J Nucl Med. 2007; 48 (Supplement 2):405P
Instrumentation & Data Analysis: Data Analysis & Management
Data Analysis & Management Posters |
Estimation of coefficients to individualize standard arterial input function without blood sampling in iodine-123-IMP autoradiographic method: Application of 2-compartment model analysis of successive two SPECT scans
Noriyuki Shuke1,
Atsutaka Okizaki1,
Junichi Sato1,
Hiroyuki Sugimori1 and
Tamio Aburano1
1 Department of Radiology, Asahikawa Medical College, Asahikawa, Japan
1702
Objectives: It is necessary to obtain an arterial blood sample to individualize the standard arterial input function for cerebral blood flow mapping with iodine-123-IMP (IMP) autoradiographic (ARG) method (J Nucl Med 1994; 35: 2019). As an alternative, individual calibration coefficient (Cf: (kBq/ml)/(CPS/g blood)) could be estimated without arterial blood sampling based on 2-compartment model analysis of dynamic SPECT data (2006 SNM abstract, 47; 362P). To make this method clinically applicable, simplified method for 2-compartment model analysis was proposed. Methods: Twenty-eight patients (60 ± 15 years old, M/F=17/11) were studied with IMP dynamic SPECT. Immediately after starting 1 min infusion of IMP (167 MBq), dynamic SPECT data acquisition (14 rotations/35 min) was performed. One-point arterial blood sampling was performed to measure Cf. In each patient, 2-compartment model analysis was done on 14-point dynamic SPECT data and 13 series of successive 2-point SPECT data generated by merging dynamic SPECT data divided between every 2 successive rotations. Using the uncalibrated standard arterial input function, 14-point and 2-point data were fitted to 2-compartment model with nonlinear regression and Newton-Raphson’s methods to obtain a product of influx rate constant (K1: ml/min) and Cf (CfK1), and efflux rate constant (k2: 1/min). Dividing CfK1 by k2, the quotient (CfK1/k2) gave a product of Cf and Vd (CfVd). CfVds obtained from 2-point data analyses were compared with that from 14-point data analysis as a reference to determine the best dividing time point in terms of their correlations. Results: In correlation with CfVd from 14-point data analysis, the best correlation was obtained from the 2-point data divided at 10 min after injection (R2=0.989, P < 0.0001). Based on this result, linear regression equation between Measured Cf and estimated CfVd from 2-point data analysis was as follows; Cf=0.0238 CfVd +0.316 (R2=0.665, RMSE=0.0671, P < 0.0001). Conclusions: Two-compartment model analysis of successive two SPECT scans could be applied to estimation of the calibration coefficients for the standard arterial input function without blood sampling in IMP ARG method.
