The role of PET for predicting nodal response in locally advanced (LA) head and neck squamous cell carcinoma (HNSCC) treated with chemoradiotherapy (CRT): Results of a prospective multicenter trial.

Abstract

6013 Background: Controversy about neck management after CRT in patients with LA HNSCC persists due to low accuracy of CT/MR to asses the neck. As already demonstrated (Mehanna, NEJM 374, 2016), PET is an alternative to planned neck dissection (ND) thanks to its high negative predictive value (NPV). However, no conclusion could be drawn for patients (pts) with equivocal response (e.g. suspicion of residual disease on CT/MR but negative PET) because pathologic confirmation was lacking. Methods: Multicenter, prospective, nonrandomized trial including pts with LA HNSCC of oral cavity, oro- hypopharynx, larynx, staged N1, N2, N3, treated with CRT and evaluated 12 weeks after CRT by overall assessment (OA): clinical examination (CE), PET and CT/MR. ND was performed in incomplete regional response based on at least 1 positive evaluation method. Pathologic analyses (HE and KI67) were performed on ND samples. Primary objective was to determine the NPV and accuracy of PET as a single examination in the post CRT nodal assessment. Primary outcome was 2-year regional recurrence free survival rate (RRFSR). Results: 264/318 pts included completed full treatment and had post CRT OA. Median follow up was 40 months. No ND was proposed in 119 patients because of a negative OA; 145 patients had ND. The presence of viable cells was reported in 27 ND (18.6%). Sensitivity, specificity, PPV, NPV, accuracy of OA were 90.0%, 49.6%, 18.6%, 97.5%, 54.2% vs 69.7%, 75.3.%, 28.8%, 94.6%, 74.6% for PET alone. Kappa coefficient was of 0.838, indicating an almost perfect agreement. In pts with negative OA, RRFSR was 61.3% vs. 56.6% in pts with positive OA and ND (p=0.45). Using post CRT assessment with PET alone, RRFSR in pts with negative PET was 63.0% vs. 48.8% in pts with positive PET (p=0.04). Using PET assessment alone, 65/145 ND (44.8%) could have been avoided without compromising RRFSR. Conclusions: NPV using PET alone is 94.6%. Post CRT evaluation using only PET would have resulted in considerably fewer ND without jeopardizing neck control. PET alone is more accurate and more discriminant for predicting pts outcome. Clinical trial information: NCT00634777.