Using Artificial Neural Network to Discriminate Parkinson’s Disease from Other Parkinsonisms by Focusing on Putamen of Dopamine Transporter SPECT Images

Abstract

Background: The challenge of differentiating, at an early stage, Parkinson’s disease from parkinsonism caused by other disorders remains unsolved. We proposed using an artificial neural network (ANN) to process images of dopamine transporter single-photon emission computed tomography (DAT-SPECT). Methods: Abnormal DAT-SPECT images of subjects with Parkinson’s disease and parkinsonism caused by other disorders were divided into training and test sets. Striatal regions of the images were segmented by using an active contour model and were used as the data to perform transfer learning on a pre-trained ANN to discriminate Parkinson’s disease from parkinsonism caused by other disorders. A support vector machine trained using parameters of semi-quantitative measurements including specific binding ratio and asymmetry index was used for comparison. Results: The predictive accuracy of the ANN classifier (86%) was higher than that of the support vector machine classifier (68%). The sensitivity and specificity of the ANN classifier in predicting Parkinson’s disease were 81.8% and 88.6%, respectively. Conclusions: The ANN classifier outperformed classical biomarkers in differentiating Parkinson’s disease from parkinsonism caused by other disorders. This classifier can be readily included into standalone computer software for clinical application.

Explainable AI to improve acceptance of convolutional neural networks for automatic classification of dopamine transporter SPECT in the diagnosis of clinically uncertain parkinsonian syndromes

“…The custom CNN used in the present study achieved high overall accuracy of 95.8%, in line with previous studies demonstrating excellent performance of artificial networks for automatic classification of DAT-SPECT [18,[47][48][49][50][51][52][53][54][55][56][57][58]. Specificity was somewhat higher than sensitivity.…”

“…Conventional machine learning methods using support vector machines [33][34][35][36][37][38][39][40][41][42][43], decision trees [44,45], or cluster analyses [46] based on a (small) set of predefined image-derived features have been proposed for this purpose. However, recent work suggests that artificial neural networks, particularly deep CNN, outperform conventional approaches for the automatic classification of DAT-SPECT [18,[47][48][49][50][51][52][53][54][55][56][57][58], partly because artificial neural networks can be less sensitive to camera-and site-specific variability of image quality (e.g., with respect to spatial resolution) [18]. Thus, deep CNN are very promising to support interpretation of DAT-SPECT in clinical routine so that there is a high clinical need for methods to explain CNN-based classification in individual patients.…”

Artificial Intelligence Algorithm-Based Ultrasound Image Segmentation Technology in the Diagnosis of Breast Cancer Axillary Lymph Node Metastasis

“…e presence or absence of breast cancer with axillary lymph node metastasis not only affects the choice of treatment options, but also has important implications for prognostic evaluation. Chien et al [18] found that about 60% of breast cancer patients had regional lymph node metastasis (axillary, internal breast, and supraclavicular), and 40% of them had axillary lymph node metastasis, and most of them were sentinel lymph node metastasis. What is more, the 10-year survival rate of patients without axillary lymph node metastasis was higher than the rate of patients with axillary lymph node metastasis.…”

Will Artificial Intelligence Outperform the Clinical Neurologist in the Near Future? Yes

“…Indeed, recent studies have provided promising results demonstrating that ML improves classification performance of dopamine transporter single photon emission computed tomography images for the differentiation of PD from other parkinsonian disorders, potentially shortening the diagnostic delay in our case of YOPD. 11,12 This application could be implemented in underserved regions assisting a remote radiologist to interpret advanced neuroimages, leading to an earlier and more accurate diagnosis of PD, especially in the early stage when characteristic features to confirm or refute the diagnosis are quite subtle and neither sensitive enough to be observed by clinical examination (eg, subtle red flags) nor specific enough for clinical interpretation (eg, a less defined dopaminergic response). 13…”