High resolution reconstruction of PET images using the iterative OSEM algorithm

3D-Segmentation of the 18F-choline PET Signal for Target Volume Definition in Radiation Therapy of the Prostate

Volumetric assessment of PET signals becomes increasingly relevant for radiotherapy (RT) planning. Here, we investigate the utility of 18F-choline PET signals to serve as a structure for semi-automatic segmentation for forward treatment planning of prostate cancer. 18F-choline PET and CT scans of ten patients with histologically proven prostate cancer without extracapsular growth were acquired using a combined PET/CT scanner. Target volumes were manually delineated on CT images using standard software. Volumes were also obtained from 18F-choline PET images using an asymmetrical segmentation algorithm. PTVs were derived from CT 18F-choline PET based clinical target volumes (CTVs) by automatic expansion and comparative planning was performed. As a read-out for dose given to non-target structures, dose to the rectal wall was assessed. Planning target volumes (PTVs) derived from CT and 18F-choline PET yielded comparable results. Optimal matching of CT and 18F-choline PET derived volumes in the lateral and cranial-caudal directions was obtained using a background-subtracted signal thresholds of 23.0+/-2.6%. In antero-posterior direction, where adaptation compensating for rectal signal overflow was required, optimal matching was achieved with a threshold of 49.5+/-4.6%. 3D-conformal planning with CT or 18F-choline PET resulted in comparable doses to the rectal wall. Choline PET signals of the prostate provide adequate spatial information amendable to standardized asymmetrical region growing algorithms for PET-based target volume definition for external beam RT.

Clinical evaluation of PET image reconstruction using a spatial resolution model

PURPOSE

PET image resolution is variable across the measured field-of-view and described by the point spread function (PSF). When accounting for the PSF during PET image reconstruction image resolution is improved and partial volume effects are reduced. Here, we evaluate the effect of PSF-based reconstruction on lesion quantification in routine clinical whole-body (WB) PET/CT imaging.

MATERIALS AND METHODS

41 oncology patients were referred for a WB-PET/CT examination (Biograph 40 TruePoint). Emission data were acquired at 2.5 min/bed at 1 hpi of 400 MBq [18F]-FDG. Attenuation-corrected PET images were reconstructed on 336 × 336-matrices using: (R1) standard AW-OSEM (4 iter, 8 subsets, 4 mm Gaussian) and (R2) AW-OSEM with PSF (3 iter, 21 subsets, 2 mm). Blinded and randomised reading of R1- and R2-PET images was performed. Individual lesions were located and counted independently on both sets of images. The relative change in PET quantification (SUVmax, SUVmean, volume) of lesions seen on R1 and R2 is reported as (R2-R1)/R1. Furthermore, SUVmax and SUVmean was measured for a 3 cm spherical norm region in the right lobe of the healthy liver for R1 and R2.

RESULTS

Clinical reading revealed 91 and 103 positive lesions for R1 and R2, respectively. For all lesions SUVmax (R2) was higher than SUVmax (R1). Regression analysis indicated that the relative increase in SUVmax (and SUVmean) decreased with lesion size, whilst it increased with increasing radial distance from the centre of the field of view (FOV). There was no significant difference in SUVmean in homogenous liver tissue between R1 and R2.

CONCLUSION

In whole-body FDG-PET/CT using routine clinical protocols, PSF-based PET reconstruction increases lesion detection and affects SUVmax measurements compared to standard AW-OSEM PET reconstruction.

Small animal tumour imaging with MRI and the ECAT EXACT scanner: application of partial volume correction and comparison with microPET data

OBJECTIVE

Partial volume effects caused by limited spatial resolution of conventional positron emission tomography (PET) scanners result in an underestimation of the activity concentration in small tumours. The aim of the study was to evaluate the feasibility of small animal tumour imaging with the clinical PET scanner ECAT EXACT after partial volume correction based on MRI calculations. The same tumour model was examined additionally with the small animal PET system, microPET focus 120.

METHODS

Before the ECAT EXACT studies recovery coefficients for different sphere volumes were generated with phantom experiments. For the following in-vivo study DS-sarcoma cells were implanted on both hind foot dorsum of male Sprague-Dawley rats. In-vivo tumour volume calculations were done with the high-resolution MRI system, Magnetom Vision Experimental. Dynamic F-fluorodeoxyglucose (FDG) PET was performed with the scanner ECAT EXACT (5 MBq intravenous, two-dimensional mode, n = 16 tumours) or with the microPET focus 120 (20 MBq intravenous, two-dimensional mode, n = 10 tumours). The animals were then killed, the tumours rapidly explanted, weighed and homogenized. The concentration of F-FDG was measured with a gamma counter and decay corrected; the ex-vivo F-FDG concentration was compared with the mean tumour activity concentration of the PET data.

RESULTS

Using the ECAT EXACT mean underestimation of actual tumour F-FDG concentration was 35.4%, for partial volume-corrected data this error decreased to 1.7%. In addition, after partial volume correction congruence and linear correlation between the regions of interest-based activity concentration and ex-vivo measurements were excellent (r = 0.98). These results were quite similar to the microPET experiments without partial volume correction: r = 0.99.

CONCLUSION

These data indicate that partial volume correction might allow use of the clinical PET system, ECAT EXACT, for the metabolic assessment of small animal tumours >/=10 mm with sufficient accuracy if no dedicated animal PET is available.

Retrospective analysis of radiation load during examination of cancer patients by (68)Ga-DOTATOC and (18)F-FDG positron emission tomography

We present retrospective analysis of the results of examinations of 1,338 cancer patients by (68)Ga-DOTATOC and (18)F-FDG positron emission and computer-aided tomography. It was shown that complex devices for positron emission and computer-aided tomography provide more informative data than individual methods. The protocol for examination by methods of positron emission and computer-aided tomography in each case is determined by clinical requirements and risk of extra exposure of the patient.

Analytical calculation of volumes-of-intersection for iterative, fully 3-D PET reconstruction

Use of iterative algorithms to reconstruct three-dimensional (3-D) positron emission tomography (PET) data requires the computation of the system probability matrix. The pure geometrical contribution can easily be approximated by the length-of-intersection (LOI) between lines-of-response (LOR) and individual voxels. However, more accurate geometrical projectors are desirable. Therefore, we have developed a fast method for the analytical calculation of the 3-D shape and volume of volumes-of-intersection (VOI). This method provides an alternative robust projector with a uniformly continuous sampling of the image space. The enhanced calculation effort is facilitated by using several speedup techniques. Exploiting intrinsic symmetry relations and the sparseness of the system matrix allows to create an efficiently compressed matrix which can be precomputed and completely stored in memory. In addition, a new voxel addressing scheme has been implemented. This scheme avoids time-consuming symmetry transformations of voxel addresses by using an octant-wise symmetrically ordered field of voxels. The above methods have been applied for a fully 3-D, iterative reconstruction of 3-D sinograms recorded with a Siemens/CTI ECAT HR+ PET scanner. A comparison of the performance of the reconstruction using LOI weighting and VOI weighting is presented.

[Image reconstruction and quantification in emission tomography]

Radionuclide tomographic imaging has substantially benefited from the introduction of statistical image reconstruction. Although the main concepts of these iterative algorithms were published decades ago, their widespread use in clinical routine only became available with faster computers for image processing in the last few years. This article gives an overview of data acquisition and iterative reconstruction in emission tomography, deals with the popular maximum likelihood algorithm, and describes the basics of the maximum a posteriori reconstruction. Prerequisites and corrections necessary for quantification are discussed in the second part of the article on the basis of positron emission tomography. Improvements in technical equipment are expected to stimulate future research into image reconstruction.