Reliability of MRICloud automatic segmentation for measuring brain ventricular volumes in Parkinson’s disease patients

Abstract Background Parkinson’s disease is the second most common neurodegenerative disease. In Parkinson’s disease, enlargement of the ventricles is expected, and when the normal anatomical shape of the ventricles is distorted, automatic segmentation methods may not be able to make accurate measurements. Therefore, it is important to evaluate the reliability of automatic segmentation methods. Our aim in this study is to obtain volumetric measurements of brain ventricles in Parkinson’s disease patients with MRICloud and compare these with the manual method in relation to healthy controls. Methods 40 idiopathic Parkinson’s disease patients and 44 healthy controls were included in our study retrospectively. Brain ventricle volumes of both groups were measured using MRICloud and manual segmentation method. The volumes of both groups were compared. The reliability of MRICloud was evaluated by comparing it with manual segmentation, which is considered the gold standard. Results We found statistically significant volumetric differences in several regions, such as the atrial part of the right lateral ventricle ( p = .030), the atrial part of the left lateral ventricle ( p = .030), and the occipital horn of the right lateral ventricle ( p = .030) between the PD and control groups. In the PD group, the intraclass correlation coefficients (ICC) value is 0.936, 0.837, 0.642, and 0.853 for the right lateral ventricle, left lateral ventricle, 3rd ventricle, and 4th ventricle, respectively. In the control group, the ICC value is. 999, 0.998, 0.912, and 0.815 for the right lateral ventricle, left lateral ventricle, 3rd ventricle, and 4th ventricle, respectively. Conclusions Our findings suggest that MRICloud is a good method for detecting the volumes of normal and pathological brain ventricles, except for the 3rd ventricle. In addition to clinical criteria, we think MRICloud is a reliable and useful method to help diagnose Parkinson’s disease and track the progression of the disease.