DRAFT Changes between Protocol and review

We stated we would contact authors and industry and this has not yet been done. This is now stated in the 'Methods' section.

We limited the accepted reference standard to EORTC(‐like) criteria. We originally stated it would be either autopsy, combined with a positive culture or with histopathological evidence, or the EORTC/MSG criteria, or the demonstration of hyphal invasion in biopsies, combined with a positive culture. The rationale for this is that autopsy is almost never done and that biopsy and culture are included in the EORTC/MSG criteria.

Items 2, 8 and 9 of the QUADAS‐checklist, as mentioned in the protocol and in the 'Methods' section, were reported in the 'Characteristics of included studies' table and not in the 'Assessment of methodological quality' table.

We did not calculate Likelihood Ratios and Odds Ratios, as described in the protocol, the reason for this is that we think that the value of this test is better described by explaining the consequences of false positive (1‐specificity) and false negative (1‐sensitivity) results.

Some extra explanation about independence of index and reference tests was added to the Methodology section (under ‘Assessment of Methodological Quality’).

In the protocol we stated that we would investigate the effect of: cut‐off values, reference standard, distinctive groups of patients, children versus adults, and use of antifungal therapy. In the review we did investigate the effects of: cut‐off values, reference standard and clinical subgroups (children versus adults; distinctive groups of patients (high risk versus low risk); use of antifungal prophylaxis; use of antifungal therapy).

In the protocol we stated that the main purpose for a test for invasive aspergillosis would be to guide therapy. During the review process, we discovered that the test is used in many different ways and in most studies it is not used to guide therapy (although a test that could guide therapy would still be ideal). We have therefore changed the text in such a way that there is less focus on guidance of therapy.