Quantitative SPECT with In-111: Comparison of collimators with different resolution and sensitivity

Abstract

1757

Objectives: Development of SPECT-CT and incorporation of various corrections into image reconstruction and analysis methods have made it possible to achieve accurate activity concentration measurements with SPECT. To optimize our use of SPECT to measure the biodistribution of In-111 labeled monoclonal antibodies (MAbs), we compared activity quantitation from images of an anthropomorphic phantom obtained with two available types of medium-energy collimators. Methods: A human torso phantom was configured with simulated lungs, liver, left kidney and tumors (5, internal diameter 1.0 to 3.3 cm) and filled with known In-111 concentrations approximating the biodistribution of radiolabel 48 h after injection of 185 MBq of an In-111 anti-CEA MAb. Relative activity concentrations were as follows: liver 1, lungs 0.08, kidney 0.3, tumors 0.5, background 0.03. SPECT-CT scans (dual 20%-wide energy windows) were obtained with each set of collimators [measured system FWHM at 0 cm = 4.4 (Set 1) or 5.4 (Set 2) mm; nominal septal penetration 0.14% (Set 1) or 1.2% (Set 2)] using a GE Hawkeye^TM tomograph (120 views, 30 s/view). True-count sensitivities for In-111 as determined by Monte Carlo simulation were 88 and 126 cps/kBq for Sets 1 and 2, respectively. Images were reconstructed using the OS-EM algorithm (30 iterations, 24 subsets) including corrections for attenuation, scatter and collimator-detector response. A separately-obtained high-resolution (2.5 mm slice thickness) CT scan was coregistered with the SPECT image sets and used to define regions of interest and measure object volumes for SPECT quantitation. Results: Overall, errors in SPECT-derived activity concentration measurements were smaller for collimator Set 1 than Set 2. {% Error [100(SPECT/true-1)] Set1/Set 2: liver -3.6%/8.6%; lungs 2.7%/12%; kidney -7.7%/-3.6%; 1.0 cm tumor (axilla) -5.9%/-20%; 1.5 cm tumor (abdomen near liver) 21%/36%; 2.1 cm tumor (left lung) -7.8%/2.7%; 3.3 cm tumor (abdomen) -7.1%/8.0%; 3.2 cm tumor (liver) 36%/45%}. Conclusions: Collimator-detector response compensation notwithstanding, benefits to SPECT quantitation from the 40% higher sensitivity of collimator Set 2 were more than offset by its 20% worse spatial resolution and higher septal penetration compared with collimator Set 1.