Validation of a practical manual method of accurate positron emission tomography (PET) tumour size depiction for...

Positron emission tomography/computed tomography (PET/CT) guidance for tumour ablation procedures offers unique advantages, including intraprocedural margin assessment and visualisation of tumours not visible with other modalities. However, accurate tumour size depiction on PET is necessary. This paper validates a manual method for accurately depicting liver and lung tumour sizes on F18-fluorodeoxyglucose (FDG) PET compared to anatomic imaging with magnetic resonance imaging (MRI) and CT.

This retrospective study included 63 patients with 105 tumours (75 liver and 30 lung) imaged between 2017 and 2021. Inclusion required FDG-avid tumours in the liver or lung, visibility on both PET/CT and MRI (liver tumours) or CT (lung tumours), tumour diameter 0.8-6.0 cm, and anatomical imaging within 7 days of PET/CT. Tumours were windowed on PET images using a manual segmentation method requiring no special software. Interreader agreement and pairwise difference between PET and anatomical tumour measurements were studied using linear mixed-effects models.

Intraclass correlation coefficients for interreader agreement were 0.92 for PET and 0.93 for anatomical tumour measurements. The overall average tumour size difference between PET and anatomical imaging was -1.6 mm (P<.001). The average size difference was -0.6 mm for lung tumours (P=.41) and -2.0 mm for liver tumours (P<.001). The average size difference for tumours ≤2 cm was -0.2 mm and -2.6 mm for tumours >2 cm.

A manual windowing method for PET tumour images provided comparable liver and lung tumour size depictions compared to MRI and CT with high interreader correlation.