Sensitivity and specificity of dopamine transporter imaging with 123I-FP-CIT SPECT in dementia with Lewy bodies: a phase III, multicentre study

Background Dementia with Lewy bodies (DLB) needs to be distinguished from other types of dementia because of important differences in patient management and outcome. Current clinically based diagnostic criteria for DLB have limited accuracy. Severe nigrostriatal dopaminergic degeneration occurs in DLB, but not in Alzheimer’s disease or most other dementia subtypes, offering a potential system for a biological diagnostic marker. The primary aim of this study was to investigate the sensitivity and specificity, in the ante-mortem differentiation of probable DLB from other causes of dementia, of single photon emission computed tomography (SPECT) brain imaging with the ligand 123I-2β- carbometoxy-3β-(4-iodophenyl)-N-(3-fluoropropyl) nortropane (123I-FP-CIT), which binds to the dopamine transporter (DAT) reuptake site. Diagnostic accuracy, positive and negative predictive values, and inter-reader agreement were the secondary endpoints and a subgroup of possible DLB patients was also included. Methods We did a phase III study in which we used a 123I-FP-CIT SPECT scan to assess 326 patients with clinical diagnoses of probable (n=94) or possible (n=57) DLB or non-DLB dementia (n=147) established by a consensus panel (in 28 patients no diagnosis could be made). Three readers, unaware of the clinical diagnosis, classified the images as normal or abnormal by visual inspection. The study had 90% power to detect the differences between our anticipated sensitivity (0·80) and specificity (0·85) targets and prespecified lower thresholds (sensitivity 0·65, specificity 0·73) using one-sided binomial tests with a significance level of α=0·025. Findings Abnormal scans had a mean sensitivity of 77·7% for detecting clinical probable DLB, with specificity of 90·4% for excluding non-DLB dementia, which was predominantly due to Alzheimer’s disease. A mean value of 85·7% was achieved for overall diagnostic accuracy, 82·4% for positive predictive value, and 87·5% for negative predictive value. Inter-reader agreement for rating scans as normal or abnormal was high (Cohen’s κ=0·87). The procedure was well tolerated with few adverse events. Interpretation A revision of the International Consensus Criteria for DLB has recommended that low DAT uptake in the basal ganglia, as shown by SPECT or PET imaging, be a suggestive feature for diagnosis. Our findings confirm the high correlation between abnormal (low binding) DAT activity measured with 123I-FP-CIT SPECT and a clinical diagnosis of probable DLB. The diagnostic accuracy is sufficiently high for this technique to be clinically useful in distinguishing DLB from Alzheimer’s disease.