FP-CIT SPECT evaluation: time to go beyond visual assessment!

Nowadays, FP-CIT single photon emission computed tomography (SPECT) is used in the assessment of patients with extrapyramidal syndrome. Meta-analysis confirmed the role of radionuclide imaging in this setting [1] and further cost-effectiveness studies showed the economic benefits of this diagnostic procedure [2, 3]. Recently, we read with great interest the paper of Badiavas and colleagues [4] published in your journal. The authors presented a stimulating review about SPECT imaging in movement disorders and focused on the semi-quantitative evaluation after referring to tracer kinetic models. They reviewed several semiquantitative methods: manual positioning of anatomical regions of interest (ROIs), an expedient approach to avoid the partial volume effect (PVE), advanced automated quantification methods (in particular BasGan) and voxel-based statistical methods (such as SPM and NEUROSTAT). In conclusion it was remarked that, despite both the need of experience for a correct visual assessment and strong European Association of Nuclear Medicine (EANM) encouragement for semi-quantitative evaluation, no semi-quantitative method has been accepted as a standard. Consequently, qualitative evaluation is still the standard. In the authors’ view, a perfect software should be PC-based, automated (i.e. would analyse raw or mildly processed data), should give reproducible output and should not depend on a specific gamma camera. An ideal software will also perform a comparison with a database of values from normal subjects. BasGan is a PC-based software freely available from the Associazione Italiana di Medicina Nucleare (AIMN) site. It is not dependent on a specific gamma camera (it analyses DICOM images after transaxial reconstruction) and performs a comparison with a database of normal results. Furthermore, the BasGan creators announced at the last EANM congress the final phase of a European project aiming to realize a database with values from healthy subjects acquired from all around the European continent [5]. On the other hand, to the best of our knowledge, no published study evaluated BasGan in a patient population, except for a small population study performed by the above-mentioned group. So, the utility of BasGan needs to be confirmed in a larger population from multicentre trials. These should validate normal references and demonstrate their usefulness to correctly establish the presence of disease. Indeed, BasGan should be useful to improve FP-CIT SPECT reports in different centres, from several countries, acquired with distinct gamma camera models and reconstructed with diverse processing applications. In our institution we evaluated the role of BasGan in 59 consecutive patients (mean age 68 years, 32 men and 27 women). FP-CIT SPECTs obtained in 2006 and 2007 were analysed with the above-mentioned software and then correlated with final diagnosis. Disease (if any) was established by the neurologists in light of all relevant data; however, given the long follow-up (at least 3 years), we think that the diagnosis was independent from the FP-CIT scan report. Results from this software showed good correlation with the final diagnosis. Logistic regression demonstrated a significant correlation between final diagnosis and both caudate values (p<0.001) and putamen values (p<0.001). Receiver-operating characteristic (ROC) curves (accuracies for both caudate and putamen semi-quantitative values) showed high AUC values (range 0.82–0.87,p≤0.0001 for all occurrences) [6]. In conclusion, in our centre, BasGan allows good accuracy in FP-CIT SPECT reporting. This accuracy was not dependent on the operator’s experience and showed a good reproducibility. In the near future, we will study the accuracy and reproducibilityofFP-CITSPECTevaluatedbyBasGanina multicentre trial. This could be the proof of concept for a successful computer-aided report of FP-CIT SPECT