Staging of head and neck cancer with PET/MR and PET/CT - initial results

Objectives To compare the diagnostic accuracy of whole-body PET/MR with that of whole-body PET/CT in determining the stage of head and neck cancer.

Methods In this prospective study, sequential FDG-PET/CT-MR was performed in 27 patients (median age 66 years, range 45 to 95 years, 16 males) with newly diagnosed head and neck cancer. Contrast-enhanced CT was performed in all subjects but those with allergy. MR (Whole body: LAVA-Flex; Head and neck: ax T2w, ax and cor T1w with and without contrast, DWI) was acquired, roughly matching the acquisition time of PET (15 minutes). PET/CT and PET/MR were evaluated separately, and a tumor-node-metastasis (TNM) stage was assigned based upon the image analysis. Nodal stations were identified according to the mapping system published by Som et al. (1). The standard of reference was histopathology for tumor and nodal stage in 14 patients, and follow-up with imaging in 13 patients. The staging classification was rated as correct (score of 2), equivocal (score of 1), or incorrect (score of 0). Wilcoxon signed ranks test was used to compare PET/CT and PET/MR.

Results The T/N/M staging by PET/MR was correct in 22 patients (81.5%) / 21 (77.8%) / 27 (100%), equivocal in 4 patients (14.8%) / 2 (7.4%) / 0 (0%), and incorrect in 1 patient (3.7%) / 4 (14.6%) / 0 (0%). The T/N/M staging by PET/CT was correct in 21 patients (77.8%) / 23 (85.2%) / 24 (88.9%), equivocal in 2 patient (7.4%) / 0 (0%) / 0 (0%), and incorrect in 4 patients (14.8%) / 4 (14.8%) / 3 (11.1%). Consistently, TNM staging by PET/MR and PET/CT was of equal accuracy (T: p = 0.271, N: p = 0.746, M: p = 0.083).

Conclusions Whole-body staging with PET/MR with 15 minutes acquisition time yields equal diagnostic accuracy in determining the stage of head and neck cancer.

Research Support GE Healthcare