Improved SPECT quantification of 111-In distribution by compensation for collimator geometry effects

1515 Objectives To determine whether proprietary compensation for collimator geometry ("resolution recovery" (RR)) improves the quantification of In-111 uptake by SPECT imaging. Methods Resolution of In-111 SPECT was measured using three line sources of (each 10MBq In-111) in air. SPECT data was acquired in 60 projections over 360° using a GE Discovery 670 SPECT-CT system fitted with a MEGP collimator. Data was acquired using two peak windows (171keV±10% and 245keV±10%) and one scatter window (140keV±9%). Recovery Coefficient was measured for a 16.5ml (r=16mm) sphere filled with 20MBq In-111, in a cylindrical water filled phantom (r=111mm) using the same acquisition parameters. Transaxial images were reconstructed using OSEM (10 subsets, 6 iterations, no post filtering) using standard CT attenuation and scatter correction; two sets of images were created, one with RR1 and one without. Resolution was estimated from x and y line profiles for each line on three SPECT non-consecutive slices, the profiles were filled to a Gaussian model and the full width half maximum (FWHM) values calculated. Mean FWHM values were calculated with and without RR. Recovery coefficient (RC) was taken as the number of counts seen inside the sphere as a percentage of total counts in the reconstruction. For the counts in the sphere a spherical VOI was defined using the CT image. Results Resolution: The line profiles from the RR data were observed to fit the Gaussian model better than the non-RR data, FWHM without RR was 11.7±1.8mm and 8.1±0.9mm with RR, (p Conclusions The employed resolution recovery technique significantly improves the resolution of In-111 SPECT imaging with a consequent improvement in the Recovery Coefficient. This result is important for the visualisation of small lesions with In-111 labelled compounds and for the prospect of predictive dosimetry with In-111 for molecular radiotherapy with Y-90 labelled agents. Research Support This work is supported by UCL Cancer Institute Research Trust (AJG & JD) and UCLH NHS Foundation Trust (WW)