J. van den Hoff

An accurate method for correction of head movement in PET

A method is presented to correct positron emission tomography (PET) data for head motion during data acquisition. The method is based on simultaneous acquisition of PET data in list mode and monitoring of the patients head movements with a motion tracking system. According to the measured head motion, the line of response (LOR) of each single detected PET event is spatially transformed, resulting in a spatially fully corrected data set. The basic algorithm for spatial transformation of LORs is based on a number of assumptions which can lead to spatial artifacts and quantitative inaccuracies in the resulting images. These deficiencies are discussed, demonstrated and methods for improvement are presented. Using different kinds of phantoms the validity and accuracy of the correction method is tested and its applicability to human studies is demonstrated as well.

An automatic method for accurate volume delineation of heterogeneous tumors in PET

PURPOSE Accurate volumetric tumor delineation is of increasing importance in radiation treatment planning. Many tumors exhibit only moderate tracer uptake heterogeneity and delineation methods using an adaptive threshold lead to robust results. These methods use a tumor reference value R (e.g., ROI maximum) and the tumor background Bg to compute the volume reproducing threshold. This threshold corresponds to an isocontour which defines the tumor boundary. However, the boundaries of strongly heterogeneous tumors can not be described by an isocontour anymore and therefore conventional threshold methods are not suitable for accurate delineation. The aim of this work is the development and validation of a delineation method for heterogeneous tumors. METHODS The new method (voxel-specific threshold method, VTM) can be considered as an extension of an adaptive threshold method (lesion-specific threshold method, LTM), where instead of a lesion-specific threshold for the whole ROI, a voxel-specific threshold is computed by determining for each voxel Bg and R in the close vicinity of the voxel. The absolute threshold for the considered voxel is then given by Tabs=T×(R-Bg)+Bg, where T=0.39 was determined with phantom measurements. VALIDATION 30 clinical datasets from patients with non-small-cell lung cancer were used to generate 30 realistic anthropomorphic software phantoms of tumors with different heterogeneities and well-known volumes and boundaries. Volume delineation was performed with VTM and LTM and compared with the known lesion volumes and boundaries. RESULTS In contrast to LTM, VTM was able to reproduce the true tumor boundaries accurately, independent of the heterogeneity. The deviation of the determined volume from the true volume was (0.8±4.2)% for VTM and (11.0±16.4)% for LTM. CONCLUSIONS In anthropomorphic software phantoms, the new method leads to promising results and to a clear improvement of volume delineation in comparison to conventional background-corrected thresholding. In the next step, the suitability for clinical routine will be further investigated.

Use of targeted therapy for refractory ALK-positive anaplastic large cell lymphoma as a bridging strategy prior to allogeneic transplantation

Dear Editor, Anaplastic lymphoma kinase (ALK)-positive anaplastic large cell lymphoma (ALCL) that is refractory following salvage therapy has a poor prognosis. Crizotinib is an ALK-specific tyrosine kinase inhibitor that was recently approved by the Food and Drug Administration (FDA) for the treatment of lung cancer associated with ALK gene rearrangements. Impressive response rates were reported using Crizotinib in lung cancer patients with ALK gene rearrangements as well as patients with ALK-positive anaplastic large cell lymphoma [1–3]. Brentuximab Vedotin (SGN-35) is a CD-30 specific monoclonal antibody attached to the antitubulin agent monomethyl auristatin E. Brentuximab is FDA approved for the treatment of relapsed or refractory Hodgkin’s lymphoma and systemic anaplastic large cell lymphoma, inducing tumor regression in a considerable proportion of patients [4]. Immunotherapy using allogeneic stem cell transplantation is also a promising treatment option for patients with lymphoma that has failed first-line therapy [5]. A 29-year-old man with anaplastic large cell lymphoma received six cycles of cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP-21) and reached only a partial response for 1 month. Despite treatment with standard salvage combination chemotherapy regimens (DHAP, Dexa-BEAM) the patient continued to show signs of disease progression—B symptoms, increasing LDH levels, and adenopathy. Positron emission tomography–magnetic resonance imaging (PETMRI) revealed infiltration of cervical, para-aortic and iliac nodes (Fig. 1A). Due to his deteriorating clinical condition and signs of respiratory failure the patient was transferred to intensive care and mechanical ventilation was initiated. A CT scan revealed pulmonary manifestations of lymphoma. Based on the findings of GambacortiPasserini et al. [1] and the lack of further treatment options, our patient was started on Crizotinib as part of a compassionate use program via nasogastric feeding tube. The rapid improvement in clinical status of the patient after started tyrosine kinase inhibitor therapy was impressive. No steroids or other antineoplastic drugs were given at this time. The patient’s B symptoms improved and respiratory function returned within 7 days, accompanied by a significantly improved CTscan. The patient was discharged after 14 days and Crizotinib treatment was continued. R. Ordemann (*) : J. Stöhlmacher : F. Kroschinsky : J. M. Middeke :U. Platzbecker :M. Bornhäuser :G. Ehninger Medical Clinic and Policlinic I, University Hospital, Fetscherstrasse 74, 01307 Dresden, Germany e-mail: rainer.ordemann@uniklinikum-dresden.de

A volume of intersection approach for on-the-fly system matrix calculation in 3D PET image reconstruction.

The aim of this study is the evaluation of on-the-fly volume of intersection computation for systems geometry modelling in 3D PET image reconstruction. For this purpose we propose a simple geometrical model in which the cubic image voxels on the given Cartesian grid are approximated with spheres and the rectangular tubes of response (ToRs) are approximated with cylinders. The model was integrated into a fully 3D list-mode PET reconstruction for performance evaluation. In our model the volume of intersection between a voxel and the ToR is only a function of the impact parameter (the distance between voxel centre to ToR axis) but is independent of the relative orientation of voxel and ToR. This substantially reduces the computational complexity of the system matrix calculation. Based on phantom measurements it was determined that adjusting the diameters of the spherical voxel size and the ToR in such a way that the actual voxel and ToR volumes are conserved leads to the best compromise between high spatial resolution, low noise, and suppression of Gibbs artefacts in the reconstructed images. Phantom as well as clinical datasets from two different PET systems (Siemens ECAT HR(+) and Philips Ingenuity-TF PET/MR) were processed using the developed and the respective vendor-provided (line of intersection related) reconstruction algorithms. A comparison of the reconstructed images demonstrated very good performance of the new approach. The evaluation showed the respective vendor-provided reconstruction algorithms to possess 34-41% lower resolution compared to the developed one while exhibiting comparable noise levels. Contrary to explicit point spread function modelling our model has a simple straight-forward implementation and it should be easy to integrate into existing reconstruction software, making it competitive to other existing resolution recovery techniques.

Locally adaptive filtering for edge preserving noise reduction on images with low SNR in PET

As well known, the signal-to-noise ratio (SNR) of PET images can be considerably low. This is especially true for whole-body examinations of heavy patients, for respiratory-gated studies, and dynamic studies with short frames. In these cases moving average filters (MAF) such as a Gaussian filter are applied in order to achieve an acceptable SNR. Image resolution is, however, considerably reduced by these MAFs. This affects detectability and quantification of small structures. Interesting alternatives to MAFs are non-linear, locally adaptive filters (NLF), which enable noise reduction while preserving sharp edges.

Comment on “Transconvolution and the virtual positron emission tomograph: A new method for cross calibration in quantitative PET/CT imaging” [Med. Phys. 40, 062503 (15pp.) (2013)]: Transconvolution and the virtual positron emission tomograph

To the Editor, With great interest, we have read the recent paper by Prenosil et al.1 entitled “Transconvolution and the virtual positron emission tomograph: A new method for cross calibration in quantitative PET/CT imaging.” The authors present a method (transconvolution) which allows to transform data acquired with different PET scanners exhibiting different spatial resolution and, therefore, different partial volume effects (PVE), to a common, virtual PET scanner. After transformation all data exhibit approximately the same PVE independent of the scanner with which they were originally acquired. The proposed method requires knowledge of the scanner’s point-spread-function (PSF). The authors compute the PSF by deconvoluting images from phantom measurements with hot spheres in a cold background. The authors show that with transconvolution the comparability of measurements performed with different scanners can be significantly improved. This is an important result especially in the context of multicenter trials. We also appreciate that the authors are using spheres in a scatter medium for determination of PSF instead of point sources in air, as is otherwise often done. Extended objects like spheres are obviously closer to the clinical situation than point sources and the non-negligible object dependence of the performed image reconstruction and scatter correction (and the consequent influence on the resulting spatial resolution) is accounted for much better than would be the case for point or line sources in air. In fact, the presented method for computing the PSF is very similar to the method we have published in 20102 where also hot spheres (in cold and warm background) were used. Unfortunately, this is not properly recognized in the present paper.1 Rather, the authors cite our work in Sec. 2.A.2 incorrectly claiming that our paper is concerned with resolution measurements using point sources in air. This is quite misleading and we think this point needs clarification. Actually, in Ref. 2 we generated radial profiles by transforming a sphere and its vicinity to spherical coordinates relative to the sphere’s center. The analytic solution for the radial activity profile of the imaged sphere (resulting from analytical computation of the convolution of the original sphere with an isotropic 3D Gaussian PSF) was then fitted to the measured profiles. The analytic solution has five parameters: signal (true activity within sphere), background level, FWHM of the PSF, and the radius as well as the (cold) wall thickness of the spherical inserts. The wall thickness was fixed to its known value. The remaining four parameters were determined by nonlinear least squares fits. This is of course not exactly the same approach as presented in Ref. 1 where explicit deconvolution was used. However, our fitting procedure, too, directly provides all parameters of the PSF as well as the actual object radius which naturally implies that de facto a deconvolution (separation of object and PSF) is performed (although it is not explicitly denoted in this way in our paper). The only point where our work differs from the approach taken by Prenosil and co-workers1 is the chosen model for the PSF. While we have used an isotropic Gaussian PSF, in Ref. 1 a more general shape is chosen, namely, an anisotropic PSF obtained by convolving a Gaussian with some power of a decaying exponential. We thus think that it is important to put the degree of novelty of the approach chosen for PSF determination in the present work into proper perspective. Taking into account this necessary clarification, we feel that the authors altogether have performed an important investigation. However, we are not convinced that the presented results are sufficient to adequately support the rather far reaching conclusions drawn by the authors. For example, with our approach we were able to determine the radii of the investigated spheres with an accuracy of ≈0.5 mm. This was achieved for six different spheres (diameter: 17–37 mm) and nine different contrast ratios (range: 3.3–∞). The quality of our fits and the very accurate reproduction of the true sphere sizes in our opinion clearly shows that an isotropic Gaussian PSF is a quite good approximation of the actual PSF and that the (undoubtedly present) residual deviations of the PSF from the Gaussian are so small that they seem mostly negligible (at least locally). This finding seems to contradict the position taken in Ref. 1 where it is stated that the exponential term of their PSF model is essential for a good approximation of the actual PSF. However, no data supporting this statement are presented. Notably, no direct comparison of a pure Gaussian PSF and the proposed non-Gaussian PSF was performed. Such a comparison seems mandatory in order to decide whether the new approach is really superior to a Gaussian approach. We, therefore, suggest to repeat the evaluation of the transconvolution method using a purely Gaussian PSF (in which case the method trivially reduces to the fact that the convolution of two Gaussians again produces a Gaussian whose variance (squared standard deviation) is the sum of the variances of the two contributing Gaussians) and compare the results with the results achieved with a non-Gaussian PSF. It would also be instructive to see a presentation of results where a somewhat

PET und Bestrahlungsplanung - technische Aspekte

Integration of PET into radiation treatment planning causes several technical problems, which have not yet been solved satisfactory up to now, but which are currently investigated intensively. First solutions of some aspects such as geometrically correct ROI delineation and the translation of this information into DICOM-RT Structure Sets for import into the radiation treatment planning systems have already been developed, but are not yet generally available. It is to be expected, however, that the dynamic technical development will continue in the next years and lead to solution of the currently persisting problems.