RESUMO
Polymicrogyria (PMG) is a disorder of cortical organization mainly seen in children, which can be associated with seizures, developmental delay and motor weakness. PMG is typically diagnosed on magnetic resonance imaging (MRI) but some cases can be challenging to detect even for experienced radiologists. In this study, we create an open pediatric MRI dataset (PPMR) containing both PMG and control cases from the Children's Hospital of Eastern Ontario (CHEO), Ottawa, Canada. The differences between PMG and control MRIs are subtle and the true distribution of the features of the disease is unknown. This makes automatic detection of potential PMG cases in MRI difficult. To enable the automatic detection of potential PMG cases, we propose an anomaly detection method based on a novel center-based deep contrastive metric learning loss function (cDCM). Despite working with a small and imbalanced dataset our method achieves 88.07% recall at 71.86% precision. This will facilitate a computer-aided tool for radiologists to select potential PMG MRIs. To the best of our knowledge, our research is the first to apply machine learning techniques to identify PMG solely from MRI. Our code is available at: https://github.com/RichardChangCA/Deep-Contrastive-Metric-Learning-Method-to-Detect-Polymicrogyria-in-Pediatric-Brain-MRI. Our pediatric MRI dataset is available at: https://www.kaggle.com/datasets/lingfengzhang/pediatric-polymicrogyria-mri-dataset.
Assuntos
Polimicrogiria , Criança , Humanos , Polimicrogiria/complicações , Polimicrogiria/patologia , Encéfalo , Imageamento por Ressonância Magnética , CanadáRESUMO
Emerging evidence suggests that advanced neuroimaging modalities such as arterial spin labelling (ASL) might have prognostic utility for pediatric concussion. This study aimed to: 1) examine group differences in global and regional brain perfusion in youth with concussion or orthopedic injury (OI) at 72 h and 4 weeks post-injury; 2) examine patterns of abnormal brain perfusion within both groups and their recovery; 3) investigate the association between perfusion and symptom burden within concussed and OI youths at both time-points; and 4) explore perfusion between symptomatic and asymptomatic concussed and OI youths. Youths ages 10.00-17.99 years presenting to the emergency department with an acute concussion or OI were enrolled. ASL-magnetic resonance imaging scans were conducted at 72 h and 4 weeks post-injury to measure brain perfusion, along with completion of the Health Behavior Inventory (HBI) to measure symptoms. Abnormal perfusion clusters were identified using voxel-based z-score analysis at each visit. First, mixed analyses of covariance (ANCOVAs) investigated the Group*Time interaction on global and regional perfusion. Post hoc region of interest (ROI) analyses were performed on significant regions. Second, within-group generalized estimating equations investigated the recovery of abnormal perfusion at an individual level. Third, multiple regressions at each time-point examined the association between HBI and regional perfusion, and between HBI and abnormal perfusion volumes within the concussion group. Fourth, whole-brain one-way ANCOVAs explored differences in regional and abnormal perfusion based on symptomatic status (symptomatic vs. asymptomatic) and OIs at each time-point. A total of 70 youths with a concussion [median age (interquartile range; IQR) = 12.70 (11.67-14.35), 47.1% female] and 29 with an OI [median age (IQR) = 12.05 (11.18-13.89), 41.4% female] were included. Although no Group effect was found in global perfusion, the concussion group showed greater adjusted perfusion within the anterior cingulate cortex/middle frontal gyrus (MFG) and right MFG compared with the OI group across time-points (ps ≤ 0.004). The concussion group showed lower perfusion within the right superior temporal gyrus at both time-points and bilateral occipital gyrus at 4 weeks, (ps ≤ 0.006). The number of hypoperfused clusters was increased at 72 h compared with 4 weeks in the concussion youths (p < 0.001), but not in the OIs. Moreover, Group moderated the HBI-perfusion association within the left precuneus and superior frontal gyrus at both time-points, (ps ≤ 0.001). No association was found between HBI and abnormal perfusion volume within the concussion group at any visits. At 4 weeks, the symptomatic sub-group (n = 10) showed lower adjusted perfusion within the right cerebellum and lingual gyrus, while the asymptomatic sub-group (n = 59) showed lower adjusted perfusion within the left calcarine, but greater perfusion within the left medial orbitofrontal cortex, right middle frontal gyrus, and bilateral caudate compared with OIs. Yet, no group differences were observed in the number of abnormal perfusion clusters or volumes at any visit. The present study suggests that symptoms may be associated with changes in regional perfusion, but not abnormal perfusion levels.