Background and purpose: The role of FDG-PET in radiotherapy target volume definition of the neck was evaluated by comparing eight methods of FDG-PET segmentation to the current CT-based practice of lymph node assessment in head-and-neck cancer patients.
Materials and methods: Seventy-eight head-and-neck cancer patients underwent coregistered CT- and FDG-PET scans. Lymph nodes were classified as "enlarged" if the shortest axial diameter on CT was 10mm, and as "marginally enlarged" if it was 7-10mm. Subsequently, lymph nodes were assessed on FDG-PET applying eight segmentation methods: visual interpretation (PET(VIS)), applying fixed thresholds at a standardized uptake value (SUV) of 2.5 and at 40% and 50% of the maximum signal intensity of the primary tumor (PET(SUV), PET(40%), PET(50%)) and applying a variable threshold based on the signal-to-background ratio (PET(SBR)). Finally, PET(40%N), PET(50%N) and PET(SBRN) were acquired using the signal of the lymph node as the threshold reference.
Results: Of 108 nodes classified as "enlarged" on CT, 75% were also identified by PET(VIS), 59% by PET(40%), 43% by PET(50%) and 43% by PET(SBR). Of 100 nodes classified as "marginally enlarged", only a minority were visualized by FDG-PET. The respective numbers were 26%, 10%, 7% and 8% for PET(VIS), PET(40%), PET(50%) and PET(SBR). PET(40%N), PET(50%N) and PET(SBRN), respectively, identified 66%, 82% and 96% of the PET(VIS)-positive nodes.
Conclusions: Many lymph nodes that are enlarged and considered metastatic by standard CT-based criteria appear to be negative on FDG-PET scan. Alternately, a small proportion of marginally enlarged nodes are positive on FDG-PET scan. However, the results are largely dependent on the PET segmentation tool used, and until proper validation FDG-PET is not recommended for target volume definition of metastatic lymph nodes in routine practice.