Artificial Neural Network-Based Prediction of Outcome in Parkinson's Disease Patients Using DaTscan SPECT Imaging Features.

PURPOSE: Quantitative analysis of dopamine transporter (DAT) single-photon emission computed tomography (SPECT) images can enhance diagnostic confidence and improve their potential as a biomarker to monitor the progression of Parkinson's disease (PD). In the present work, we aim to predict motor outcome from baseline DAT SPECT imaging radiomic features and clinical measures using machine learning techniques. PROCEDURES: We designed and trained artificial neural networks (ANNs) to analyze the data from 69 patients within the Parkinson's Progressive Marker Initiative (PPMI) database. The task was to predict the unified PD rating scale (UPDRS) part III motor score in year 4 from 92 imaging features extracted on 12 different regions as well as 6 non-imaging measures at baseline (year 0). We first performed univariate screening (including the adjustment for false discovery) to select 4 regions each having 10 features with significant performance in classifying year 4 motor outcome into two classes of patients (divided by the UPDRS III threshold of 30). The leave-one-out strategy was then applied to train and test the ANNs for individual and combinations of features. The prediction statistics were calculated from 100 rounds of experiments, and the accuracy in appropriate prediction (classification of year 4 outcome) was quantified. RESULTS: Out of the baseline non-imaging features, only the UPDRS III (at year 0) was predictive, while multiple imaging features depicted significance. The different selected features reached a predictive accuracy of 70 % if used individually. Combining the top imaging features from the selected regions significantly improved the prediction accuracy to 75 % (p < 0.01). The combination of imaging features with the year 0 UPDRS III score also improved the prediction accuracy to 75 %. CONCLUSION: This study demonstrated the added predictive value of radiomic features extracted from DAT SPECT images in serving as a biomarker for PD progression tracking.

Data-driven, voxel-based analysis of brain PET images: Application of PCA and LASSO methods to visualize and quantify patterns of neurodegeneration.

Spatial patterns of radiotracer binding in positron emission tomography (PET) images may convey information related to the disease topology. However, this information is not captured by the standard PET image analysis that quantifies the mean radiotracer uptake within a region of interest (ROI). On the other hand, spatial analyses that use more advanced radiomic features may be difficult to interpret. Here we propose an alternative data-driven, voxel-based approach to spatial pattern analysis in brain PET, which can be easily interpreted. We apply principal component analysis (PCA) to identify voxel covariance patterns, and optimally combine several patterns using the least absolute shrinkage and selection operator (LASSO). The resulting models predict clinical disease metrics from raw voxel values, allowing for inclusion of clinical covariates. The analysis is performed on high-resolution PET images from healthy controls and subjects affected by Parkinson's disease (PD), acquired with a pre-synaptic and a post-synaptic dopaminergic PET tracer. We demonstrate that PCA identifies robust and tracer-specific binding patterns in sub-cortical brain structures; the patterns evolve as a function of disease progression. Principal component LASSO (PC-LASSO) models of clinical disease metrics achieve higher predictive accuracy compared to the mean tracer binding ratio (BR) alone: the cross-validated test mean squared error of adjusted disease duration (motor impairment score) was 16.3 ± 0.17 years2 (9.7 ± 0.15) with mean BR, versus 14.4 ± 0.18 years2 (8.9 ± 0.16) with PC-LASSO. We interpret the best-performing PC-LASSO models in the spatial sense and discuss them with reference to the PD pathology and somatotopic organization of the striatum. PC-LASSO is thus shown to be a useful method to analyze clinically-relevant tracer binding patterns, and to construct interpretable, imaging-based predictive models of clinical metrics.

Improved prediction of outcome in Parkinson's disease using radiomics analysis of longitudinal DAT SPECT images.

No disease modifying therapies for Parkinson's disease (PD) have been found effective to date. To properly power clinical trials for discovery of such therapies, the ability to predict outcome in PD is critical, and there is a significant need for discovery of prognostic biomarkers of PD. Dopamine transporter (DAT) SPECT imaging is widely used for diagnostic purposes in PD. In the present work, we aimed to evaluate whether longitudinal DAT SPECT imaging can significantly improve prediction of outcome in PD patients. In particular, we investigated whether radiomics analysis of DAT SPECT images, in addition to use of conventional non-imaging and imaging measures, could be used to predict motor severity at year 4 in PD subjects. We selected 64 PD subjects (38 male, 26 female; age at baseline (year 0): 61.9 ± 7.3, range [46,78]) from the Parkinson's Progressive Marker Initiative (PPMI) database. Inclusion criteria included (i) having had at least 2 SPECT scans at years 0 and 1 acquired on a similar scanner, (ii) having undergone a high-resolution 3 T MRI scan, and (iii) having motor assessment (MDS-UPDRS-III) available in year 4 used as outcome measure. Image analysis included automatic region-of-interest (ROI) extraction on MRI images, registration of SPECT images onto the corresponding MRI images, and extraction of radiomic features. Non-imaging predictors included demographics, disease duration as well as motor and non-motor clinical measures in years 0 and 1. The image predictors included 92 radiomic features extracted from the caudate, putamen, and ventral striatum of DAT SPECT images at years 0 and 1 to quantify heterogeneity and texture in uptake. Random forest (RF) analysis with 5000 trees was used to combine both non-imaging and imaging variables to predict motor outcome (UPDRS-III: 27.3 ± 14.7, range [3,77]). The RF prediction was evaluated using leave-one-out cross-validation. Our results demonstrated that addition of radiomic features to conventional measures significantly improved (p < 0.001) prediction of outcome, reducing the absolute error of predicting MDS-UPDRS-III from 9.00 ± 0.88 to 4.12 ± 0.43. This shows that radiomics analysis of DAT SPECT images has a significant potential towards development of effective prognostic biomarkers in PD.