Quantifying the calcification of abdominal aorta and major side branches with deep learning.

AIM: To explore the possibility of a neural network-based method for quantifying calcifications of the abdominal aorta and its branches. MATERIALS AND METHODS: In total, 58 computed tomography (CT) angiography volumes were selected from a dataset of 609 to represent different stages of sclerosis. The ground truth segmentations of the abdominal aorta, coeliac trunk, superior mesenteric artery, renal arteries, common iliac arteries, and their calcifications were delineated manually. Two V-Net ensemble models were trained, one for segmenting arteries of interest and another for calcifications. The branches of interest were shortened algorithmically. The volumes of calcification were then evaluated from the arteries of interest. RESULTS: The results indicate that automatic detection is possible with a high correlation to the ground truth. The scores for the ensemble calcification model were dice score of 0.69 and volumetric similarity (VS) of 0.80 and for the arteries of interest segmentations: aorta: dice 0.96, VS 0.98; aortic branches: dice 0.74, VS 0.87; and common iliac arteries: dice 0.72, VS 0.91. CONCLUSIONS: The presented neural network model is the first to be capable of automatically segmenting, in addition to calcification, both the aorta and its branches from contrast-enhanced CT angiography. This technology shows promise in addressing limitations inherent in earlier methods that relied solely on plain CT.

Automatic detection of Crohn's disease using quantified motility in magnetic resonance enterography: initial experiences.

AIM: To report initial experiences of automatic detection of Crohn's disease (CD) using quantified motility in magnetic resonance enterography (MRE). MATERIALS AND METHODS: From 302 patients, three datasets with roughly equal proportions of CD and non-CD cases with various illnesses were drawn for testing and neural network training and validation. All datasets had unique MRE parameter configurations and were performed in free breathing. Nine neural networks were devised for automatic generation of three different regions of interests (ROI): small bowel, all bowel, and non-bowel. Additionally, a full-image ROI was tested. The motility in an MRE series was quantified via a registration procedure, which, accompanied with given ROIs, resulted in three motility indices (MI). A subset of the indices was used as an input for a binary logistic regression classifier, which predicted whether the MRE series represented CD. RESULTS: The highest mean area under the curve (AUC) score, 0.78, was reached using the full-image ROI and with the dataset with the highest cine series length. The best AUC scores for the other two datasets were only 0.54 and 0.49. CONCLUSION: The automatic system was able to detect CD in the group of MRE studies with lower temporal resolution and longer cine series showing potential in primary bowel disorder diagnostics. Larger ROI selections and utilising all available cine series for motility registration yielded slight performance improvements.