[en] Our conclusions are that for the production of optimum hard copy digital imaging techniques are essential. For routine imaging each image can be correctly exposed and windowed to ensure accurate diagnosis. Digital imaging is ideal in difficult low activity examinations such as gallium-67 studies, labelled monoclonal antibodies or MIBG imaging. The correct choice of matrix size is important. For high information density imaging the 256x256 matrix size with a large field of view camera seems to be optimum for most types of nuclear medicine examinations. In the low information density situation it is probably better to use a 128x128 matrix with some computer smoothing. An algorithm which modulated the intensity of individual pixels based on the average counting rate along the x- or y-axis would help in accentuating small changes in radioactivity. From our experiments in digitising high photon images it is obvious that there should be no edges, lines or empty space visible on the image. To overcome this problem some form of spot wobble is suggested which will only marginally degrade the spacial information on the image. The optimum form of hard copy has yet to be found. So far all forms of paper output have yielded less than satisfactory results. Transparent films appear to be most popular. For this form of output, digital imaging is ideal since the computer can be adjusted so that the end image directly reflects what has been seen on the digital camera monitor. While instant prints are valuable for including in the patients notes, probably the ideal medium is instant hard copy in the form of a transparent image. (orig.)

[en] There is no doubt that for routine imagine the ability to post-process images greatly assists diagnosis and facilitates the production of high quality hard copy. In addition it gives increased confidence in diagnosis of many pathological abnormalities especially in situations where it is difficult to predict what the end image will be like. It is felt that this facility should in the long-term greatly improve clinical diagnostic rates and, therefore, it is expected that future gamma cameras will be built around computers and digital imaging will become as important for production of the images as it is for the processing of dynamic studies. (orig.)

[en] Full text of publication follows. Aim: intravenous administration of Re-186 hydroxyethylidene-diphosphonate (HEDP) is used for metastatic bone pain palliation in hormone refractory prostate cancer patients. Dosimetry for bone seeking radionuclides is challenging due to the complex structure with osteoblastic, osteolytic and mixed lesions. The aim of this study was to perform image-based patient-specific 3D convolution dosimetry to obtain a distribution of the absorbed doses to each lesion and estimate inter- and intra-patient variations. Materials and methods: 28 patients received a fixed 5 GBq activity of Re-186 HEDP followed by peripheral blood stem cell rescue at 14 days in a phase II trial. A FORTE dual-headed gamma camera was used to acquire sequential Single-Photon-Emission Computed Tomography (SPECT) data of the thorax and pelvis area at 1, 4, 24, 48 and 72 hours following administration. The projection data were reconstructed using filtered-back projection and were corrected for attenuation and scatter. Voxelised cumulated activity distributions were obtained with two different methods. First, the scans were co-registered and the time-activity curves were obtained on a voxel-by-voxel basis. Second, the clearance curve was obtained from the mean number of counts in each individual lesion and used to scale the uptake distribution taken at 24 hours. The calibration factors required for image quantification were obtained from a phantom experiment. An in-house developed EGSnrc Monte Carlo code was used for the calculation of dose voxel kernels for soft-tissue and cortical/trabecular bone used to perform convolution dosimetry. Cumulative dose-volume histograms were produced and mean absorbed doses calculated for each spinal and pelvic lesion. Results: preliminary results show that the lesion mean absorbed doses ranged from 25 to 55 Gy when the medium was soft tissue and decreased by 40% if bone was considered. The use of the cumulated activity distribution obtained from the scan acquired at 24 hours following administration reduced the number of artefacts introduced by the registration and voxelised cumulated activity calculations. Conclusion: patient-specific convolution dosimetry calculations show that the absorbed dose to each lesion changes significantly depending on the medium density considered. This suggests that specific lesion and surrounding tissue compositions should be considered to overcome the limitations of convolution dosimetry, which could explain the range of absorbed doses observed. Future work will include the correlation of absorbed dose with patient outcome. (authors)