科研速览 · Science Skim继续刷下去 · Keep skimming →
◆ Head & Neck2026-06-16· Margin (machine learning)

Early Feasibility of Registration of Micro‐ <scp>PET</scp> / <scp>CT</scp> Scans to Annotated <scp>3D</scp> Specimen Models

Joaquin Austerlitz, Andreja Radevic, Ashtyn McAdoo, Kyrionna Golliday, Cheng Ye, Daniel Fabbri, Adam J. Rosenberg, Marcus Balanky, Anas Alabkaa, Ely Kr, Mitra Mehrad, Spencer Roark, Eben Rosenthal, Michael C. Topf

原始摘要(英文原文)· Original abstract
BACKGROUND: F-FDG) micro-positron emission tomography/computed tomography (micro-PET/CT) is an emerging modality for margin assessment. Prior to clinical use, micro-PET/CT margin distances must be correlated with gold-standard histopathology. METHODS: Micro-PET/CT uptake was co-registered to 3D specimen models annotated with margin sampling sites using principal component analysis and iterative closest point surface registration. Margin distances were measured on the fused models using Gaussian Mixed Modeling and compared with histopathology. RESULTS: Fused micro-PET/CT and 3D specimen models were successfully generated for five patients with head and neck cancer. Among the 16 measured margins, the margin distance on the fused models was within 1 mm of histopathology in 15/16 cases. The Bland-Altman plot demonstrated excellent agreement (±1.25 mm). CONCLUSION: We report a novel technique to register micro-PET/CT data to annotated 3D specimen models. Margin distance measured on the micro-PET/CT scans correlated closely with histopathology. Intraoperative micro-PET/CT warrants further investigation in head and neck margin assessment.
读原文 · Read the paper ↗

AI 追问PRO

登录后使用 AI 追问

讨论区

登录后参与讨论

相关论文 · Related

Early Feasibility of Registration of Micro‐ <scp>PET</scp> / <scp>CT</scp> Scans to Annotated <scp>3D</scp> Specimen Models — 科研速览 Science Skim