Radiation Protection Dosimetry (2008), Vol. 129, No. 1–3, pp. 288–290 Advance Access publication 24 April 2008 doi:10.1093/rpd/ncn153 COUNT-RATE ANALYSIS FROM CLINICAL SCANS IN PET WITH LSO DETECTORS F. Bonutti1, *, E. Cattaruzzi2, E. Cragnolini1, M. Floreani1, C. Foti1, M. R. Malisan1, E. Moretti1, O. Geatti2 and R. Padovani2 1 Medical Physics Department, University Hospital, Udine, Italy 2 Nuclear Medicine Department, University Hospital, Udine, Italy INTRODUCTION The optimisation of the acquisition parameters in positron emission tomography (PET) has the purpose to improve the quality of the diagnostic images. Optimisation parameters deal with the administered activity, time interval between the radiopharmaceutical injection and the PET scan, duration of the scan, and the patient size. Optimisation can be done by maximising the signalto-noise ratio (SNR), employing noise equivalent count rate (NECR) calculations that in turn depend on the coincidence rate. As already proposed in the literature(1,2) it is possible to set personalised values of the acquisition parameters for each patient. METHODS In ‘conventional’ nuclear medicine single photon emission computed tomography (SPECT), the SNR goes as the square root of the total amount of counts in the whole image, and the single photon events are at the same time source of signal and source of noise. In PET this is not valid anymore. In fact, while the ‘useful’ signal comes from the true coincidences, the noise has three different sources: the true, the random and the scattering coincidences. This has led to defining the NEC (noise equivalent count) as those fictitious coincidences that alone, and therefore as unique source of signal and noise, result in the SNR of the acquired images. In terms of the *Corresponding author: bonutti.faustino@aoud.sanita.fvg.it coincidence rates, the NEC became NECR, which can be simplified to PNECR ( pseudo NECR)(1). This work comprises the results from a retrospective analysis of the clinical coincidence rates collected from 500 scans performed with an lutetium orthosilicate (LSO)-based PET/CT scanner (Biograph Duo, Siemens). For each bed position, the scanner records the coincidences and the single counts in a log-file, which can be analysed off-line later. In order to optimise the injected activity, for each patient the functions T ¼ T(s), R ¼ R(s), S ¼ S(s) (where T, R, S and s are, respectively, the true, random, scattering and single count rates) are established by fitting the NEMA (National Electrical Manufacturers Association) 70 cm phantom(3) count rate curves to the measured clinical points, in order to analytically calculate the personalised PNECR(s) curve(1). RESULTS The PNECR(s) curves relative to the central bed ( position) for seven-bed scans are reported in Figure 1, where the single count rates are significantly far from the PNECR maximum (PNECRmax). Assuming that the ratio between the activity at PNECRmax and activity actually administered is the same for patient and phantom, it is possible to calculate the required activity to get PNECRmax. For each bed, Figure 2 shows the percentage of missing activity to get PNECRmax. For the central bed (position number 4), missing activity of 70% is estimated, but the improvement # The Author 2008. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org Downloaded from http://rpd.oxfordjournals.org/ at Azienda Ospedaliera on October 17, 2012 The purpose of optimising the acquisition parameters in positron emission tomography is to improve the quality of the diagnostic images. Optimisation can be done by maximising the noise equivalent count rate (NECR) that in turn depends on the coincidence rate. For each bed position the scanner records coincidences and singles rates. For each patient, the true, random and scattered coincidences as functions of the single count rate(s) are determined by fitting the NEMA (National Electrical Manufacturers Association) 70 cm phantom count rate curves to measured clinical points. This enables analytical calculation of the personalised PNECR [ pseudo NECR(s)] curve, linked to the NECR curve. For central bed positions, missing activity of 70% is estimated to get maximum PNECR (PNECRmax), but the improvement in terms of signal-toznoise ratio would be 15%. The correlation between patient weight and PNECRmax is also estimated to determine the optimal scan duration of a single bed position as a function of patient weight at the same PNEC. Normalising the counts at PNECRmax for the 70 kg patient, the bed duration for a 90 kg patient should be 230 s, which is 30% longer. Although the analysis indicates that the fast scanner electronics allow using higher administered activities, this would involve poor improvement in terms of NECR. Instead, attending to higher bed duration for heavier patients may be more useful. COUNT-RATE ANALYSIS Figure 2. Percentage missing activity to get the maximum PNECR value for various bed positions. in the PNECR that could be obtained using such activity is 15%. Since the gain in terms of PNECR is so low, from the patient exposure point of view it could seem unjustified to inject higher activities than those currently administered. As proposed before(1), it is also interesting to estimate the correlation between patient weight and PNECRmax. Such correlation allows for estimating, which should be the scan duration of a single bed, as a function of patient weight to acquire the same PNEC. If the counts are normalised to PNECRmax for the 70 kg patient, it is found that with increasing patient weight, the duration of the single bed has to increase proportionally (Figure 3). Based on this analysis, the bed duration for a 90 kg patient should be 230 s, i.e. 30 % longer than the 180 s bed duration conventionally adopted. Finally, in Figure 4, the correlation is shown between the percentual distance of the measured PNECR from the PNECRmax and the time interval between activity injection and scan ( patient weight between 60 and 70 kg, with a mean injected activity of 363 MBq, s.d. ¼ 22 MBq). As there is a direct Figure 3. Dependence of the duration of single bed-scan on patient weight. Figure 4. Dependence of the percentage distance from PNECRmax on the time interval (h.mm.ss) between activity injection and scan. proportionality between single counts and activity (A), it is expected that the PNECR gets worse in the time (t) according to the relation PNECRfs[A(t)]g as shown in dashed line. It can be useful to interpret the diagram of Figure 4 as an indication of how much the SNR goes away from its maximum value as the time between the activity administration and the scan increases. CONCLUSIONS Although the analysis indicates that the fast electronics implemented in the scanner allows the use of higher administered activities, it was found that this would involve a poor improvement in terms of NECR. Instead, it would be better to focus attention on the usefulness of higher bed duration for heavier patients. FUNDING This work, part of the Sentinel Project, was partially supported by the European Commission, Euratom Research and Training Programme on Nuclear Energy, contract no. 01 2909. 289 Downloaded from http://rpd.oxfordjournals.org/ at Azienda Ospedaliera on October 17, 2012 Figure 1. Clinical PNECR and coincidence rates of patients of 52–62 kg weight and 155 –170 cm height. F. BONUTTI ET AL. REFERENCES 1. Watson, C. C., Casey, M. F., Beyer, T., Bruckbauer, T., Townsend, D. W. and Brasse, D. Evaluation of clinical PET count rate performance. In: Nuclear Science Symposium Conference Record, 2002 IEEE, Vol. 3, pp.1406– 1410 (2002). 2. Lartizien, C., Comtat, C., Kinahan, P. E., Ferreira, N., Bendriem, B. and Trébossen, R. Optimization of injected dose based on noise equivalent count rates for 2- and 3dimensional whole-body PET. J. Nucl. Med. 43, 1268– 1278 (2002). 3. National Electrical Manufacturers Association. NEMA Standards Publication NU 2-2001: Performance Measurements of Positron Emission Tomographs (Rosslyn, VA: National Electrical Manufacturers Association) (2001). Downloaded from http://rpd.oxfordjournals.org/ at Azienda Ospedaliera on October 17, 2012 290