specific radiologic findings). Most patients had either no prior or 1–2 days of mold-active antifungal agents. There were 43 cases with positive BAL GM and 11 cases of positive Aspergillus BAL isolates; 5 patients had both. Of the positive BAL GM cases, 27 (63%) were associated with radiologic findings consistent with IPA. The remaining cases were associated with non-specific radiologic findings. Of the patients with Aspergillus isolates on BAL, 55% (6/11) had radiologic features consistent with IPA, while the remaining cases had non-specific radiologic findings.

We next evaluated the sensitivity of serum GM. Of 34 patients with BAL GM positivity who had concomitant serum GM testing, 4 (12%) had positive serum GM. Of 8 patients with Aspergillus species isolated on BAL who had serum GM performed, only 1 (12.5%) had positive serum GM. In contrast, 5/8 patients (63%) with Aspergillus isolates on BAL had a positive BAL GM. The combined sensitivity of serum GM was 12% (5/42).

Although the majority of patients with positive BAL for Aspergillus (+GM and/or isolates) had characteristic radiologic findings on CT scan, the absence of such findings did not exclude this diagnosis, as over one third had only non-specific radiologic findings. Serum GM had very low sensitivity in this population and should not be used in isolation to diagnose IPA.

No relevant conflicts of interest to declare.