Diagnostic Accuracy of Combined FP-CIT, IBZM, and MIBG Scintigraphy in the Differential Diagnosis of Degenerative Parkinsonism: A Multidimensional Statistical Approach

In vivo molecular imaging of pre- and postsynaptic nigrostriatal neuronal degeneration and sympathetic cardiac innervation with SPECT is used to distinguish idiopathic Parkinson disease (PD) from atypical parkinsonian disorder (APD). However, the diagnostic accuracy of these imaging approaches as stand-alone procedures is often unsatisfying. The aim of this study was therefore to evaluate to which extent diagnostic accuracy can be increased by their combined use together with a multidimensional statistical algorithm. Methods: The SPECT radiotracers (123)I-(S)-2-hydroxy-3-iodo-6-methoxy-N-[1-ethyl-2-pyrrodinyl)-methyl]benzamide (IBZM), (123)I-N-omega-fluoropropyl-2 beta-carbomethoxy-3 beta-(4-iodophenyl)nortropan (FP-CIT), and meta-(123)I-iodobenzylguanidine (MIBG) were used to assess striatal postsynaptic D(2) receptor binding, striatal presynaptic dopamine transporter binding, and myocardial adrenergic innervation, respectively. Thirty-one PD and 17 APD patients were prospectively investigated. PD and APD diagnoses were established using consensus criteria and reevaluated after 37.4 +/- 12.4 and 26 +/- 11.6 mo in PD and APD, respectively. Test accuracy (TA) for PD-APD differentiation was computed for all logical (Boolean) combinations of imaging modalities by receiver-operating-characteristic analysis-that is, after multidimensional optimization of cutoff values. Results: Analysis showed moderate TA for PD-APD differentiation using each molecular approach alone (IBZM, 79%; MIBG, 73%; and FP-CIT, 73%). For combined use, the highest TA resulted under the assumption that at least 2 of the 3 biologic markers had to be positive for APD using the following cutoff values: 1.46 or less for IBZM, less than 2.10 for FP-CIT, and greater than 1.43 for MIBG. This algorithm distinguished APD from PD with a sensitivity of 94%, specificity of 94% (TA, 94%), positive predictive value of 89%, and negative predictive value of 97%. Conclusion: Results suggest that the multidimensional combination of FP-CIT, IBZM, and MIBG scintigraphy is likely to significantly increase TA in differentiating PD from APD. The differential diagnosis of degenerative parkinsonism may thus be facilitated.