Minimizing image misregistration during PET/CT-guided percutaneous interventions with monitored breath-hold PET and CT acquisitions

Abstract

Purpose: To validate a monitored, breath-hold positron emission tomography (PET)/computed tomography (CT) acquisition technique for the minimization of respiratory PET/CT image misregistration and lesion distortion during PET/CT-guided percutaneous interventional procedures.

Materials and methods: Eleven patients referred for percutaneous biopsy or thermal ablation of tumors near the diaphragm were prospectively enrolled. Initial PET/CT scanning was performed by using a bellows device and monitored, same-level breath-holds for PET and CT acquisitions. Breath-hold PET consisted of nine 20-second breath-hold frames, yielding a 3-minute equivalent PET dataset. A second PET/CT scan was obtained without monitoring by using end-expiration breath-hold CT and free-breathing PET. PET/CT tumor misregistration and craniocaudal tumor diameter were measured on monitored and unmonitored PET/CT datasets. Data were analyzed by using nonparametric, two-sided, signed-rank statistical tests.

Results: Mean PET/CT image misregistrations in the craniocaudal, anteroposterior, and transverse planes were 2.6 mm (range, 0-7 mm), 3.3 mm (range, 1-8 mm), and 2.7 mm (range, 0-8 mm) with monitoring and 14.7 mm (range, 0-49 mm), 7.6 mm (range, 1-24 mm), and 4.0 mm (range, 0-12 mm) without monitoring, respectively. Differences were significant for craniocaudal (P = .0087) and anteroposterior (P = .014) planes, but not for the transverse plane (P = .23). Mean craniocaudal target diameter was 2.5 mm (range, -2 to 9 mm) larger (ie, distorted) for unmonitored versus monitored PET (P = .061).

Conclusions: Acquiring PET/CT datasets with respiratory bellows-assisted, monitored breath-holds improves PET/CT image registration versus unmonitored PET/CT and may facilitate accurate targeting during PET/CT-guided interventions in anatomic regions subject to respiratory motion.