Abstract Body: Background: Up to 5% of patients with infective endocarditis (IE) develop intracranial mycotic aneurysms (MA) with nearly 25% of these patients having multiple intracranial aneurysms. Mortality rates have been reported up to 30% for unruptured and 80% for unruptured MA. Therefore, accurate diagnosis of intracerebral aneurysms is critical. CT angiography (CTA) is widely used to detect mycotic aneurysms (MA), yet data are sparse in the literature regarding diagnostic accuracy of CTA compared with the gold standard of 3D rotational digital subtraction angiography (DSA). In this study, we compare CTA to DSA particularly focusing on how the size of aneurysms may change sensitivity of CTA. Based on these results, we offer a practical strategy for the detection and monitoring of MAs.

Methods: This retrospective chart review included all patients admitted to Boston Medical Center between January 1, 2014 and December 31, 2023. 883 patients with IE were identified by ICD-9 and ICD-10 code. 77 intracranial aneurysms were identified by DSA with accompanying CTA within 1 month. An experienced neuroradiologist reviewed each CTA and DSA. Using DSA as the gold standard, sensitivity, specificity, positive and negative predictive values, and likelihood ratios were calculated. To avoid bias and increase clinical relevance, these calculations were based on the final clinical interpretation at the time of the study. For the analysis based on aneurysm size, the review of our study neuroradiologist was used.

Results: Of the 77 aneurysms identified by CTA or DSA, test results were as follows: 53 true negative (CTA -, DSA-), 13 true positive (CTA +, DSA +), 5 false positive (CTA +, DSA -), and 8 false negative (CTA -, DSA +). Including all aneurysms, the sensitivity of CTA was (62%), and the specificity was (91%). The positive and negative predictive values were 72% and 87%, respectively. The sensitivity of CTA for aneurysms ≥ 5mm was 100%. For aneurysms ≥ 4, the sensitivity was 88%. For aneurysms ≤ 3mm, the sensitivity of CTA was 46%.

Conclusions: Sensitivity of CTA for intracranial aneurysms is low, however, it varies depending on size and is excellent for aneurysms ≥ 5mm. Based on this information, we present a strategy that uses a combination of DSA and CTA to detect and monitor MAs.