Beyond CT: Attenuation correction for stand-alone brain PET.

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Radiophysics

Background:

• Quantitative brain PET imaging requires accurate attenuation correction (AC) for reliable results.
• Dedicated PET-only systems lack CT or MRI, posing a significant challenge for AC.
• Developing CT-less AC methods is vital for stand-alone brain PET systems.

Purpose of the Study:

• To provide a comprehensive review of CT-less AC methods for dedicated brain PET systems.
• To categorize and analyze existing AC approaches and their limitations.
• To identify the gap between research advancements and clinical implementation.

Main Methods:

• Review and analysis of segmentation-based, atlas/template-driven, and simultaneous reconstruction AC techniques.
• Evaluation of data-driven deep learning approaches for attenuation mapping.
• Survey of AC methods currently used in commercial brain PET systems.

Main Results:

• Template-based methods are efficient but struggle with atypical anatomy.
• Simultaneous reconstruction methods are powerful but sensitive to noise and require TOF.
• Deep learning shows high potential for patient-specific AC but needs large datasets and faces generalizability issues.
• Clinical practice still favors simpler, established AC techniques over advanced research methods.

Conclusions:

• There is an ongoing need for CT-less AC solutions that balance accuracy, robustness, and clinical feasibility.
• Bridging the gap between research and clinical practice is essential for advancing stand-alone brain PET.
• Further development is required to overcome limitations of current CT-less AC methods for widespread adoption.