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

Abstract

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.

Keywords: DAT SPECT; Longitudinal; Outcome prediction; Parkinson's disease; Radiomics; Textural features.

Fig. 1

3D volume rendering of six segmentations (caudate, putamen and VS; both right and left) for a typical study, as well as transaxial, coronal and sagittal slices through the DAT SPECT image with superimposed segmentations.

Fig. 2

(left) A decision tree, with six leaves, for prediction of UPDRS-III motor outcome. (right) The performance of the tree on the data is shown, which is sub-optimal, given that only one of six outcomes can be arrived at, at the leaves. Use of random forest of decision trees aims to improve this performance. Radiomic features, such as difference Entropy, SZHGE and LZLGE as seen above, are elaborated in the supplement. (m) and (l) refer to the more and less affected sides, respectively (e.g. caudate(m) is the more affected caudate).

Fig. 3

Plots of outcome prediction, when using (top) only demographics and clinical measures, (middle) addition of conventional features as extracted from DAT SPECT images, and (top) addition of radiomics features as extracted from the images. Data from years 0 and 1 are both utilized to predict motor performance in year 4.

Fig. 4

Plots of outcome prediction, when using (top) only demographics and clinical measures, (middle) addition of conventional features as extracted from DAT SPECT images, and (top) addition of radiomics features as extracted from the images. Data from only year 0 (baseline) are utilized to predict motor performance in year 4.

Fig. 5

Relative contribution of different predictors. %IncMSE is incremental % change in mean square error (MSE) by exclusion of a single feature. Top two predictors are the clinical motor measures (UPDRS-III) in years 1 and 0. But as seen in fig. 3, additional use of the radiomic features is necessarily to significantly improve prediction of outcome. (m) and (l) refer to the more and less affected sides of a structure: putamen, caudate or VS (ventral striatum). The radiomic features are elaborated in the supplement.