Generating lung tumor internal target volumes from 4D-PET maximum intensity projections

Area of Science:

• Nuclear Medicine
• Radiotherapy Planning
• Medical Imaging

Background:

• Positron emission tomography (PET) for lung tumors is blurred by breathing motion.
• Respiratory-correlated PET visualizes functional uptake but has limited use in radiotherapy planning.

Purpose of the Study:

• To propose and evaluate a gated PET maximum intensity projection (MIP) technique.
• To quantitatively and efficiently incorporate respiratory-correlated PET into radiotherapy planning.

Main Methods:

• Acquired 4D-CT and respiratory-gated FDG-PET for four small lung tumors.
• Generated internal target volumes (ITVs) from PET-MIP, ungated PET, and 4D-CT.
• Compared volumetric overlap and relative volumes of ITVs using various segmentation thresholds.

Main Results:

• PET-MIP ITVs showed significantly better overlap and volume agreement with CT-MIP ITVs compared to ungated PET ITVs.
• Mean normalized overlap for PET-MIP was 0.68 ± 0.07 versus 0.47 ± 0.12 for ungated PET.
• Mean relative volume for PET-MIP was 1.07 ± 0.42 versus 0.69 ± 0.56 for ungated PET.

Conclusions:

• Gated PET-MIP images provide a better match to CT-MIP images for small, CT-visible lung tumors.
• PET-MIP is an efficient method for radiation oncologists to integrate 4D-PET data into lung tumor contouring.
• This technique requires minimal implementation effort in treatment planning software.