Computer Program Distinguishes Invasive Aspergillosis From Aspergillus Colonisation

Ghent University Hospital
All of use breathe in the fungus Aspergillus every day as it is plentiful in the air around us, indoor or outdoors. Many people who are vulnerable to infection will be unable to clear their lungs of inhaled Aspergillus spores in the normal way most of us do and instead it will start to grow where it lands in their airways. These people tend to be the critically ill who have major health problems other than Aspergillus infection and are generally resident in hospital for the duration of their illness e.g. people undergoing treatment for some cancers.

At this stage there is limited threat to the health of the patient as the fungus tends to simply grow on the surface of the larger airways and cause no further problems, possibly even clearing up after the patient has sufficiently recovered their immune function. A few patients will not be so lucky and the infection will progress to a fully invasive aspergillosis which needs urgent treatment with antifungal drugs.

There is some suggestion that patients who are at risk should receive antifungal drugs in order to prevent infection - colonisation or invasive - but that is wasteful and expensive and may needlessly expose patients to the many side effects of antifungal drugs, some of which are quite severe. We need a way to tell the difference between patients who have colonised with Aspergillus and those who have a serious invasive infection, and a research group in Belgium (Blot 2012) has attempted to provide an answer.

The researchers have developed a clinical algorithm using a number of diagnostic criteria using a group of 115 patients with proven aspergillosis with the aim of distinguishing between patients who are colonised and those with invasive infection. The algorithm was tested on a group of more than 500 patients and found to be able to identify which group a patient with an Aspergillus infection fell into with 92% accuracy. Those without an Aspergillus infection had a 62% probability of being correctly identified. In those people where an Aspergillus infection is known to be present this is claimed to be clinically useful - perhaps enabling us to judge far more efficiently which patients need early intervention with antifungal drugs. Given that all time saved will help improve the outcome for the patients this finding could well help us improve the numbers of patients treated successfully.