Massachusetts General Hospital / Brigham and Women's Hospital (Center for Clinical Data Science)
Training code and model weights not public. Contact creators for more information.
Deep learning pipeline for detection and quantification of abdominal aortic aneurysm (AAA) -- a typically asymptomatic condition often missed incidentally by radiologists -- from abdominal-pelvic CT. A modified 3D U-Net segments the aorta on both contrast and non-contrast CT volumes with a variable number of images, after which an ellipse-fitting post-processing step measures the aortic cross-sectional diameter along the vessel to detect aneurysmal dilation. Trained and validated on 321 abdominal-pelvic CT examinations from Massachusetts General Hospital, the model reached a sensitivity/specificity of 0.91/0.95 on the primary validation set, and 0.85/1.0 on a separate 57-exam generalization test set with different patient demographics and acquisition characteristics; the authors report that DeepAAA exceeded literature-reported radiologist performance for incidental AAA detection.
Architecture
CNN (3D)
Modified 3D U-Net for aorta segmentation on CT volumes, followed by classical ellipse-fitting post-processing along the segmented vessel to measure aortic diameter and detect aneurysmal dilation
Added to catalog
2026-08-12
DeepAAA External Generalization Test Cohort
A separate set of 57 abdominal-pelvic CT examinations with differing patient demographics and acquisition characteristics than the primary MGH training cohort, used to test the generalizability of the DeepAAA model.
MGH/BWH Abdominal-Pelvic CT Cohort (DeepAAA)
Abdominal-pelvic CT examinations (contrast and non-contrast) performed at Massachusetts General Hospital, used to train and validate a 3D aorta segmentation and abdominal aortic aneurysm (AAA) detection model.
3D segmentation of the aorta on abdominal-pelvic CT, combined with ellipse fitting along the vessel to quantify aortic cross-sectional diameter
Patient-level binary detection of abdominal aortic aneurysm (AAA) based on the measured aortic diameter derived from the segmentation