Multifocal regression and pathologic response predicts recurrence after neoadjuvant chemotherapy in head and neck squamous cell carcinoma

Objectives: Complete pathological response after neoadjuvant chemotherapy (NAC) in head and neck squamous cell carcinomas (HNSCC) is a good prognostic factor. Multifocal regression post-NAC in breast cancer has proven to impact locoregional control (LRC) but has not been evaluated in HNSCC. We evaluate the impact of multifocal regression and major pathologic response (MPR) on survival indices in HNSCC. Materials and Methods: Retrospective review of HNSCC patients receiving NAC followed by surgery with curative intent between March 2016 to March 2019 at MD Anderson Cancer Center. Tumor focality (uni- or multifocal), pathologic response and other pathologic data were collected. MPR was defined as <= 10% residual tumor. Overall survival (OS) and LRC were analyzed and multivariate Cox regression analysis was performed. Results: 101 patients were analyzed, with 18.8% pathologic complete response, 18.8% with 1-10% viable tumor and 60.4% with > 10% viable tumor. 61 (60.4%) had unifocal disease while 19 (18.8%) had multifocal disease. Tumor focality was significantly associated with LRC but not OS, where the 3-year LRC was 82%, 69% and 52% (p = 0.015) for no viable tumor, unifocal disease and multifocal disease respectively. On multivariate analysis, multifocal disease (HR 10.43; 95 %CI 1.24-87.5) and extranodal extension (HR 4.4; 95 %CI 1.60-12.07) continued to be significant independent predictors of LRC. MPR group displayed significantly better 3-year OS (75% vs 51%, p = 0.041) and 3-year LRC (80% vs 62%, p = 0.011) than those with > 10% viable tumor. Conclusion: Multifocal regression and less than MPR after NAC in HNSCC predicts for locoregional recurrence and should be routinely reported.