3663 - Predicting the Risk of Local Failure after Treatment for Head and Neck Squamous Cell Carcinoma Using Clinical and Post-Radiotherapy PET/CT Imaging Features

R. T. Hughes¹, E. A. Restler², S. Smith³, R. F. Shenker⁴, C. M. Lanier⁵, and J. A. Schroeder⁶; ¹Department of Radiation Oncology, Wake Forest University School of Medicine, Winston Salem, NC, ²Wake Forest University School of Medicine, Winston Salem, NC, ³Department of Biostatistics and Data Science, Wake Forest University School of Medicine, Winston Salem, NC, ⁴Department of Radiation Oncology, Duke University Medical Center, Durham, NC, ⁵Department of Radiation Oncology, Wake Forest University School of Medicine, Winston-Salem, NC, ⁶Department of Radiology, University of North Carolina School of Medicine, Chapel Hill, NC

Purpose/Objective(s): Post-radiotherapy (RT) fludeoxyglucose-18 (FDG) positron emission tomography/computed tomography (PET/CT) is routinely used for disease response 3-months after treatment of head and neck squamous cell carcinoma (HNSCC). While evidence supports the diagnostic accuracy of PET/CT for residual/recurrent disease, models predicting risk of recurrence based on clinical and imaging features are needed. Materials/

Methods: Patients with HNSCC of the oral cavity, oropharynx, larynx, hypopharynx, nasopharynx or paranasal sinuses treated with definitive or adjuvant RT between 2014-2020 were identified. Exclusion criteria included: recurrent/metastatic disease, no post-RT PET/CT within 1 year of RT and reirradiation. The following features were collected from review of PET/CT images: maximum standardized uptake value (SUVmax) of the primary site on initial (if available) and post-RT PET/CT, post-RT focal or asymmetric FDG uptake at the primary site, post-RT uptake contralateral to the primary site (for lateralized tumors), and post-RT liver and blood pool uptake. Ratios of FDG uptake at various sites were calculated, including post-RT primary site ipsilateral to contralateral, post-RT primary site to liver, post-RT to pre-RT (n=180 with pre-RT PET/CT available for review). The primary endpoint was LF at the primary site within 2 years after PET/CT. Clinical factors (T and N stage, smoking status, definitive v. adjuvant RT, HPV status, and concurrent chemotherapy) and PET/CT features (above) were included into a backward selection multivariable logistic regression model.
Results: 280 patients were included. Median age was 60, 76% were males and 75% were current/former smokers, 42% were diagnosed with HPV-associated oropharyngeal cancer. AJCC 7^th Ed. Stage III-IVB disease was present in 89%; 50% T1-2 and 89% N2-3. RT was definitive in 67% and adjuvant in 33%; 88% received concurrent chemotherapy. The median time from RT to PET/CT was 91 days (interquartile range [IQR] 85-101). Median follow-up was 31.3 months (IQR 17.5-51.5). On univariable analyses, post-RT factors associated with LF were focal FDG uptake, asymmetric FDG uptake, SUVmax at the primary site, ipsilateral primary site:contralateral ratio, and primary site:liver ratio. Multivariable analysis identified T3-4 tumor (odds ratio [OR] 6.48, p=0.01), current/former smoking versus never smoking (OR 8.42, p=0.04), asymmetric uptake on post-RT PET/CT (OR 5.90, p=0.02), and the ratio of post-RT:pre-RT uptake at the primary site (OR 7.14, p=0.07) as the most predictive model of local failure incorporating both clinical and PET/CT factors. An additional model including T3-4, current/former smoking, and post-RT primary site uptake performed similarly.
Conclusion: We identify practical models predicting LF after RT for patients with HNSCC, which may inform clinical practice and indicate additional surveillance measures for patients identified at high risk. Further validation is warranted.