Limitations of PET/CT in determining need for neck dissection after primary chemoradiation for advanced head and neck squamous cell carcinoma

Objective: The purpose of our study was to evaluate PET/CT in predicting residual nodal disease after primary chemoradiation (CRT) for head and neck cancer (HNSCC) with N2 disease or greater. Design: A retrospective cohort analysis was conducted. Thirty-eight patients received primary CRT for HNSCC with N2 or greater disease, PET/CT after treatment and neck dissection from January 2003 to December 2006. PET/CT results were correlated with pathology results from neck dissection specimens for each respective side. Results: Forty-six neck regions were analyzed. Nine were determined to have either PET/CT result reports (n = 5) or pathology reports (n = 4) that were indeterminate. PET/CT demonstrated a sensitivity and specificity of 57.1 and 73.9%, respectively, an accuracy of 67.5%, a positive predictive value of 57.1% and a negative predictive value of 73.9% when compared to the final pathology. Conclusion: PET/CT is neither highly sensitive nor highly specific for identifying residual nodal metastases after CRT for advanced-stage HNSCC. Physicians should not rely solely on PET/CT to determine interventions on the neck after CRT. A standardization of language for reporting findings and risk of residual disease is needed.