ABSTRACT
BACKGROUND AND AIMS: We aimed to identify serum metabolites associated with mucosal and transmural inflammation in pediatric Crohn disease (pCD). METHODS: Fifty-six pCD patients were included through a pre-planned sub-study of the multicenter, prospective, ImageKids cohort, designed to develop the Pediatric Inflammatory Crohn's MRE Index (PICMI). Children were included throughout their disease course when undergoing ileocolonoscopy and magnetic resonance enterography (MRE) and followed for 18 months when MRE was repeated. Serum metabolites were identified using liquid chromatography/mass spectroscopy. Outcomes included: PICMI, the simple endoscopic score (SES), faecal calprotectin (FCP), and C-reactive protein (CRP), to assess transmural, mucosal, and systemic inflammation, respectively. Random forest models were built by outcome. Maximum relevance minimum redundancy (mRMR) feature selection with a j-fold cross validation scheme identified the best subset of features and hyperparameter settings. RESULTS: Tryptophan and glutarylcarnitine were the top common mRMR metabolites linked to pCD inflammation. Random forest models established that amino acids and amines were among the most influential metabolites for predicting transmural and mucosal inflammation. Predictive models performed well, each with an area under the curve (AUC) > 70%. In addition, serum metabolites linked with pCD inflammation mainly related to perturbations in citrate cycle (TCA cycle), aminoacyl-tRNA biosynthesis, tryptophan metabolism, butanoate metabolism, and tyrosine metabolism. CONCLUSIONS: We extend on recent studies, observing differences in serum metabolite between healthy controls and Crohn disease patients, and suggest various associations of serum metabolites with transmural and mucosal inflammation. These metabolites could improve the understanding of pCD pathogenesis and assess disease severity.
ABSTRACT
BACKGROUND: Deep learning using convolutional neural networks (CNNs) has shown great promise in advancing neuroscience research. However, the ability to interpret the CNNs lags far behind, confounding their clinical translation. NEW METHOD: We interrogated 3 heatmap-generating techniques that have increasing generalizability for CNN interpretation: class activation mapping (CAM), gradient (Grad)-CAM, and Grad-CAM++. To investigate the impact of CNNs on heatmap generation, we also examined 6 different models trained to classify brain magnetic resonance imaging into 3 types: relapsing-remitting multiple sclerosis (RRMS), secondary progressive MS (SPMS), and control. Further, we designed novel methods to visualize and quantify the heatmaps to improve interpretability. RESULTS: Grad-CAM showed the best heatmap localizing ability, and CNNs with a global average pooling layer and pretrained weights had the best classification performance. Based on the best-performing CNN model, called VGG19, the 95th percentile values of Grad-CAM in SPMS were significantly higher than RRMS, indicating greater heterogeneity. Further, voxel-wise analysis of the thresholded Grad-CAM confirmed the difference identified visually between RRMS and SPMS in discriminative brain regions: occipital versus frontal and occipital, or temporal/parietal. COMPARISON WITH EXISTING METHODS: No study has examined the CAM methods together using clinical images. There is also lack of study on the impact of CNN architecture on heatmap outcomes, and of technologies to quantify heatmap patterns in clinical settings. CONCLUSIONS: Grad-CAM outperforms CAM and Grad-CAM++. Integrating Grad-CAM, novel heatmap quantification approaches, and robust CNN models may be an effective strategy in identifying the most crucial brain areas underlying disease development in MS.
