Acoustic detection of intracranial aneurysms: a decision analysis

Abstract

We present a further evaluation of an improved recording method for the acoustic detection of intracranial aneurysms (ADA). A sensor was applied to the patient's eyes. Two measures were derived to summarize the power spectral density function of the sound frequencies that were obtained from each patient: the power median (PM), the median of the power spectral density function, and the mean difference error (MDE), a measure of the difference between the normalized, logarithmically transformed spectra of the patient and a template, the normal spectrum. The capability of these two measures (alone or combined) to discriminate between patients with and without an intracranial aneurysm was tested in a series of 89 patients harbouring a total of 109 aneurysms and 73 controls, using multiple logistic regression analysis. When PM and MDE were combined, the accuracy of the predictions amounted to 79%. Individualized threshold values of the likelihood ratio of harbouring an aneurysm, for ordering four-vessel angiography are suggested, depending on the prior probability of harbouring an aneurysm, the risks of unnecessary angiography and the risk of living with an undetected aneurysm. Our decision analysis suggests that using these recommendations, employing acoustic detection results in a small gain in quality adjusted life expectancy (0.01 life year) for patients aged between 40 and 60, compared to no diagnostic testing, and 0.02 life year compared to angiography, which cannot be recommended. For patients with a three times increased prior risk of harbouring an intracranial aneurysm, the benefit of ADA compared to angiography increases to 0.05 life year. We conclude that acoustic detection has the potential of becoming a useful tool in the non-invasive diagnosis of occult, asymptomatic intracranial aneurysms.