The Influence of Attenuation and Scatter Corrections on Quantitative Analysis of ¹²³I-FP-CIT SPECT Brain Imaging

抄録

Objective: [¹²³I] N-ω-fluoropropyl-2β-carbomethoxy-3β-(4-[¹²³I]iodophenyl) nortropane (¹²³I-FP-CIT) or dopamine transporter (DAT) SPECT scan quantitatively depicts striatum dopamine transporter density. The specific binding ratio (SBR) is important for Parkinson's disease (PD) diagnosis. Correction techniques may improve ¹²³I-FP-CIT image quality, but they are currently not standardized. We aim to determine attenuation correction (AC) and scatter correction (SC) influence on ¹²³I-FP-CIT quantitative analysis.

Methods: We preformed SPECT imaging on a striatal phantom representing striatum-to-background ratios ranging from 10:1 to 1:1. We used the Chang method for AC and triple energy window for SC. Commercially available software semi-automatically calculated SBR. We performed correlation analysis between SBR and the actual concentration ratio in four groups: no correction (NC), SC, AC, and combined correction (CC), paying particular attention to SBR values below the clinical cut-off (4.5). We applied the same imaging conditions, reconstruction, and corrections on 49 human ¹²³I-FP-CIT SPECT studies. Asymmetric index values (AI) were also analyzed. Receiver operating characteristic (ROC) analysis provided AI diagnostic cut-off values in PD patients. PD patient population whose AI values were above these cut-off value were subject to a correlation study to confirm the association between their clinical symptoms and ¹²³I-FP-CIT SPECT.

Results: Phantom study. AC or CC increased SBR ¹²³I-FP-CIT SPECT, making them very close to the standard ratios (AC: y=1.19x - 0.7; R²=0.98; CC: y=1.4x - 1.2; R²=0.98), while NC and SC underestimated SBR at lower values (1 to 4). Clinical study. There was no difference in SBR between NC and SC, while AC and CC produced significantly higher SBR than NC (p<0.0001). Only SC affected AI values (p<0.05). The association between clinical symptoms and ¹²³I-FP-CIT imaging results were maintained only when either NC (n=12, p<0.05) or CC (n=14, p<0.05) was applied.

Conclusions: AC increased SBR more than other corrections, while SC increased AI value more than other corrections. The association between clinical symptom and SPECT findings were maintained when either NC or CC was applied. These correction methods should be carefully selected for ¹²³I-FP-CIT SPECT brain imaging.