A multiresolution image based approach for correction of partial volume effects in emission tomography.

Partial volume effects (PVEs) are consequences of the limited spatial resolution in emission tomography. They lead to a loss of signal in tissues of size similar to the point spread function and induce activity spillover between regions. Although PVE can be corrected for by using algorithms that provide the correct radioactivity concentration in a series of regions of interest (ROIs), so far little attention has been given to the possibility of creating improved images as a result of PVE correction. Potential advantages of PVE-corrected images include the ability to accurately delineate functional volumes as well as improving tumour-to-background ratio, resulting in an associated improvement in the analysis of response to therapy studies and diagnostic examinations, respectively. The objective of our study was therefore to develop a methodology for PVE correction not only to enable the accurate recuperation of activity concentrations, but also to generate PVE-corrected images. In the multiresolution analysis that we define here, details of a high-resolution image H (MRI or CT) are extracted, transformed and integrated in a low-resolution image L (PET or SPECT). A discrete wavelet transform of both H and L images is performed by using the "à trous" algorithm, which allows the spatial frequencies (details, edges, textures) to be obtained easily at a level of resolution common to H and L. A model is then inferred to build the lacking details of L from the high-frequency details in H. The process was successfully tested on synthetic and simulated data, proving the ability to obtain accurately corrected images. Quantitative PVE correction was found to be comparable with a method considered as a reference but limited to ROI analyses. Visual improvement and quantitative correction were also obtained in two examples of clinical images, the first using a combined PET/CT scanner with a lymphoma patient and the second using a FDG brain PET and corresponding T1-weighted MRI in an epileptic patient.

Validation of a Monte Carlo simulation of the Philips Allegro/GEMINI PET systems using GATE.

A newly developed simulation toolkit, GATE (Geant4 Application for Tomographic Emission), was used to develop a Monte Carlo simulation of a fully three-dimensional (3D) clinical PET scanner. The Philips Allegro/GEMINI PET systems were simulated in order to (a) allow a detailed study of the parameters affecting the system's performance under various imaging conditions, (b) study the optimization and quantitative accuracy of emission acquisition protocols for dynamic and static imaging, and (c) further validate the potential of GATE for the simulation of clinical PET systems. A model of the detection system and its geometry was developed. The accuracy of the developed detection model was tested through the comparison of simulated and measured results obtained with the Allegro/GEMINI systems for a number of NEMA NU2-2001 performance protocols including spatial resolution, sensitivity and scatter fraction. In addition, an approximate model of the system's dead time at the level of detected single events and coincidences was developed in an attempt to simulate the count rate related performance characteristics of the scanner. The developed dead-time model was assessed under different imaging conditions using the count rate loss and noise equivalent count rates performance protocols of standard and modified NEMA NU2-2001 (whole body imaging conditions) and NEMA NU2-1994 (brain imaging conditions) comparing simulated with experimental measurements obtained with the Allegro/GEMINI PET systems. Finally, a reconstructed image quality protocol was used to assess the overall performance of the developed model. An agreement of <3% was obtained in scatter fraction, with a difference between 4% and 10% in the true and random coincidence count rates respectively, throughout a range of activity concentrations and under various imaging conditions, resulting in <8% differences between simulated and measured noise equivalent count rates performance. Finally, the image quality validation study revealed a good agreement in signal-to-noise ratio and contrast recovery coefficients for a number of different volume spheres and two different (clinical level based) tumour-to-background ratios. In conclusion, these results support the accurate modelling of the Philips Allegro/GEMINI PET systems using GATE in combination with a dead-time model for the signal flow description, which leads to an agreement of <10% in coincidence count rates under different imaging conditions and clinically relevant activity concentration levels.

FDG PET in epithelioid hemangioendothelioma.

Epithelioid hemangioendothelioma (EH) is an uncommon tumor of endothelial origin. It can develop in any tissue and can be multicentric or metastatic. The usual course is a slow progression. Imaging techniques are generally useful in determining the extent of the disease. A case of EH involving bone marrow and mediastinum is described. We discuss the use of FDG PET scanning in EH, showing its use in detecting bone marrow involvement and determining the extent of the disease.

Technology related parameters affecting quantification in positron emission tomography imaging.

Some of the issues associated with positron emission tomography (PET) technology which still pose challenges for the recovery of quantitative images are discussed. Through these issues reference to what is today considered as the 'gold standard' in quantitative PET imaging is also presented. A brief comparison of 2-D and 3-D PET is given, together with a short discussion of combined PET/CT imaging devices.

Post-synchronized scintigraphic data to estimate antral motility.

We evaluated an improved dynamic antral scintigraphy (DAS) technique, without any frequency filtering or computation of an autocorrelation function. This DAS was performed in 15 consecutive patients and 10 healthy volunteers. Antral frequency was first estimated and was given as an input parameter to compute phase and amplitude values in each antral pixel. Motility indices were calculated by multiplying the frequency by a normalized amplitude in the whole antrum. In addition, a gastric emptying (GE) test was performed. Only 10 patients had a delayed GE when using a cut-off value of the mean of half emptying time (T +2SD) obtained in controls. Antral frequencies were significantly increased but motility indices were significantly lower in patients than in controls. These results in patients were accounted for by a retention of food in the antrum. Therefore, amplitude normalization by the antral mean count activity of each set of data was essential for discriminating between patients and controls and normalized indices appeared early predictors of hypomotility in patients with normal GE. This improved DAS technique should be a useful tool to assess antral dysmotility noninvasively, and may be of physiological and clinical interest.

Three-dimensional attenuation map reconstruction using geometrical models and free-form deformations.

We address the issue of using deformable models to reconstruct an unknown attenuation map of the torso from a set of transmission scans. We assume the three-dimensional (3-D) distribution of attenuation coefficients to be piecewise uniform. We represent the unknown distribution by a set of closed surfaces defining regions having the same attenuating properties. The methods of reconstruction published so far tend to directly deform the surfaces, the parameters being the surface elements. Rather than deforming the surfaces, we explore the possibility of deforming the space in which the geometrical primitives are contained. We focus on the use of free-form deformations (FFD's) to describe the continuous transformation of space used to match a set of transmission measurements. We illustrate this approach by reconstructing realistically simulated transmission scans of the torso with various noise levels and compare the results to standard reconstruction methods.

Gastric emptying of solids: estimates of lag phase and constant emptying times.

There is no consensus regarding the best way to estimate the lag phase time (Tlag) and the constant emptying time (TRE) of the gastric emptying of solids. Furthermore, biphasic gastric emptying is usually described by the modified power exponential function of either Elashoff or Siegel. In an attempt to test the validity of the power exponential functions and to identify relevant parameters of biphasic gastric emptying, we followed an approach which consists of describing the power exponential function by two straight lines. The first line is horizontal and represents Tlag. The second line is tangential to the constant emptying [tangent at the maximum slope (MS) or at the half-emptying value]. Scintigraphic data of 132 patients and 15 controls were fitted by both power exponential functions. Each corresponding half-emptying time, Tlag and TRE estimated from the Elashof and Siegel power exponential functions were strongly correlated (0.93 < r < 1, P < 0.0001). The Bland and Altman statistical method demonstrated good agreement (<5% outliers). The half-emptying tangent method sometimes gave negative Tlag and should be abandoned. Tlag(MS) and TRE(MS) did not correlate and therefore were independent parameters. We conclude that the Elashoff and Siegel functions are equivalent and that the maximum slope tangent method allows a reliable description of the two independent phases of gastric emptying.

Post-synchronization of dynamic images of periodically moving organs.

The motion of periodically moving organs can be studied by acquisition of dynamic image series. When time is short, it is necessary either to find a sync signal to sum synchronous images in real time or to acquire a regular time series and to synchronize a posteriori. Dynamic acquisitions were performed (gastric and lung studies). The activity in each pixel of the moving organ can be expressed as h(t)=a0 + a1 cos(omegao0t - phi). The time-activity curve u(t) over a region of interest (ROI) of the considered organ was computed. When the ROI is well chosen, the power spectrum of u(t) exhibits a sharp peak near the characteristic frequency of the periodic motion. The DC component, amplitude and phase in each pixel can be then estimated by minimizing the following function: J=sigma[h(t) - g(t)]2, where g(t) is a noisy measurement of h(t). It is then easy to reconstruct an a posteriori gated time series by computing h(t) for various times over a single period. This approach was successful in characterizing lung and gastric motions. Dynamic series were acquired as for gastric emptying studies. The characteristic frequency of antral motility was easily and unambiguously estimated and DC, amplitude and phase images were computed. Dynamic pulmonary functional imaging was performed with 81Krm. The characteristic frequency was also easily estimated from the time-activity curve power spectrum using a ROI drawn over the lower part of the lungs. The DC, amplitude and phase images were then computed from the dynamic series and the characteristic frequency. In conclusion, a posteriori gating of dynamic series of periodically moving organs can be achieved in a simple fashion. This approach overcomes the difficulty of direct analysis of time-activity curves and provides amplitude and phase images.

13C-octanoic acid breath test for gastric emptying measurement.

OBJECTIVES: The purpose of this multicentre study was to assess the accuracy and reproducibility of the 13C-octanoic acid breath test compared to scintigraphy for measurement of gastric emptying. METHODS: Sixty-nine healthy subjects (40 men, 29 women; mean age 30 years, range 21-61) were studied at least once by using the 13C-octanoic breath test. In 34 healthy subjects, gastric emptying was simultaneously measured by gastric scintigraphy, and the 13C-octanoic breath test was then repeated in 18 of these cases. Fifty-four patients (30 men, 24 women; mean age 46 years, range 13-74) with dyspeptic or reflux symptoms were studied according to the same procedure. RESULTS: The correlation between breath test and scintigraphic values was highly significant (r = 0.744, P<0.001). The concordance of results of scintigraphy and breath test (normal versus abnormal) and the reproducibility (Bland and Altman method) of the breath test were good (CVinter = 24%, CVintra = 15%). As compared to scintigraphy, breath test detected an abnormal gastric emptying with 67% sensitivity and 80% specificity (ROC analysis). CONCLUSION: These results confirm the value of breath test as an accurate measurement of gastric emptying. Its excellent reproducibility makes it a method of choice for pharmacological studies. However, at least when scintigraphy is considered the gold standard, breath test sensitivity may be insufficient for the detection of gastroparesis in an individual patient.

Use of left ventricular pacing in heart failure: evaluation by gated blood pool imaging.

BACKGROUND: Left ventricular (LV) pacing has been suggested to complement other forms of therapy in patients with heart failure. METHODS AND RESULTS: We investigated 17 patients (15 men, 2 women, aged 68 +/- 6 years, 10 ischemic and 7 primary dilated cardiomyopathy) with heart failure (13 were in New York Heart Association class IV and 4 in class III). One month after LV pacer implantation, 12 patients reported clinical improvement (mean class 3.7 before pacing vs 2.6 with LV pacing; P = .001). We report the results of 3 equilibrium-gated blood pool studies performed in each patient, 1 before pacing and 2 after pacer implantation (1 with pacing on, and 1 after turning off the pacer). LV pacing did not modify LV ejection fraction. Phase analysis demonstrated a significant decrease of the interventricular phase shift (delta(pi)) with LV pacing (no pacing, delta(pi) = 8.99 degrees +/- 19.05 degrees; delta7n= -0.97 degrees +/- 27.85 degrees with LV pacing). Clinical improvement was observed in patients with an initial positive delta(pi) that decreased with pacing and/or an initial LV phase standard deviation >50 degrees that decreased with pacing. CONCLUSION: LV pacing induces interventricular and intraventricular synchronization. A decrease of the interventricular phase shift seems to be the most important predictor of functional recovery for paced patients with heart failure.

Evaluation of a new, rapid, and quantitative D-Dimer test in patients with suspected pulmonary embolism.

Previous studies have suggested the utility of D-Dimer ELISA assays in eliminating a diagnosis of pulmonary embolism (PE). Our objectives were to evaluate the performance of a new, rapid, quantitative, and automated Liatest D-Dimer Assay in patients with suspected PE. Three hundred eighty-six consecutive patients referred to our institution between March 1992 and December 1996 for clinically suspected PE, with recent clinical signs not exceeding 1 wk, were included in this study. Diagnosis of PE was based on clinical evaluation, radionuclide lung imaging, lower limb examination, and, when required, pulmonary angiography. D-Dimer performances, for both Liatest D-Dimer and standard D-Dimer ELISA (Asserachrom DDi), assays, were assessed at the end of the study. Among the 386 patients tested, 146 (37.8%) were classified as PE-positive. Liatest D-Dimer assay had a 100% sensitivity (95% confidence interval, 97 to 100%) and a negative predictive value of 100% (95% confidence interval, 94 to 100%). A normal result, below the cutoff of 500 ng/ml, occurred in 83 of the 386 (21%) patients. There was a strong agreement between Liatest D-Dimer and Asserachrom DDi analyses. These findings suggest that this rapid, quantitative, and automated D-Dimer assay provides a useful diagnostic tool for the clinician with regard to exclusion of PE.

Standardization in the field of medical image management: the contribution of the MIMOSA model.

This paper deals with the development of standards in the field of medical imaging and picture archiving and communication systems (PACS's), and notably concerning the interworking between PACS's and hospital information systems (HIS). It explains, in detail, how a conceptual model of the management of medical images, such as the medical image management in an open system architecture (MIMOSA) model, can contribute to the development of standards for medical image management and PACS's. This contribution is twofold: 1) Since the model lists and structures the concepts and resources involved to make the images available to the users when and where they are required, and describes the interactions between PACS components and HIS, the MIMOSA work helps by defining a reference architecture which includes an external description of the various components of a PACS, and a logical structure for assembling them. 2) The model and the implementation of a demonstrator based on this model allow the relevance of the Digital Imaging and Communications in Medicine (DICOM) standard with respect to image management issues to be assessed, highlighting some current limitations of this standard and proposing extensions. Such a twofold action is necessary in order both to bring solutions, even partial, in the short term, and to allow for the convergence, in the long term, of the standards developed by independent standardization groups in medical informatics (e.g., those within Technical Committee 251 of CEN: Comité Européen de Normalisation).

MIMOSA: a functional model of the management of medical images.

The concept of picture archiving and communication system (PACS) appeared in 1982. Twelve years later most PACS prototypes still do not satisfy the medical community requirements: too much emphasis has been put on technical solutions and little effort devoted to the management of images itself. A common approach to this problem is needed more than ever to allow cost-effective systems to be developed. Modelling is a way to reach this objective, and the MIMOSA topic within the AIM/EurlPACS project aims at defining such a model of medical image management. This paper presents the MIMOSA approach with the three underlying constraints, i.e. genericity, implementation independence and performance. The MIMOSA group is building three interrelated models: a data model, a functional model and a dynamic model. However, in this report we focus only on the functional model, and explain how it was abstracted from various clinical scenarios in the first phase of the project. The availability level concept is introduced as a customizable and implementation independent solution to the problem of managing the delay of access to the information. The so-called contextual diagram and the acquisition of a new examination which are two representative parts of the model are detailed. The validation aspects and the relationships to projects working on the management of patient records are addressed as a conclusion.

Bandlimited and Haar filtered back-projection reconstructions.

A new way to discretize the filtered back-projection (FBP) algorithm is presented. The function basis is the Haar system (2D product of rectangular windows). This scheme allows one to derive the optimal shape of the apodisation window, which is angle varying, and the oversampling ratio between the pixel and the projection cell size. The discrete equivalent filter is also derived. The comparison of standard radial band-limited and separable Haar reconstructions shows that improvements, in terms of linearity, shift-invariance and aliasing, can be obtained even for the case of a limited number of views. Considerations of projection degradations are then analyzed according to a specific imaging device to derive the optimum oversampling ratio.

Inter- and intrasubject variability of solid and liquid gastric emptying parameters. A scintigraphic study in healthy subjects and diabetic patients.

This study was performed to assess: (1) the inter- and intrasubject variability of gastric emptying measurements by a scintigraphic method in 12 healthy subjects and 14 diabetic patients, and (2) the reproducibility of diagnosis of either the presence or absence of gastroparesis. To address this issue, radiolabeled solid-liquid meals were ingested by all subjects on two separate days. High intersubject variability of gastric emptying measurements was shown in both populations. Intrasubject variability was higher in diabetic patients than in healthy subjects. However, there was no significant difference between the means of any parameters obtained on two separate days. The reproducibility of the diagnosis of gastroparesis was excellent for all parameters; the solid half emptying time was the most reproducible parameter (92% in healthy subjects and 93% in diabetic patients). The means +/- SD of the difference between the two separate days' half emptying time results were -10.6 +/- 41.3 min for solids and -4.8 +/- 36.6 min for liquids and were not statistically different from zero. Only one difference between the half emptying time results on two separate days was not in the 95% confidence interval for both solids and liquids; however, this result came from a diabetic patient with obvious gastroparesis on both days. Thus: (1) in spite of high inter- and intrasubject variability, the scintigraphic method of measuring gastric emptying is highly reproducible for the diagnosis of gastroparesis; (2) reproducibility is better in healthy subjects than in diabetic patients; and (3) the knowledge of intrasubject variability allows assessment of the required sample sizes for pharmacological studies using prokinetic medications.

Simultaneous assessment of liquid emptying and proximal gastric tone in humans.

BACKGROUND: Barostat is the only technique that allows assessment of gastric tone in humans. Our aim was to simultaneously assess gastric emptying and relaxation in response to a liquid meal. METHODS: Gastric tone was monitored using an electronic barostat in six healthy subjects after three liquid meals (200 mL, 400 mL, 600 mL, 1 kcal/mL). Scintigraphic imaging was obtained by using double isotopic labeling (technetium 99m for liquid of the 200 mL meal and xenon 133 for air into the barostat). RESULTS: Profound gastric relaxation was detected in every subject. The duration of proximal gastric relaxation increased with meal size. The proximal stomach remained relaxed through the duration of gastric emptying. Gastric tone returned to the fasting values simultaneously with completion of liquid emptying. Repeated measures after the 200-mL meal showed that amplitude (i.e., maximal volume change) and duration of relaxations were reproducible. However, the presence of the bag slightly accelerated gastric emptying and modified the intragastric distribution of the meal. CONCLUSIONS: The barostat is a sensitive and reproducible technique to measure gastric relaxation following liquid meals in humans. The results also suggest that the role of gastric tone as the driving force of gastric emptying of liquids has been overestimated.

Iterative reconstruction methods for nonuniform attenuation distribution in SPECT.

Two iterative methods, a generalization of the Chang method and a projection precorrection, were investigated to determine whether the use of an attenuation map could improve nonuniform attenuation compensation. After a detailed description of the methods, results obtained with simulated and phantom data were compared. This study demonstrates that projection precorrection provides accurate quantification and good image quality as early as the precorrection step, whereas the generalized Chang method requires computation of one more iteration.

How does iliac crest bone marrow biopsy compare with imaging in the detection of bone metastases in small cell lung cancer?

Iliac crest bone marrow biopsy (BMB) has often been used as the gold standard for the detection of bone marrow metastases in small cell lung cancer (SCLC). However, it is likely to lead to numerous false-negative results. For this reason, we compared the results of bone scintigraphy (BS), magnetic resonance imaging (MRI), and BMB in 48 sequential patients affected with pathologically confirmed SCLC (47 were evaluable; mean age, 58.4 years). The three procedures were carried out within 1 week, no treatment being performed during this period. Whole-body scans and spot views were obtained in the anterior and posterior projections. For MRI, only the thoracolumbar spine, the sternum and the pelvis were scanned, using spin-echo T1-weighted sequences, resulting in an acquisition time of less than 45 min. Only five BMBs were rated as positive. In these cases, both BS and MRI were also positive. The other 42 biopsies were negative. Among them, in ten cases both BS and MRI were positive. In 21 cases, both BS and MRI were negative. In five cases MRI was positive while BS was negative. Finally, in six cases MRI was negative whilst BS was positive. In most cases in which either BS or MRI was positive, follow-up scans confirmed the initial findings. This study suggests that BMB is more invasive and less sensitive than BS or MRI in detecting bone metastases. MRI seems to be more sensitive than BS in detecting small spinal or pelvic metastases. Whole-body bone scintigraphy is more sensitive in detecting skull, costal or peripheral metastases.(ABSTRACT TRUNCATED AT 250 WORDS)

[Electrogastrography: responses to food and sham feeding and relationships with gastric emptying of solids in men]. / Electrogastrographie: réponses à l'alimentation et au repas fictif et relations avec la vidange gastrique des solides chez l'homme.

Cutaneous electrogastrography is a non invasive method to study gastric electromechanical activity. The aim of this work was to determine the amplitude of electrogastrographic (EGG) activity a) during fasting, b) after a meal, c) following vagal stimulation by sham feeding, and to determine the relationship between gastric emptying of solids and EGG activity. EGG activity was recorded in eight healthy subjects in various experimental conditions, twice after sham feeding, twice after a meal, and once during the simultaneous scintigraphic assessment of gastric emptying of the solid component of a meal. During one of the sham feeding tests, subjects were intubated and acid secretion was measured. The EGG signal amplitude was continuously monitored and an intercorrelation function (IF) was calculated using the Fast Fourier Transforms of electrical activity recorded by 2 cutaneous electrodes placed on the epigastric area. During fasting, IF was usually of low amplitude with occasional short increases of amplitude. Sham feeding without intubation rarely induced an early and brief increase in IF amplitude (2 of 8 subjects). Sham feeding-induced acid secretion was negatively correlated with IF amplitude (r = 0.78, P = 0.02) suggesting a motor inhibition associated with vagally stimulated acid secretion. When given orally, meals induced an increase in IF amplitude, but there was major intra- and interindividual variations. There was no significant correlation between the IF increase and the half emptying time for solids (r = 0.62, P = 0.10). This study shows that the high variability of EGG activity during fasting considerably hampers the analysis of changes induced by any stimulus.(ABSTRACT TRUNCATED AT 250 WORDS)