The primary finding of this study is that DAT-SPECT with the brain MPH collimator improves intra- and between-rater agreement, and the certainty of the visual interpretation of the SPECT images compared to DAT-SPECT with a conventional LEHRHS collimator (Fig. 3). The improvement of intra-rater agreement was observed in each individual rater (supplementary Fig. 1), suggesting that it is not restricted to less experienced raters (after clinical reading of about 200 DAT-SPECT), but also experienced (about 2000 DAT-SPECT) and very experienced (about 5000 DAT-SPECT) raters benefit from the improved image quality.
In clinical routine, interpretation of DAT-SPECT is restricted to the binary decision as ‘reduced’ (indicating nigrostriatal degeneration) or ‘normal’ (indicating a secondary parkinsonian syndrome without nigrostriatal degeneration). It is particularly relevant, therefore, that the dichotomized Likert 6-score showed perfect intra- and between-rater agreement for each individual rater and for each pair of raters with the MPH collimators only (supplementary Figure S1). With the LEHRHS-SPECT images, the raters still achieved very good intra-rater and between-rater agreement (kappa about 0.95 and about 0.90, respectively), but not perfect. There was no clear difference between LEHRHS-OSEM and LEHRHS-FBP to favor one over the other with respect to intra- and between-rater agreement of the visual interpretation, although iterative reconstruction with attenuation and scatter correction and resolution recovery provided better image quality (Fig. 1), in line with previous reports .
The improvement of rater confidence in the visual interpretation of DAT-SPECT by MPH collimators was mainly driven by improved confidence in excluding nigrostriatal degeneration (Fig. 4, supplementary Figure S2). This might be explained by the fact that DAT loss associated with nigrostriatal degeneration in neurodegenerative parkinsonian syndromes typically starts in the posterior part of the putamen, which is thinner than its anterior part. Thus, the posterior part of the putamen is particularly prone to partial volume effects caused by limited spatial resolution in the SPECT images. Improved spatial resolution of MPH-SPECT results in less pronounced partial volume effects and, consequently, more reliable depiction of the posterior putamen (Fig. 6). The impact of MPH-SPECT on the confidence of the detection of nigrostriatal degeneration was small (Fig. 4, supplementary Figure S2).
We hypothesize that the improvement of image quality with respect to both spatial resolution and statistical noise by MPH-SPECT also improves the confidence of the referring neurologist in DAT-SPECT findings, because the improved image quality with better anatomical delineation of the striatum and its substructures (i) makes the SPECT images more appealing for readers more familiar with structural MRI and (ii) makes it easier for non-experts to comprehend their interpretation by the nuclear medicine physician.
Improved spatial resolution of MPH-SPECT resulted in considerably higher estimates of the putamen SBR compared to LEHRHS-SPECT (Fig. 5). However, the effect was similar in normal and reduced DAT-SPECT so that the effect size of the difference between normal and reduced DAT-SPECT was practically the same in all settings. Thus, the impact of MPH collimators on semi-quantitative analysis in DAT-SPECT most likely is small. This might be explained by the fact that averaging intensities of a rather large number of voxels in the striatal region-of-interest (or in the reference region) levels out differences with respect to spatial resolution and statistical noise to large extent.
However, there are early signs of Parkinson’s disease such as smell loss and rapid eye movement sleep and behavior disorder that precede motor symptoms. As soon as disease-modifying drugs for treatment of Parkinson’s disease will be available, it will be important to detect Parkinson’s disease also in these early stages, when the loss of (unilateral) putaminal DAT is considerably below the 50% threshold for the occurrence of motor symptoms [27–29]. MPH collimators might be particularly useful to improve sensitivity and specificity for the detection of mild nigrostriatal degeneration at early (pre-motor) stages of the disease.
The following limitations of the study should be noted. First, potential impact of MPH-SPECT on the diagnostic accuracy could not be assessed, because there was no independent standard-of-truth diagnosis available. All included patients were referred to DAT-SPECT to support the etiological diagnosis of a clinically uncertain parkinsonian syndrome. Thus, a sufficiently reliable clinical diagnosis was not available. However, amongst the 67 of the 71 patients in which the three raters agreed with respect to the binary classification of the iteratively reconstructed LEHRHS-SPECT images there was no one with discrepant binary classification from the MPH-SPECT images. This suggests that the impact of the brain MPH collimators on the binary classification of DAT-SPECT is small. Second, the patient sample was slightly unbalanced with respect to the ordering of the SPECT acquisition. MPH-SPECT was performed first in about two thirds of the cases. As a consequence, there might have been a bias in favor of MPH-SPECT, as head motion might have been more frequent during the second acquisition due to possible exhaustion of patients. The impact of the difference in the delay between injection of 123I-FP-CIT and the start of the acquisition between MPH-SPECT and LEHRHS-SPECT most likely was small, since all acquisitions were performed within the recommended time window 3–6 h after injection [4, 30]. Finally, practice guidelines on DAT-SPECT recommend the use of fan-beam collimators , which provide a 20–40% gain in count sensitivity at similar spatial resolution compared to LEHRHS collimators [7, 31]. There are no fan-beam collimators available for the triple head SPECT system used in this study. We decided not to use a different SPECT camera equipped with fan-beam collimators in order to avoid ‘contamination’ of collimator effects by camera effects.
In conclusion, MPH collimators improve intra- and between-rater agreement as well as the certainty of the visual interpretation in DAT-SPECT, particularly for the exclusion of nigrostriatal degeneration, and therefore can be recommended for routine clinical use.