Comparison of visual interpretation of [I-123] FP-CIT SPECT scans versus reference-based quantitative analysis utilizing a Japanese normal database.

Abstract

Objective: Dopamine transporter single-photon emission computed tomography (DAT-SPECT) plays an important role in diagnosing parkinsonism. Recently, a reference-based quantitative analysis utilizing a Japanese normal database for DAT-SPECT was developed. This study aimed to investigate the frequency and trends of cases wherein the analysis- and physician-based diagnoses diverged.

Methods: Two physicians performed an interpretation task twice on 195 DAT-SPECT scans. After assessing intra- and intertester agreements, disagreements were resolved by consensus. For the reference-based quantitative analysis, the calibrated specific binding ratio (cSBR), calibrated asymmetry index (cAI), and Z-scores were measured. Images were grouped according to physician consensus and the negative-positive difference from thresholds (Z-score of less than -2.0 and/or cAI of more than 12.22) as follows: group 1 (physician, normal; quantitative analysis, normal; n  = 70), group 2 (abnormal; normal; n  = 4), group 3 (normal; abnormal; n  = 31), and group 4 (abnormal; abnormal; n  = 90).

Results: Median cSBRs and Z-scores decreased in order from group 1 to group 4. Median cAI values increased in the order of groups 1, 3, 2, and 4. Significant differences were observed between groups 1 and 2 for cSBRs and cAIs; groups 2 and 3 for Z-scores; groups 2 and 4 for cSBRs and Z-scores; and groups 1 and 3, 1 and 4, and 3 and 4 for all parameters (Kruskal-Wallis and Steel-Dwass tests).

Conclusion: In approximately 18% of cases, the visual interpretation of physicians diverged from the reference-based quantitative analysis based on a Japanese normal database. It is crucial to appropriately utilize DAT-SPECT reference-based quantitative analysis as a diagnostic aid.