The identification of the causative organism in invasive pulmonary aspergillosis (IPA) is recommended. We investigated whether a mycologic diagnostic strategy could be optimized based on patient characteristics. Fifty-five patients were enrolled in a prospective study. The presence of Aspergillus in respiratory samples occurred more frequently in non-acute leukemia (AL) patients than in AL patients (P = .0003), and in patients with leukocyte counts more than 100/mm(3) (P = .002). In a logistic regression model, these 2 factors appeared to be independent, with an adjusted odds ratio of 7.14 (95% confidence interval, 1.40-36.5) for non-AL patients and an adjusted odds ratio of 6.97 (95% confidence interval, 1.33-36.5) for patients with leukocyte counts more than 100/mm(3). A positive mycologic result was also more frequent among patients with lung CT scan signs of airway-invasive disease than among other patients (P = .043). Airway-invasive signs were more frequent among non-AL patients (P = .049), whereas angioinvasive disease was more frequent among both AL patients (P = .01) and patients with leukocyte counts less than 100/mm(3) (P = .001). A concomitant pulmonary infection was identified more frequently among non-AL patients (P = .005 vs allogeneic hematopoietic stem cell transplant and P = .048 vs others). Our results suggest that different strategies for diagnosing IPA should be considered based on the underlying condition.