Verification of image quality improvement of low-count bone scintigraphy using deep learning.

To improve image quality for low-count bone scintigraphy using deep learning and evaluate their clinical applicability. Six hundred patients (training, 500; validation, 50; evaluation, 50) were included in this study. Low-count original images (75%, 50%, 25%, 10%, and 5% counts) were generated from reference images (100% counts) using Poisson resampling. Output (DL-filtered) images were obtained after training with U-Net using reference images as teacher data. Gaussian-filtered images were generated for comparison. Peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) to the reference image were calculated to determine image quality. Artificial neural network (ANN) value, bone scan index (BSI), and number of hotspots (Hs) were computed using BONENAVI analysis to assess diagnostic performance. Accuracy of bone metastasis detection and area under the curve (AUC) were calculated. PSNR and SSIM for DL-filtered images were highest in all count percentages. BONENAVI analysis values for DL-filtered images did not differ significantly, regardless of the presence or absence of bone metastases. BONENAVI analysis values for original and Gaussian-filtered images differed significantly at ≦25% counts in patients without bone metastases. In patients with bone metastases, BSI and Hs for original and Gaussian-filtered images differed significantly at ≦10% counts, whereas ANN values did not. The accuracy of bone metastasis detection was highest for DL-filtered images in all count percentages; the AUC did not differ significantly. The deep learning method improved image quality and bone metastasis detection accuracy for low-count bone scintigraphy, suggesting its clinical applicability.

Optimization of the Attenuation Coefficient for Chang Attenuation Correction in 123I Brain Perfusion SPECT.

N-isopropyl-p-123I-iodoamphetamine brain perfusion SPECT has been used with various attenuation coefficients (µ-values); however, optimization is required. This study aimed to determine the optimal µ-value (µopt-value) for Chang attenuation correction (AC) using clinical data by comparing the Chang method and CT-based AC. Methods: We used 100 patients (reference group, 60; disease group, 40) who underwent N-isopropyl-p-123I-iodoamphetamine SPECT. SPECT images of the reference group were obtained to calculate the AC using the Chang method (µ-values, 0.07-0.20; 0.005 interval) and the CT-based method, both without scatter correction (SC) and with SC. The µopt-value with the smallest mean percentage error for the brain regions of the reference group was calculated. Agreement between the Chang and CT-based methods applying the µopt-value was evaluated using Bland-Altman analysis. Additionally, the percentage error in the region of hypoperfusion in the diseased group was compared with the percentage error in the same region in the reference group when the µopt-value was applied. Results: The µopt-values were 0.140 for Chang without SC and 0.160 for Chang with SC. In the Chang method, with the µopt-value applied, fixed and proportional biases were observed in the Bland-Altman analysis (both P < 0.05), and there was a tendency for the percentage error to be underestimated in the limbic regions and overestimated in the central brain regions. There was no significant difference between the disease group and the reference group in the region of hypoperfusion in either Chang without SC or Chang with SC. Conclusion: The present study revealed that the µopt-values of the Chang method are 0.140 without SC and 0.160 with SC.