ABSTRACT
The current era witnesses a highly dynamic development of Artificial Intelligence (AI) applications, impacting various human activities. Medical imaging techniques are no exception. AI can find application in image acquisition, image processing and augmentation, as well as in the actual interpretation of images. Moreover, within the domain of radiomics, AI can be instrumental in advanced analysis surpassing the capacities of the human eye and experience. While several certified commercial solutions are available, the validation and accumulation of sufficient evidence regarding their positive impact on healthcare is currently constrained. The role of AI presently leans towards being assistive, yet further evolution is anticipated. Risks and disadvantages encompass dependency on computational power, the quality of input data, and their annotation for learning purposes. The transparency of algorithmic functioning is lacking, and issues pertaining to portability may arise. The integration and utilization of AI introduce entirely new ethical and legislative aspects. Predicting the future development of AI in imaging methods is challenging, with a further increase in implementation appearing more probable.
Subject(s)
Artificial Intelligence , Diagnostic Imaging , Humans , Diagnostic Imaging/methods , Image Processing, Computer-Assisted/methodsABSTRACT
Isolated REM sleep behavior disorder (iRBD) is an early stage of synucleinopathy with most patients progressing to Parkinson's disease (PD) or related conditions. Quantitative susceptibility mapping (QSM) in PD has identified pathological iron accumulation in the substantia nigra (SN) and variably also in basal ganglia and cortex. Analyzing whole-brain QSM across iRBD, PD, and healthy controls (HC) may help to ascertain the extent of neurodegeneration in prodromal synucleinopathy. 70 de novo PD patients, 70 iRBD patients, and 60 HCs underwent 3 T MRI. T1 and susceptibility-weighted images were acquired and processed to space standardized QSM. Voxel-based analyses of grey matter magnetic susceptibility differences comparing all groups were performed on the whole brain and upper brainstem levels with the statistical threshold set at family-wise error-corrected p-values <.05. Whole-brain analysis showed increased susceptibility in the bilateral fronto-parietal cortex of iRBD patients compared to both PD and HC. This was not associated with cortical thinning according to the cortical thickness analysis. Compared to iRBD, PD patients had increased susceptibility in the left amygdala and hippocampal region. Upper brainstem analysis revealed increased susceptibility within the bilateral SN for both PD and iRBD compared to HC; changes were located predominantly in nigrosome 1 in the former and nigrosome 2 in the latter group. In the iRBD group, abnormal dopamine transporter SPECT was associated with increased susceptibility in nigrosome 1. iRBD patients display greater fronto-parietal cortex involvement than incidental early-stage PD cohort indicating more widespread subclinical neuropathology. Dopaminergic degeneration in the substantia nigra is paralleled by susceptibility increase, mainly in nigrosome 1.
Subject(s)
Parkinson Disease , REM Sleep Behavior Disorder , Synucleinopathies , Humans , REM Sleep Behavior Disorder/diagnostic imaging , Synucleinopathies/complications , Synucleinopathies/pathology , Brain/diagnostic imaging , Brain/pathology , Substantia Nigra/diagnostic imaging , Substantia Nigra/pathology , Parkinson Disease/complications , IronABSTRACT
OBJECTIVE: To apply a machine learning analysis to clinical and presynaptic dopaminergic imaging data of patients with rapid eye movement (REM) sleep behavior disorder (RBD) to predict the development of Parkinson disease (PD) and dementia with Lewy bodies (DLB). METHODS: In this multicenter study of the International RBD study group, 173 patients (mean age 70.5 ± 6.3 years, 70.5% males) with polysomnography-confirmed RBD who eventually phenoconverted to overt alpha-synucleinopathy (RBD due to synucleinopathy) were enrolled, and underwent baseline presynaptic dopaminergic imaging and clinical assessment, including motor, cognitive, olfaction, and constipation evaluation. For comparison, 232 RBD non-phenoconvertor patients (67.6 ± 7.1 years, 78.4% males) and 160 controls (68.2 ± 7.2 years, 53.1% males) were enrolled. Imaging and clinical features were analyzed by machine learning to determine predictors of phenoconversion. RESULTS: Machine learning analysis showed that clinical data alone poorly predicted phenoconversion. Presynaptic dopaminergic imaging significantly improved the prediction, especially in combination with clinical data, with 77% sensitivity and 85% specificity in differentiating RBD due to synucleinopathy from non phenoconverted RBD patients, and 85% sensitivity and 86% specificity in discriminating PD-converters from DLB-converters. Quantification of presynaptic dopaminergic imaging showed that an empirical z-score cutoff of -1.0 at the most affected hemisphere putamen characterized RBD due to synucleinopathy patients, while a cutoff of -1.0 at the most affected hemisphere putamen/caudate ratio characterized PD-converters. INTERPRETATION: Clinical data alone poorly predicted phenoconversion in RBD due to synucleinopathy patients. Conversely, presynaptic dopaminergic imaging allows a good prediction of forthcoming phenoconversion diagnosis. This finding may be used in designing future disease-modifying trials. ANN NEUROL 2024;95:1178-1192.