Assessment of splenic switch-off with arterial spin labeling in adenosine perfusion cardiac MRI

Abstract

Background: In patients with suspected coronary artery disease (CAD), myocardial perfusion is assessed under rest and pharmacological stress to identify ischemia. Splenic switch-off, defined as the stress to rest splenic perfusion attenuation in response to adenosine, has been proposed as an indicator of stress adequacy. Its occurrence has been previously assessed in first-pass perfusion images, but the use of noncontrast techniques would be highly beneficial. Purpose: To explore the ability of pseudo-continuous arterial spin labeling (PCASL) to identify splenic switch-off in patients with suspected CAD. Study Type: Prospective. Population: Five healthy volunteers (age 24.8 ± 3.8 years) and 32 patients (age 66.4 ± 8.2 years) with suspected CAD. Field strength/Sequence: A 1.5-T/PCASL (spin-echo) and first-pass imaging (gradient-echo). Assessment: In healthy subjects, multi-delay PCASL data (500–2000 msec) were acquired to quantify splenic blood flow (SBF) and determine the adequate postlabeling delay (PLD) for single-delay acquisitions (PLD > arterial transit time). In patients, single-delay PCASL (1200 msec) and first-pass perfusion images were acquired under rest and adenosine conditions. PCASL data were used to compute SBF maps and SBF stress-to-rest ratios. Three observers classified patients into “switch-off” and “failed switch-off” groups by visually comparing rest-stress perfusion data acquired with PCASL and first-pass, independently. First-pass categories were used as reference to evaluate the accuracy of quantitative classification. Statistical Tests: Wilcoxon signed-rank, Pearson correlation, kappa, percentage agreement, Generalized Linear Mixed Model, Mann–Whitney, Pearson Chi-squared, receiver operating characteristic, area-under-the-curve (AUC) and confusion matrix. Significance: P value < 0.05. Results: A total of 27 patients (84.4%) experienced splenic switch-off according to first-pass categories. Comparison of PCASL-derived SBF maps during stress and rest allowed assessment of splenic switch-off, reflected in a reduction of SBF values during stress. SBF stress-to-rest ratios showed a 97% accuracy (sensitivity = 80%, specificity = 100%, AUC = 85.2%). Data Conclusion: This study could demonstrate the feasibility of PCASL to identify splenic switch-off during adenosine perfusion MRI, both by qualitative and quantitative assessments.