Current head and neck epidemiology demonstrates a steadily increasing incidence of p16+ human papillomavirus-related oropharynx squamous cell cancer (OPSCC). This distinct tumor subtype is associated with better survival outcomes. There is a growing recognition of the need to define management regimens that take into account the inherent patho-biological attributes of these cancers and provide optimum oncological control with minimum morbidity. This is facilitated by a clear understanding of the prognostic variables that predict disease outcome in patients with p16+ OPSCC. To provide prognostic estimates, pathological staging and histopathological parameters are usually superior to clinical staging. However, knowledge of pathological predictors is sparse, mainly because of commonly employed nonsurgical management policies utilizing chemoradiotherapy. Minimally invasive approaches to the oropharynx, particularly transoral laser microsurgery (TLM), are well-reported effective primary treatments for oropharynx cancers. From such series, it is feasible to conduct a detailed appraisal based on pathologic information from surgical specimens of both the primary and neck, to establish prognosticators unique to p16+ oropharynx cancer patients.
A prospectively assembled database of oropharynx cancer patients treated with primary TLM ± neck dissection ± adjuvant therapy from 1996 to 2010, analyzed retrospectively for survival and recurrence.
The fundamental inclusion criteria were: 1) previously untreated biopsy-proven OPSCC treated with primary TLM ± neck dissection, 2) diffuse p16 positivity in the surgical specimen, 3) availability for adjuvant therapy, if indicated, and 4), minimum follow-up of 12 months or to death. Cox proportional hazard regression analyses were used to identify variables that were prognostic for disease-free survival (DFS), the primary end point of the study, as well as disease-specific survival (DSS) and overall survival. Kaplan-Meier survival estimates and patterns of disease recurrence were also assessed. We also explored concordance for T and N staging, when assessed by clinical (cT, cN) and pathological (cT, pT) measures.
Of 211 patients in the TLM database, 171 met all the eligibility criteria. The median follow-up was 47 months. The 3- and 5-year Kaplan-Meier estimates for DFS were 91% and 88%, respectively, whereas for DSS they were 95.5% and 94.4%, respectively. A total of 12 (7%) recurrences occurred: two local, four regional, and six distant. Of all T-stage categories, pT4 tumors were strongest predictors of poorer DFS. cT4 tonsil primaries, ever smoking status, three or more metastatic nodes, pN2b+ stage, and radiation-based adjuvant therapy were other prognosticators for DFS. Angioinvasion and T3-T4 tumors were prognostic for reduced DSS, although smoking parameters were not. Extracapsular spread, N stage, and margins were nonprognosticators. Recursive partitioning analysis defined high- and low-risk groupings of prognosticators. Downstaging of clinical T stage was observed for 31% of tumors on application of pathological classification.
We document a well-delineated set of prognostic variables that specifically and accurately identify individuals at risk of reduced outcomes in an otherwise good prognosis p16+ OPSCC cohort. Based on these prognosticators, appropriate patient counseling, adjuvant treatment recommendations, and stratification for trials can more accurately be made. We also observed an additional edge conferred by TLM toward more accurate clinical as well as pathological T staging.