PT - JOURNAL ARTICLE AU - Christopher Leatherday AU - Jan Boucek AU - Michael McCarthy TI - Comparison of Methods for Striatal/Occipital FDOPA PET Uptake Ratio Formulation DP - 2020 May 01 TA - Journal of Nuclear Medicine PG - 152--152 VI - 61 IP - supplement 1 4099 - http://jnm.snmjournals.org/content/61/supplement_1/152.short 4100 - http://jnm.snmjournals.org/content/61/supplement_1/152.full SO - J Nucl Med2020 May 01; 61 AB - 152Introduction: 18F-Fluorodopa (FDOPA) PET imaging can accurately reflect the degeneration of dopaminergic receptors that occurs in the basal ganglia of patients with Parkinson’s disease (PD)1. One option for assessment is the striatal/occipital uptake ratio (SOR), which has been shown to have utility in delineating between PD patients and healthy controls. The methodology of defining ROIs can vary depending on the availability of complementary imaging; previous work has shown that manually marking ROIs on a PET-fused T1 weighted MRI improves diagnostic accuracy over using PET alone2. We compared SOR values obtained from manually defined ROIs on fused PET/CT or MRI with values from automated volumetric methods utilizing PET only or MRI-based segmentation. Methods: Analysis was conducted on 12 patients with suspected PD who were referred for FDOPA Brain PET imaging to either Sir Charles Gairdner or Fiona Stanley hospital and had a suitable MRI (T1 weighted head with resolution ≤1mm isotropic). Four different SOR quantification methods were compared, two manual and two automated. The manual methods have been previously defined; ROIs are drawn on a single axial slice of either the fused PET/CT or PET/MR2. Neurostat3 and FreeSurfer4 were used to implement PET and MRI-based volumetric segmentation methods respectively. Figure 1 shows examples of the MRI-based and manual and marking methods. Results: Figure 2 shows the comparative SOR values for each of the methods for each subject. Approximate ranges for healthy controls, early stage PD , and advanced PD reported elsewhere5 are overlaid on the graph. There is enough variation between methods for a third of the patients to change their diagnostic category (5, 7, 9, 10). Figure 3 shows the values for each method after subtracting the mean for each patient. Tukey’s Honestly Significant Difference test found that the values from manual CT marking were significantly higher than the Neurostat (p < 0.01) and FreeSurfer methods (p < 0.001). Conclusions: SOR values for several patients in this cohort varied between the tested methods in a clinically significant magnitude. The patient mean-subtracted values from the manual CT-based marking method were significantly higher than the automated methods, which may be due to the lack of soft tissue contrast resulting in the selection of areas containing the highest uptake rather than areas that correspond to tissue structure directly. A longitudinal study that includes clinical PD assessment would be useful for determining if there are significant and systematic differences in diagnostic accuracy between the methods, particularly if used in conjunction with advanced MRI methods such as R2* or quantitative susceptibility mapping for improved basal ganglia segmentation accuracy. References: 1Pavese N, Rivero-Bosch M, Lewis SJ et al. 2011. Neuroimage 56:1463-8 2Struck AF, Hall LT, Kusmirek JE et al. 2012. Am J Nucl Med Mol Imaging 2(4):475-82. 3Minoshima S, Frey KA, Koeppe RA et al. 1995. J Nucl Med 36:1238-48 4Fischl B, Salat DH, Busa E et al. 2002. Neuron 33:341-55. 5Eshuis SA, Jager PL, Maguire RP et al. 2009. Eur J Nucl Med Mol Imaging 36(3):454-62 Figure 1: (Top row) A fused FDOPA PET/MRI showing the occipital grey matter, caudate, and putamen VOIs produced by FreeSurfer segmentation (blue); (Bottom row) A manually marked fused PET/CT and MRI. Figure 2: A comparison of the SOR values for each patient using each of the four methods. The shaded regions indicate the boundaries of the healthy control (green), early stage PD (yellow), and advanced PD (red) groups previously reported5. Figure 3: SOR for each method after subtraction from the patient mean.