Relative increase in leukemia-specific DNA in peripheral blood plasma from patients with acute myeloid leukemia and myelodysplasia. Blood, 103, 2799Blood, 103, -2801 Keywords: polycythemia vera, idiopathic myelofibrosis, essential thrombocythemia, JAK2, mutations doi:10.1111/j. 1365-2141.2006.06174.x Incidence of invasive fungal infection in adult haematological malignancy: a prospective validation of a risk stratification scheme Invasive fungal infections (IFI) are important infectious complications in haemato-oncology patients, leading to direct high morbidity and mortality. The diagnosis of IFI is frequently established postmortem because conventional diagnostic methods are insensitive and the definition of proven mould infection usually requires histological examination and cultures of deep tissues, which often are impossible to carry out due to the site of the infection and the critical condition of these patients (Ascioglu et al, 2002). In a seminal review paper, Prentice et al (2000) proposed a risk group stratification of IFI in neutropenic patients based on a review of published data, and proposed the assignment of patients to high, intermediate (high and low) and low risk of developing IFI.We herein report a prospective validation study of the incidence of IFI, using, as a diagnostic tool, the detection of galactomannan (GM) (Platelia Aspergillus; Bio-Rad, Marnes La Coquette, France) on a twice-weekly screening protocol during the period of IFI risk. An index of ‡0AE5 was considered to be a positive result if it was confirmed with a subsequent sample. Semiquantitative surveillance cultures for yeast were performed weekly and conventional microbiological data were recorded. In cases of clinical suspicion of IFI or when the GM index was above 0AE5 (Maertens et al, 2004), or in cases of pulmonary infiltrates a diagnostic work-up was started; this included thoracic high resolution computed tomography scanning (HRCT), followed, when possible, by bronchoalveolar lavage and/or biopsy for bacterial, fungal and viral cultures and Legionella antigen in urine. Only proven and probable IFI were considered for this validation study, following the definitions of Ascioglu et al (2002). Local ethical committee approval was granted.From September 2004 to July 2005, 78 adult patients (90 episodes) receiving cytotoxic chemotherapy and stem cell transplantation (SCT) were prospectively studied (Table I). The sensitivity, specificity, positive and negative predictive values of GM screening were 100%, 97%, 78% and 100% respectively. We determined the relationship between the incidence of IFI (per patient) and risk group. Evidence of IFI was documented in 10 patients (13%) (six proven and four probable). There were no possible IFI cases in our population. The incidence of IFI correlated directly and significantly (v 2 P ¼ 0AE0229) with risk stratification group, with the highest proportion in the high-risk group and the percentage falling directly and significantly as the risk declined. The incidence in the low-risk gro...