Noise reduction and convergence of Bayesian algorithms with blobs based on the Huber function and median root prior.

Iterative image reconstruction algorithms have the potential to produce low noise images. Early stopping of the iteration process is problematic because some features of the image may converge slowly. On the other hand, there may be noise build-up with increased number of iterations. Therefore, we examined the stabilizing effect of using two different prior functions as well as image representation by blobs so that the number of iterations could be increased without noise build-up. Reconstruction was performed of simulated phantoms and of real data acquired by positron emission tomography. Image quality measures were calculated for images reconstructed with or without priors. Both priors stabilized the iteration process. The first prior based on the Huber function reduced the noise without significant loss of contrast recovery of small spots, but the drawback of the method was the difficulty in finding optimal values of two free parameters. The second method based on a median root prior has only one Bayesian parameter which was easy to set, but it should be taken into account that the image resolution while using that prior has to be chosen sufficiently high not to cause the complete removal of small spots. In conclusion, the Huber penalty function gives accurate and low noise images, but it may be difficult to determine the parameters. The median root prior method is not quite as accurate but may be used if image resolution is increased.

Influence of nitric oxide synthase and adrenergic inhibition on adenosine-induced myocardial hyperemia.

BACKGROUND: Myocardial perfusion during adenosine-induced hyperemia is used both in clinical diagnosis of coronary heart disease and for scientific investigations of the myocardial microcirculation. The objective of this study was to clarify whether adenosine-induced hyperemia is dependent on endothelial NO production or is influenced by adrenergic mechanisms. METHODS AND RESULTS: In 12 healthy men, myocardial perfusion was measured with PET in 2 protocols performed in random order, each including 3 perfusion measurements. First, perfusion was measured at rest. Second, either saline or the NO synthase inhibitor N(G)-nitro-L-arginine methyl ester (L-NAME, 4 mg/kg) was infused, and perfusion during adenosine-induced hyperemia was determined. Last, in both protocols, the alpha-receptor blocker phentolamine was infused, and perfusion during adenosine-induced hyperemia was determined again. Resting perfusion was similar in the 2 protocols (0.69+/-0.14 and 0.66+/-0.18 mL. min(-1). g(-1)). L-NAME increased mean arterial blood pressure by 12+/-7 mm Hg (P<0.01) and reduced heart rate by 16+/-7 bpm (P<0.01). Adenosine-induced hyperemia (1.90+/-0.33 mL. min(-1). g(-1)) was attenuated by L-NAME (1.50+/-0.55 mL. min(-1). g(-1), P<0.01). The addition of phentolamine had no effect on the adenosine-induced hyperemia (2.10+/-0.34 mL. min(-1). g(-1), P=NS). In the presence of L-NAME, however, when the adenosine response was attenuated, phentolamine was able to increase hyperemic perfusion (2.05+/-0.44 mL. min(-1). g(-1), P<0.05). CONCLUSIONS: Inhibition of endogenous NO synthesis attenuates myocardial perfusion during adenosine-induced hyperemia, indicating that coronary vasodilation by adenosine is partly endothelium dependent. alpha-Adrenergic blockade has no effect on adenosine-induced hyperemia unless NO synthesis is inhibited.

FDOPA metabolism in the adult porcine brain: influence of tracer circulation time and VOI selection on estimates of striatal DOPA decarboxylation.

Different methodologies for PET data analysis influence the magnitude of estimates of blood-brain transfer coefficients and rate constants for the metabolism of FDOPA in living striatum. We now test the effects on several kinetic parameters of automatic procedures for volume of interest (VOI) selection. We also tested the sensitivity of the estimates to dynamic frame sequence duration, and produced a standard method for minimizing the variations in physiological estimates for FDOPA kinetics in minipig brain. We used minipigs because our previous work has shown them to provide an appropriate animal model for study normal and pathological cerebral DOPA metabolism using PET. Time-activity curves in striatum of adult minipigs were acquired in VOIs defined manually on MR-images, or alternatively on the basis of the radioactivity concentration based on the most radioactive voxel in the last scan frame. For all frame sequences, the relative decarboxylase activity (k(3)(D)) declined significantly (P < 0.006) as the VOI threshold declined from 95 to 70% of the most radioactive voxel. Irrespective of VOI size, the magnitude of k(3)(D) declined significantly (P < 0.001) from 0.074+/-0.008 to 0.045+/-0.005 per min (mean+/-S.E.M.) as total sequence length increased from 60 to 120 min circulation. The method of VOI selection had no significant effect on the striatum decarboxylation index of FDOPA calculated relative to the radioactivity in cerebellum (k(3)(S)).

Acute neuroleptic stimulates DOPA decarboxylase in porcine brain in vivo.

The activity of DOPA decarboxylase measured in homogenates from rat striatum, or calculated from the rate of tracer decarboxylation measured ex vivo, is stimulated following acute treatment with antagonists of dopamine D2-like receptors. We used compartmental kinetics to test the hypothesis that utilization of the DOPA decarboxylase substrate [(18)F]fluorodopa is potentiated in living striatum following acute treatment with a typical neuroleptic. The kinetics of the tracer uptake were determined in eight anesthetized female pigs (40 kg) and in three animals receiving an infusion of haloperidol (75 microg kg(-1) h(-1)) for 1 h prior to tracer administration and throughout the 2-h positron emission recording. The relative activity of DOPA decarboxylase in striatum was increased threefold by the treatment. This potentiation of DOPA decarboxylation after pharmacological blockade of dopamine D2-like receptors may be used to optimize the utilization of exogenous DOPA in the treatment of Parkinson's disease.

Electromechanical mapping for detection of myocardial viability in patients with ischemic cardiomyopathy.

BACKGROUND: We evaluated the ability of electromechanical mapping of the left ventricle to distinguish between nonviable and viable myocardium in patients with ischemic cardiomyopathy. METHODS AND RESULTS: Unipolar voltage amplitudes and local endocardial shortening were measured in 31 patients (mean+/-SD age, 62+/-8 years) with ischemic cardiomyopathy (ejection fraction, 30+/-9%). Dysfunctional regions, identified by 3D echocardiography, were characterized as nonviable when PET revealed matched reduction of perfusion and metabolism and as viable when perfusion was reduced or normal and metabolism was preserved. Mean unipolar voltage amplitudes and local shortening differed among normal, nonviable, and viable dysfunctional segments. Coefficient of variation for local shortening exceeded differences between groups and did not allow distinction between normal and dysfunctional myocardium. Optimum nominal discriminatory unipolar voltage amplitude between nonviable and viable dysfunctional myocardium was 6.5 mV, but we observed a great overlap between groups. Individual cutoff levels calculated as a percentage of electrical activity in normal segments were more accurate in the detection of viable dysfunctional myocardium than a general nominal cutoff level. The optimum normalized discriminatory value was 68%. Sensitivity and specificity were 78% for the normalized discriminatory value compared with 69% for the nominal value (P:<0.02). CONCLUSIONS: Endocardial ECG amplitudes in patients with ischemic cardiomyopathy display a wide scatter, complicating the establishment of exact nominal values that allow distinction between viable and nonviable areas. Individual normalization of unipolar voltage amplitudes improves diagnostic accuracy. Electroanatomic mapping may enable identification of myocardial viability.

Pulmonary beta adrenoceptor density in arrhythmogenic right ventricular cardiomyopathy and idiopathic tachycardia.

OBJECTIVE: In recent in vivo studies using positron emission tomography (PET) our group demonstrated that the myocardial beta adrenoceptor (betaAR) density is reduced in arrhythmogenic right ventricular cardiomyopathy (ARVC) and idiopathic right ventricular outflow tract tachycardia (RVO-VT) associated with an increased presynaptic catecholamine washout. It was hypothesised that the reduction of myocardial betaAR density is secondary to an increase of local catecholamines in the myocardium resulting from the presynaptic dysfunction since circulating plasma catecholamines were demonstrated to be unchanged in these conditions. To further prove this hypothesis of an organ-limited adrenergic nervous dysfunction of the heart, this study aimed to investigate betaAR density in another thoracic organ, the lung. METHODS: Pulmonary and myocardial betaAR density was measured in 7 ARVC patients, 8 RVO-VT patients and in a group of healthy controls (n = 13) using the non-selective beta-blocker [11C]-CGP 12177 and PET. RESULTS: Pulmonary betaAR density was similar in controls (12.4 +/- 1.7 pmol/g tissue), ARVC (11.6 +/- 1.7 pmol/g tissue, p = ns) and RVO-VT (12.8 +/- 2.0 pmol/g tissue, p = ns), whereas myocardial betaAR density was significantly reduced in ARVC (6.3 +/- 1.1 pmol/g tissue, p = 0.006) and RVO-VT (6.8 +/- 1.2 pmol/g tissue, p=0.02) as compared to controls (8.8+/-1.5 pmol/g tissue). CONCLUSION: The unchanged pulmonary betaAR density in the presence of a previously described significant reduction in myocardial betaAR density in the same patient principally supports our pathophysiological hypothesis that the myocardial betaAR density may be reduced in ARVC and RVO-VT because of an increase in local synaptic catecholamine levels due to an organ-limited presynaptic adrenergic dysfunction of the heart. Since in the present study only pulmonary betaAR density was measured, future functional studies excluding pulmonary betaAR desensitisation are required to finally prove the unchanged pulmonary sympathetic innervation in ARVC and RVO-VT.

The DaNeX study of embryonic mesencephalic, dopaminergic tissue grafted to a minipig model of Parkinson's disease: preliminary findings of effect of MPTP poisoning on striatal dopaminergic markers.

A multicenter study is under way to investigate the efficacy of allografting of embryonic mesencephalic neurons in a pig model of Parkinson's disease. We have first established that a stable parkinsonian syndrome can be established by 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) intoxication of adult male Göttingen minipigs. We are now using positron emission tomography (PET) methods for testing the physiological responses to MPTP intoxication and the time course of the response to several treatment strategies. We now report preliminary results obtained in 11 pigs employed in the initial phase of the study; the completed study shall ultimately include 30 pigs. Animals were randomly assigned to one of five groups: 1) Control, 2) MPTP intoxication, 3) MPTP intoxication followed by allograft, 4) MPTP intoxication followed by allograft with immunosuppression, and 5) MPTP intoxication followed by allograft with immunosuppression and co-grafting of immortalized HiB5 cells, which had been manipulated to secrete glia cell line-derived neurotrophic factor (GDNF) (approximately 2 ng GDNF/h/10(5) cells). MPTP was administered (1 mg/kg/day, SC) for 7-10 days until the pigs had developed mild parkinsonian symptoms of muscle rigidity, hypokinesia, and impaired coordination, especially of the hind limbs. Approximately 2 weeks after the last MPTP dose, animals received a T1-weighted magnetic resonance imaging (MRI) scan, and a series of dynamic PET recordings. After the first series of PET scans, four grafts of porcine embryonic mesencephalic tissue (E28 days) were placed in each striatum of some MPTP-intoxicated pigs, using MRI-based stereotactic techniques. Immunosuppression of some animals with cyclosporin and prednisolone began just prior to surgery. Two more series of PET scans were performed at 4-month intervals after surgery. After the last scans, pigs were killed and the brains were perfused for unbiased stereological examination of cytological and histochemical markers in striatum and substantial nigra. The behavioral impairment of the animals (the "Parkinson's score") had been evaluated throughout the 8-month period. Kinetic analysis of the first set of PET scans has indicated that the rate constant for the decarboxylation of FDOPA in catecholamine fibers was reduced by 33% in striatum of the mildly parkinsonian pigs. The rate of association of [11C]NS-2214 to catecholamine uptake sites was reduced by 62% in the same groups of pigs. No significant difference was found in the binding potential of [11C]raclopride to the dopamine D2-like receptors in striatum of the MPTP-intoxicated versus control pigs. These preliminary results are suggestive that the activity of DOPA decarboxylase may be upregulated in the partially denervated pig striatum.

Abnormalities of cardiac sympathetic innervation in arrhythmogenic right ventricular cardiomyopathy : quantitative assessment of presynaptic norepinephrine reuptake and postsynaptic beta-adrenergic receptor density with positron emission tomography.

BACKGROUND: The frequent provocation of ventricular tachycardia by stress or catecholamines and the efficacy of antiarrhythmic drugs with antiadrenergic properties suggest an involvement of the cardiac adrenergic system in arrhythmogenesis in patients with arrhythmogenic right ventricular cardiomyopathy (ARVC). Previous studies demonstrated abnormalities of the presynaptic uptake-1 assessed by (123)I-MIBG-single-photon emission computed tomography. METHODS AND RESULTS: This study investigated neuronal reuptake of norepinephrine (uptake-1) and beta-adrenergic receptor density in 8 patients with ARVC and 29 age-matched control subjects. All subjects underwent positron emission tomography with the volume of distribution (V(d)) of [(11)C]hydroxyephedrine ((11)C-HED) used to assess presynaptic norepinephrine reuptake, the maximum binding capacity (B(max)) of [(11)C]CGP-12177 ((11)C-CGP-12177) to assess postsynaptic beta-adrenergic receptor density, and [(15)O]H(2)O for quantification of myocardial blood flow. Patients with ARVC demonstrated a highly significant global reduction in postsynaptic beta-adrenergic receptor density compared with that in control subjects (B(max) of (11)C-CGP-12177: 5.9+/-1.3 vs 10.2+/-2.9 pmol/g tissue, P<0.0007), whereas the presynaptic uptake-1 tended toward reduction only (V(d) of (11)C-HED: 59.1+/-25.2 vs 71.0+/-18.8 mL/g tissue, NS). There were no differences in myocardial blood flow between the groups, and plasma norepinephrine was within normal limits in patients and control subjects. CONCLUSIONS: The findings demonstrate a significant reduction of myocardial beta-adrenergic receptor density in patients with ARVC. This may result from a secondary downregulation after increased local synaptic norepinephrine levels caused by increased firing rates of the efferent neurons or as the result of impaired presynaptic catecholamine reuptake. These findings give new insights into the pathophysiology of arrhythmogenesis in ARVC, with potential impact on diagnostic evaluation and therapeutic management.

Cerebral 6-[(18)F]fluoro-L-DOPA (FDOPA) metabolism in pig studied by positron emission tomography.

We measured 6-[(18)F]fluoro-L-DOPA (FDOPA) uptake and metabolism in the brain of 4-month-old female pigs (n = 8) using a high-resolution positron emission tomograph (PET) in 3D mode. The mean net blood-brain clearance of FDOPA (K(i)(D)) to striatum was 0.011 ml g(-1) min(-1). Correcting for the elimination of decarboxylated metabolites from striatum (k(loss) = 0.004 min(-1)) increased the apparent magnitude of the estimate of K(i)(D) by 50%, at the expense of doubling the variance of the mean estimate. The mean decarboxylation rate of FDOPA in striatum relative to the cerebellum input (k(3)(s)) was 0.008 min(-1). For multicompartmental analyses, the FDOPA partition volume (V(e)(D)) was constrained to the individual value observed in cerebellum (mean = 0.53 ml g(-1)), with correction for the presence in brain of the plasma metabolite 3-O-methyl-FDOPA (OMFD). Using the first 60 min of the dynamic PET scans, the rate constant of FDOPA decarboxylation (k(3)(D)) was estimated to be 0.037 min(-1 )in striatum, but was not significantly different than zero in frontal cortex. Fitting of a compartmental model correcting for elimination of decarboxylated metabolites to the complete PET frame-sequence (120 min) increased the variance of the estimate of k(3)(D) in striatum. The magnitude of k(3)(D) in striatum of young pig was less than values estimated previously in neonatal piglet, adult monkey, and human. MRI-based simulations predicted that recovery of radioactivity from pig striatum was highly sensitive to the volume of interest. We conclude that the spatial resolution of our tomograph reduces the apparent magnitude of k(3)(D) in striatum. However, anaesthetised pigs are an appropriate experimental model for PET studies of DOPA decarboxylation in striatum.

Effect of a full bladder on urine production in humans.

OBJECTIVE: To demonstrate the reduction in urine production in healthy humans upon bladder distension and to identify the factors responsible for this reduction. MATERIALS AND METHODS: Twelve healthy females were investigated twice in a cross-over designed experiment: once with the urinary bladder empty at the start and once pre-filled to 60% of the maximum bladder capacity. Glomerular filtration rate, effective renal plasma flow, urine content of catecholamines, blood pressure, pulse rate, plasma arginine vasopressin (AVP) and plasma concentration of renine and electrolytes were analysed together with serum osmolality. RESULTS: Three subjects failed to reach maximum bladder capacity during the "full bladder" test and were excluded. The urine production in the "full-bladder" test was significantly lower than the "empty-bladder" test (p = 0.024). In the "full-bladder" test a significant increase in mean blood pressure was found (p=0.01). No further significant changes were demonstrated. CONCLUSIONS: Acute bladder distension causes a reduction in urine production or a "pooling of urine" in the upper urinary tract in healthy humans. The mechanism is unknown.

Generation of myocardial factor images directly from the dynamic oxygen-15-water scan without use of an oxygen-15-carbon monoxide blood-pool scan.

UNLABELLED: The measurement of regional myocardial blood flow (MBF) with H2(15)O and PET requires an additional C15O blood-pool scan for the purpose of region of interest (ROI) definition. This additional scan results in a substantially increased radiation dose, study duration and risk of movement artifacts. Therefore, a method was developed to generate myocardial factor images directly from the dynamic H2(15)O study without the need for a C15O scan. METHODS: The factor sinograms were generated by means of linear dimension reduction of the dynamic sinograms, where the required variate and covariate factors (myocardial and blood time-activity curves) were modeled from the lung time-activity curve. The factor images were generated by iterative reconstruction. RESULTS: No significant difference was found between MBF values from ROIs drawn on the traditional images (using the C15O scan) and those drawn on the factor images. CONCLUSION: It is possible to generate myocardial images directly from the dynamic H2(15)O study, so that the C15O scan can be omitted from MBF studies. The proposed method is robust and results in nearly optimal signal-to-noise ratios in the factor images.

Measurement of myocardial blood flow with oxygen-15 labelled water: comparison of different administration protocols.

Positron emission tomography (PET) in conjunction with C15O2 or H215O can be used to measure myocardial blood flow (MBF) and tissue fraction (TF), i.e. the fraction of the tissue mass in the volume of the region of interest. However, with C15O2 inhalation, the tissue fraction in the septum is overestimated. Bolus injection of H215O together with arterial cannulation gives very precise results but is invasive. The purpose of this study was to develop a method which circumvents these problems. A four-parameter model with parameters for MBF, TF and spill-over fractions from both left and right ventricular cavities was developed. This method was compared with a three-parameter model (no right ventricular cavity spill-over) in both septal and non-septal regions of interest for three different administration protocols: bolus injection of H215O, infusion of H215O and inhalation of C15O2. It was found that MBF can be measured with intravenous administration of H215O without the requirement for arterial cannulation. The four-parameter protocol with bolus injection was stable in clinical studies. The four-parameter model proved essential for the septum, where it gave highly significantly better fits than did the three-parameter model (P<0.00003 in each of 15 subjects). Administration of H215O together with this four-parameter model also circumvented the problem of overestimation of TF in the septum seen with C15O2 inhalation. In addition, the radiation dose of H215O protocols is lower than that of C15O2 inhalation. Using a left atrial input curve instead of a left ventricular cavity input curve gave the same mean MBF and TF.

Cardiac sympathetic innervation in patients with idiopathic right ventricular outflow tract tachycardia.

OBJECTIVES: This study investigated the neuronal reuptake of norepinephrine (uptake-1) and the beta-adrenoceptor density in patients with idiopathic right ventricular outflow tract tachycardia (RVO-VT). BACKGROUND: Clinical findings, such as the inducibility of ventricular tachycardia by stress or catecholamine infusion, and the therapeutic efficacy of antiarrhythmic drugs with antiadrenergic properties suggest abnormalities of cardiac sympathetic innervation in patients with idiopathic RVO-VT. METHODS: Eight patients with idiopathic RVO-VT and a total of 29 age-matched control subjects were investigated by positron emission tomography using [11C]hydroxyephedrine (HED) (volume of distribution of [11C]HED) to assess presynaptic norepinephrine reuptake; [11C]CGP 12177 (maximal binding capacity of [11C]CGP 12177) to measure postsynaptic beta-adrenoceptor density; and oxygen-15-labeled water for quantification of myocardial blood flow (MBF). RESULTS: Both myocardial catecholamine reuptake and beta-adrenoceptor density were significantly reduced in patients with idiopathic RVO-VT. The volume of distribution of [11C]HED in patients with RVO-VT was (mean +/- SD) 41.0 +/- 13.5 versus 71.0 +/- 18.8 ml/g in control subjects (p < 0.002). The maximal binding capacity of the beta-adrenoceptor antagonist [11C] CGP 12177 was 6.8 +/- 1.2 pmol/g in patients with RVO-VT versus 10.2 +/- 2.9 pmol/g in control subjects (p < 0.004). There were no significant differences in MBF at rest (0.98 +/- 0.14 vs. 0.97 +/- 0.24 ml/min per g, p = NS) between patients with RVO-VT and control subjects. CONCLUSIONS: The findings of the present study suggest that myocardial beta-adrenoceptor downregulation in patients with RVO-VT occurs subsequently to increased local synaptic catecholamine levels caused by impaired catecholamine reuptake.

Myocardial presynaptic and postsynaptic autonomic dysfunction in hypertrophic cardiomyopathy.

Although hypertrophic cardiomyopathy (HCM) is genetically determined, several other factors, including autonomic dysfunction, may play a role in the phenotypic expression. A recent study using positron emission tomography with [11C]CGP 12177 ([11C]CGP) demonstrated that beta-adrenoceptor (betaAR) density is reduced in HCM and is correlated with disease progression. This present study tested the hypothesis that this downregulation is associated with reduced catecholamine reuptake (uptake 1) by myocardial sympathetic nerve terminals leading to increased local norepinephrine concentration. Myocardial presynaptic catecholamine reuptake was assessed by measuring the volume of distribution (Vd) of the catecholamine analogue [11C]hydroxyephedrine ([11C]HED) in 9 unrelated HCM patients aged 45+/-15 years. The maximum number of binding sites (Bmax) for myocardial betaAR density was measured in 13 unrelated HCM patients aged 40+/-12 years using the nonselective beta blocker [11C]CGP. Six patients were studied with both [11C]HED and [11C]CGP. Comparison was made with two groups of healthy control subjects for each ligand ([11C]HED, n=10, aged 35+/-8 years; [11C]CGP, n=19, aged 44+/-16 years). Myocardial Vd of [11C]HED (33.4+/-4.3 mL/g tissue) and betaAR density (7.3+/-2.6 pmol/g tissue) were significantly reduced in HCM patients compared with control subjects (71.0+/-18.8 mL/g tissue, P<.001, and 10.2+/-2.9 pmol/g tissue, P=.008, respectively). These results are consistent with our hypothesis that myocardial betaAR downregulation in HCM is associated with an impaired uptake-1 mechanism and hence increased local catecholamine levels.

Calculation of single detector efficiencies and extension of the normalization sinogram in PET.

A fast iterative method is presented for calculating single detector efficiencies for positron emission tomographs. These efficiencies can be used to extend the normalization scan to areas outside that covered by the normalization source. The root mean square (rms) error of the calculated single detector efficiencies decreases exponentially with the number of iterations. Thirty iterations per normalization image are sufficient and take about 1 s on a SUN Classic. The geometry factors are composed of factors which only depend on the distance from the centre of the field-of-view (FOV) and of factors which show a more complex pattern over the normalization sinogram. The geometry factors are specific to each scanner. On the ECAT 931 scanner the complex part of the geometry factors showed a diamond shaped pattern (caused by the varying sensitivity of single detectors in a detector block with varying angle of incidence) and S-shaped curves (representing attenuation caused by supporting rods for the ring source). The coefficient of variation of the diamond-shaped pattern was 4% for detectors farthest from the centre of the FOV. Extensions of the normalization scan may, therefore, contain a relative rms error of about 4% if the applied geometry factors only take the distance from the centre of the FOV into account.

Linear dimension reduction of sequences of medical images: III. Factor analysis in signal space.

A method is presented for improving the precision of factor analysis by utilizing physiological information. The first step is an optimal linear dimension reduction, whereby the data are projected onto a low-dimensional signal space. Then principal component analysis is performed in the signal space rather than in the entire data space. This improves the precision of the principal components. Unlike ordinary principal component analysis, the present method is not degraded when the time intervals are subdivided, provided that the signal space is correct. Alternatively, but with identical results, the covariance matrix can be calculated from the whole data space. The covariance matrix is then transformed and principal component analysis is performed in either a low-rank matrix or a low-dimensional submatrix instead of in the whole covariance matrix. Factor analysis using the intersection method with a theory space may be improved by employing the present method. In simulations based on a [11C]flumazenil study with 27 frames, the proposed method required only 58 per cent of the radioactivity to produce the same precision as the intersection method and only 27 per cent when compared to ordinary principal component analysis.

Linear dimension reduction of sequences of medical images: I. Optimal inner products.

A general theory is presented for minimizing noise in linear dimension reduction of sequences of medical images when the factors and the covariance matrix and mean of the noise are given. A dimension reduction is optimal when all diagonal elements in the covariance matrix of the noise in the signal (factor) space are minimized. This occurs when the noise in the signal space is uncorrelated with the residual noise. Expressions are given for the resulting covariance matrix of the noise in the signal space. Many optimal inner products exist, which all result in the same optimal dimension reduction. Given any pair of inner products for signal space and residual space, a combined inner product exists that is also optimal. If the covariance matrices of the noise in different pixel vectors are not multiples of each other, different pixel vectors may have different optimal inner products. The averaging process in generating images from tomographic projections tends to make the covariance matrices more uniform.

Linear dimension reduction of sequences of medical images: II. Direct sum decomposition.

Using unitary transformations together with a previously described statistical theory for optimal linear dimension reduction it is shown how pixels in a sequence of images can be decomposed into a sum of variates, covariates, and residual vectors, with all covariances equal to zero. It is demonstrated that this decomposition is optimal with respect to noise. In addition, it results in simplified and well conditioned equations for dimension reduction and elimination of covariates. The factor images are not degraded by subdivision of the time intervals. In contrast to traditional factor analysis, the factors can be measured directly or calculated based on physiological models. This procedure not only solves the rotation problem associated with factor analysis, but also eliminates the need for calculation of the principal components altogether. Examples are given of factor images of the heart, generated from a dynamic study using oxygen-15-labelled water and positron emission tomography. As a special application of the method, it is shown that the factor images may reveal any contamination of the blood curve derived from the original dynamic images with myocardial activity.

Switching from DXA pencil-beam to fan-beam. II: Studies in vivo.

Switching from the Hologic QDR-1000/W to the QDR-2000 DXA densitometer was critically evaluated with regard to cross-calibration and dosimetry. Studies with bone equivalent humanoid spine phantoms and patient studies were done. Fan-beam scanning with the QDR-2000 is problematic because of magnification. Mean phantom bone mineral content (BMC) and bone mineral density (BMD) were moderately but significantly different. Biological variation disguised differences between the two devices in humans, but significant differences were revealed when individual data were analyzed. Longitudinal assessments of BMC and BMD, initiated with QDR-1000/W and continued with the QDR-2000, should employ single-beam mode only and not fan-beam mode--but even if that is done, significant errors can be introduced. The new QDR-2000 should be properly cross-calibrated with the original densitometer, and one should make sure that the same software, phantom, and type of collimator are used. The radiation dose is substantially higher with QDR-2000 (fan-beam and high-resolution array mode) than with QDR-1000/W (pencil-beam mode) and QDR-2000 (pencil-beam mode), and higher than claimed by the manufacturer. The typical radiation dose given by the manufacturer was half the actual radiation dose measured (e.g., for fan-beam scan 62 microSv versus 33 microSv). High-resolution array mode does not improve precision, but augments the radiation dose to the patient.