Purpose: To examine the time-dependent diagnostic performance of FDG-PET/CT in the follow-up of head and neck cancer (HNC) and to assess the prognostic value of PET-negative and PET-inconclusive findings.
Material and methods: 279 HNC patients primarily treated with radiotherapy from 2006 to 2012 were included. The follow-up PET/CT scans were categorized as benign, malignant or inconclusive by a radiologist and a nuclear physician. The reference standard was histology or verification by progression on imaging. The outcome measures were positive (PPV) and negative predictive value (NPV), and the PET/CT scans were grouped according to time since treatment and compared. An analysis of the diagnostic accuracy was performed with the inconclusive results categorized as both benign and malignant to create ranges for the diagnostic performance.
Results: The proportion of inconclusive results declined from 26 to 8.4% and 0% after 0-3, 3-6 and 12-24 months post-treatment. The ranges for diagnostic performance after 0-3, 3-6, 6-12, 12-24 months and overall post-treatment were: PPV 27.3-50, 48.4-58.3, 71.4-100, 100 and 50.5-65.7 and NPV 75.0-84.6, 95.1-96.8, 92.9-100, 100 and 94.8-96.7. Time to recurrence was not statistically different after a PET-negative or a PET-inconclusive result.
Conclusion: The diagnostic accuracy of a surveillance PET/CT scan after HNC improves with time since treatment, and is very reliable after 1 year. However, the NPV is already high 3 months post-treatment supporting the use of PET/CT for early evaluation of head and neck cancer patients.
Keywords: FDG-PET/CT; Follow-up; Head and neck cancer; Nuclear medicine; Oncology; Radiotherapy; Surveillance.