Deep Learning Improves Sensitivity to Change in Sinus Computed Tomography: Evidence From Two Randomized Controlled Trials

J.W.M.: Employee and owned stock or stock options in AstraZeneca at time of work, and current employee of Blue Earth Diagnostics. L.H.N.: Employee and owned stock or stock options in AstraZeneca at time of work, and current employee of Vida Diagnostics. S.N., D.C., V.W., and M.J.: Employee and owned stock or stock options in AstraZeneca. R.S.: Employee of Perceptive. S.M.H.: U.S. Patent 12,482,124 assigned to National Jewish Health and unlicensed. Figure S1: – Participant selection flow diagrams for A) OSTRO and B) ORCHID. Figure S2: - Example CT images and sinus segmentations produced by the deep learning algorithm in A) baseline and B) week 56 CT in a representative study participant. 1) axial section, 2) axial section with segmentation overlays (yellow = sphenoid, fuchsia = posterior ethmoid, cyan = anterior ethmoid), 3) coronal section, 4) coronal section with segmentation overlay (blue = frontal, cyan = anterior ethmoid, red = right maxillary, green = left maxillary). Total opacification changed from 98.9% to 89.6%. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.