Purpose
Computed tomography (CT)âbased attenuation correction (CTAC) in singleâphoton emission computed tomography (SPECT) is highly accurate, but it requires hybrid SPECT/CT instruments and additional radiation exposure. To obtain attenuation correction (AC) without the need for additional CT images, a deep learning method was used to generate pseudoâCT images has previously been reported, but it is limited because of crossâmodality transformation, resulting in misalignment and modalityâspecific artifacts. This study aimed to develop a deep learningâbased approach using nonâattenuationâcorrected (NAC) images and CTACâbased images for training to yield AC images in brainâperfusion SPECT. This study also investigated whether the proposed approach is superior to conventional Changâs AC (ChangAC).
Methods
In total, 236 patients who underwent brainâperfusion SPECT were randomly divided into two groups: the training group (189 patients; 80%) and the test group (47 patients; 20%). Two models were constructed using Autoencoder (AutoencoderAC) and UâNet (UâNetAC), respectively. ChangAC, AutoencoderAC, and UâNetAC approaches were compared with CTAC using qualitative analysis (visual evaluation) and quantitative analysis (normalized mean squared error [NMSE] and the percentage error in each brain region). Statistical analyses were performed using the Wilcoxon signedârank sum test and BlandâAltman analysis.
Results
UâNetAC had the highest visual evaluation score. The NMSE results for the UâNetAC were the lowest, followed by AutoencoderAC and ChangAC (PÂ <Â 0.001). BlandâAltman analysis showed a fixed bias for ChangAC and AutoencoderAC and a proportional bias for ChangAC. ChangAC underestimated counts by 30â40% in all brain regions. AutoencoderAC and UâNetAC produced mean errors of <1% and maximum errors of 3%, respectively.
Conclusion
New deep learningâbased AC methods for AutoencoderAC and UâNetAC were developed. Their accuracy was higher than that obtained by ChangAC. UâNetAC exhibited higher qualitative and quantitative accuracy than AutoencoderAC. We generated highly accurate AC images directly from NAC images without the need for intermediate pseudoâCT images. To verify our modelsâ generalizability, external validation is required.