Aim: Presence and consequent extent of infection in patients on continuous-flow left ventricular assist devices (CF-LVADs) can be challenging with the current diagnostic tools. The present study sought to demonstrate the diagnostic power of 18F-Fluorodeoxyglucose-Positron-Emission Tomography/Computed Tomography (18F-FDG PET/CT) in detecting infection in patients supported with CF-LVAD.
Background: The present study sought to demonstrate the diagnostic power of 18F-fluorodeoxyglucose-positron-emission tomography/computed tomography (18F-FDG PET/CT) in detecting infection in patients supported with CF-LVAD.
Methods and results: Between July 2009 and April 2016, 61 PET/CT examinations were performed in 47 patients (median age 64.13 years, IQR 18.77) supported with a CF-LVAD. PET/CT assessments were performed qualitatively and quantitatively at three different levels: at the piercing site of driveline (first level), along the intracorporeal course of driveline (second level), and around the device (third level). Final diagnosis of LVAD infection was prospectively performed and was based upon microbiological samples taken at hospital admission, during the surgical revision/transplantation and recurrence of symptoms on long-term follow-up. At last follow-up a total of 40 (65.57%) final diagnoses of LVAD-infection could be ascertained. Matching the final diagnosis with the PET/CT assessments the sensitivity, specificity, and positive and negative predictive value were 90.0, 71.4, 85.71, and 78.94%, respectively. Level sub-analyses of SUV max showed an optimal discriminator power for levels 1 and 2 (AUC of level 1-0.824, P < 0.001; AUC of level 2-0.849, P < 0.001, respectively). At the third level semi-quantitative analysis showed poor discriminator power (AUC 0.589, P = 0.33). Qualitative visual analysis instead indicated a trend toward significance (P = 0.07).
Conclusions: Quantitative 18F-FDG PET/CT is an optimal diagnostic tool in detecting superficial and deep driveline infections. However, diagnostic accuracy with regard to the diagnosis of pump housing infection is limited. Here, clinical and qualitative PET/CT analyses must be better considered.