Convolutional neural network based attenuation correction for ¹²³I-FP-CIT SPECT with focused striatum imaging

Abstract

SPECT imaging with 123 I-FP-CIT is used for diagnosis of neurodegenerative disorders like Parkinson's disease. Attenuation correction (AC) can be useful for quantitative analysis of 123 I-FP-CIT SPECT. Ideally, AC would be performed based on attenuation maps (m-maps) derived from perfectly registered CT scans. Such m-maps, however, are most times not available and possible errors in image registration can induce quantitative inaccuracies in AC corrected SPECT images. Earlier, we showed that a convolutional neural network (CNN) based approach allows to estimate SPECT-aligned m-maps for full brain perfusion imaging using only emission data. Here we investigate the feasibility of similar CNN methods for axially focused 123 I-FP-CIT scans. We tested our approach on a highresolution multi-pinhole prototype clinical SPECT system in a Monte Carlo simulation study. Three CNNs that estimate m-maps in a voxel-wise, patch-wise and image-wise manner were investigated. As the added value of AC on clinical 123 I-FP-CIT scans is still debatable, the impact of AC was also reported to check in which cases CNN based AC could be beneficial. AC using the ground truth m-maps (GT-AC) and CNN estimated m-maps (CNN-AC) were compared with the case when no AC was done (No-AC). Results show that the effect of using GT-AC versus CNN-AC or No-AC on striatal shape and symmetry is minimal. Specific binding ratios (SBRs) from localized regions show a deviation from GT-AC  2.5% for all three CNN-ACs while No-AC systematically underestimates SBRs by 13.1%. A strong correlation (  r 0.99) was obtained between GT-AC based SBRs and SBRs from CNN-ACs and No-AC. Absolute quantification (in kBq ml −1 ) shows a deviation from GT-AC within 2.2% for all three CNN-ACs and of 71.7% for No-AC. To conclude, all three CNNs show comparable performance in accurate m-map estimation and 123 I-FP-CIT quantification. CNNestimated m-map can be a promising substitute for CT-based m-map.

Soft Attention Based DenseNet Model for Parkinson’s Disease Classification Using SPECT Images

“… Chen et al (2021) put focus on striatum scanning and implemented a CNN model based on attenuation correction. Monte-carlo based simulation results were drawn for a clearer visual assessment based on voxel-wise, patch-wise and image-wise imaging methods.…”

CT-free quantitative SPECT for automatic evaluation of %thyroid uptake based on deep-learning

“…Additionally, the time-consuming manual thyroid segmentation on CT canvas is challenging. In the literature, there are deep-learning-based CT-free AC studies for myocardial perfusion SPECT [ 2 , 5 ], brain perfusion SPECT [ 6 , 7 , 32 ] and dopamine-transporter brain SPECT [ 3 ]. Undoubtedly, AC using CT is essential for quantitative thyroid SPECT/CT, but thyroid-dedicated deep-learning study has not been investigated.…”

Generative adversarial network-based attenuation correction for 99mTc-TRODAT-1 brain SPECT

“…Murata et al ( 21 ) demonstrate that 2D autoencoder and U-Net-based direct DL-AC are better than NAC and Chang's AC for brain perfusion SPECT. Chen et al ( 23 ) suggest that CNN-estimated μ-map could be a promising substitute for CT-based μ-map for 123 I-FP-CIT scans. Our results are consistent with theirs in that DL-based AC is better than NAC and Chang's method.…”

“…Murata et al ( 21 ) compared Chang's AC with a 2D autoencoder and U-Net for DL-AC for brain perfusion SPECT. Chen et al have proposed CNN-based μ-map generation for brain perfusion SPECT ( 22 ) and 123 I-FP-CIT SPECT ( 23 ) using NAC SPECT input in simulations, demonstrating improved absolute quantification accuracy. A diagram explaining DL-AC μ and DL-AC is shown in Figure 1 .…”