Neural networks of high-res quantitative CT (HR-pQCT) scans allowed for 3D imaging of MCP2 to differentiate normals from #RA & #PsA. 932 of 617 ppts correct AuROC for HC 82%, RA 75%, PsA 68%. UA pts read as 86% RA,11% PsA, & 3% HC https://t.co/wSQgllYgug https://t.co/yA6g55SpeC