Comparison of region-of-interest analysis and human observers in the diagnosis of Parkinson's disease using [/sup 99m/Tc]TRODAT-1 and SPECT

Abstract

A large set of patients with Parkinson's Disease (PD), and normal healthy control subjects, were studied using the dopamine transporter tracer [ 99m Tc]TRODAT-1 and single photon emission tomography (SPECT). The sample used in this study was comprised of 81 PD patients (mean age ± SD, 63.4 ± 10.4 y; age range, 39.0 -84.2 y), and 94 healthy controls (mean age ± SD, 61.8 ± 11.0 y; age range, 40.9 -83.3 y). A standardized template containing six regions-of-interest (ROIs) was transposed onto subregions of the right and left basal ganglia. In addition, all images were coregistered to a standard [ 99m Tc]TRODAT-1 template. The three slices corresponding to the striatum were extracted and summed into a single transverse slice. These singleslice images were used in a human observer study, using four experienced investigators. Each observer was asked to rate their confidence that each randomly selected image had the appearance of a healthy control subject, or a patient with PD. The data from the observer study was analyzed using a Receiver Operating Characteristic (ROC) analysis. The area under the ROC curve (AUC) indicated the accuracy with which each observer identified each subject. In a similar manner, the ROI data was also analyzed using a ROC curve. The observer analysis gave a more accurate diagnosis (AUC = 0.93 ± 0.02) than the ROI technique (AUC = 0.90 ± 0.02). This suggests that the human observers are visually acquiring more information from the images than is contained in the quantitative striatal uptake alone. Indeed, it is likely that the pattern of uptake of the tracer is of more significance than the absolute uptake, which has implications for the analysis of radionuclide images using quantitative techniques.

Value of Semiquantitative Analysis for Clinical Reporting of123I-2-β-Carbomethoxy-3β-(4-Iodophenyl)-N-(3-Fluoropropyl)Nortropane SPECT Studies

“…The relative accuracy of diagnosis between visual analysis and a semiquantitative approach was investigated by Acton et al (10), showing that semiquantification using region-of-interest analysis gave diagnostic accuracy comparable to visual analysis. The consistency between visual and semiquantitative assessments has also been investigated, with encouraging results (11-13).…”