Reduction of SPECT acquisition time using deep learning: A phantom study.

Single photon emission computed tomography (SPECT) procedures are characterized by long acquisition time to acquire diagnostically acceptable image data. The goal of this investigation was to assess the feasibility of using a deep convolutional neural network (DCNN) to reduce the acquisition time. The DCNN was implemented using the PyTorch and trained using image data from standard SPECT quality phantoms. The under-sampled image dataset is provided to neural network as input, while missing projections were provided as targets. The network is to produce for the output the missing projections. The baseline method of calculating the missing projections as arithmetic means of adjacent ones was introduced. The obtained synthesized projections and reconstructed images were compared to original data and baseline data across several parameters using PyTorch and PyTorch Image Quality code libraries. Results obtained from comparisons of projection and reconstructed image data show the DCNN clearly outperforming the baseline method. However, subsequent analysis revealed the synthesized image data being more comparable to under-sampled than to fully-sampled image data. The results of this investigation imply that neural network can replicate coarser objects better. However, densely sampled clinical image datasets, coarse reconstruction matrices and patient data featuring coarse structures combined with a lack of baseline data generation methods will hamper the ability to analyse the neural network outputs correctly. This study calls for use of phantom image data and introduction of a baseline method in the evaluation of neural network outputs.

National survey to set diagnostic reference levels in nuclear medicine single photon emission imaging in Croatia.

PURPOSE: In order to introduce the concept of diagnostic reference levels (DRLs) in the national nuclear medicine practice a survey was proposed and completed through all nuclear medicine departments in Croatia. An additional aim was to increase the awareness of importance and full implementation of a comprehensive quality program that includes devices used in the nuclear medicine chain. METHODS: Data were collected for more than 30 nuclear medicine single photon emission procedures. National DRLs (NDRLs) as administered activity and also as administered activity per unit mass were calculated in accordance to International Commission on Radiological Protection (ICRP) recommendations. Additionally, effective doses were estimated using conversion factors published by the ICRP. RESULTS: NDRLs for nuclear medicine single photon emission procedures were proposed. For procedures performed in only one department typical values were presented as reference. Effective doses related to applied radiopharmaceuticals were calculated to estimate radiation risk related to respective nuclear medicine procedure in more detail. CONCLUSION: This work presents results of the first national survey on DRLs of nuclear medicine single photon emission procedures and proposes reliable NDRLs that represent an actual status of nuclear medicine practice in Croatia. Results have motivated departments to introduce and set their own typical values to be used, as one of the tools, for further optimization process. One of the drawbacks of the DRL concept in nuclear medicine is the lack of the image quality parameters involved. For this reason, a quantity that considers both radiation protection and image quality should be introduced.