123I-N-ω-fluoropropyl-2β-carbomethoxy-3β-(4-iodophenyl)nortropan (123I-FP-CIT) single photon emission computed tomography (SPECT) can be evaluated by both visual assessment and quantitative analysis to assess the striatal dopamine state in vivo. The aim of our study was to investigate if visual assessment according to a predefined image grading scale reflects the results of quantitative assessment techniques.
Patients, methods: 195 patients with a clinical diagnosis of idiopathic Parkinson's disease (n = 134), atypical parkinsonian syndrome (n = 47) or essential tremor (n = 14) were examined with 123I-FP-CIT SPECT and included in this retrospective study. Results were analysed according to predefined visual patterns of dopaminergic degeneration and graded as normal (grade 5) or abnormal (grade 1-4) independently by three raters. Quantitative two-dimensional (2D) operator-dependent, manual and three-dimensional (3D) operator-independent, automated approaches were used for quantitative analysis of the specific 123I-FP-CIT tracer binding ratio (SBR) for caudate and putamen.
Results: Sensitivity, specificity, PPV, NPV and diagnostic accuracy of visual assessment of 123I-FP-CIT SPECT for the diagnosis of a neurodegenerative Parkinson's syndrome were 99%, 86%, 99%, 86% and 98%, respectively. Visual assessment and quantitative analysis agreed well in evaluating the degree of dopaminergic degeneration. Significant differences (p<0.001) were found between degeneration patterns. Only between the so-called eagle wing degeneration and the normal pattern no significant differences in SBR caudate and putamen were found, neither by the quantitative manual (p = 1.00; p = 0.196) nor by the quantitative automated method (p=1.0; p = 0.785). Inter-rater agreement for visual assessment was substantial for all possible pairs of the three raters (κ = 0.70 to 0.74). Strong correlations were observed between the quantitative manual and quantitative automated methods for quantification of SBR caudatum (r = 0.920, r² = 0.846, p<0.001) and SBR putamen (r = 0.908, r²=0.824, p < 0.001).
Conclusion: Visual assessment was highly consistent with the results obtained by quantitative analysis and showed a substantial inter-rater agreement between experienced and inexperienced raters. Our findings indicate that visual assessment might be a reliable analysis approach for clinical routine.