Re-Examining The Changes In The Pathologic Nodal Classification Systems For Hpv(+) And Hpv(-) Oropharyngeal Cancer

HPV(-) and HPV(+) oropharyngeal cancer (OPC) shared identical pathologic nodal (pN) classification systems for decades. In 2018, the American Joint Commission on Cancer 8th Edition split these diseases into radically different systems based on the belief that the biological impact of nodal metastases in HPV(-) and HPV(+) OPC was completely different. pN-classification for HPV(-) OPC remained similar, relying on lymph node (LN) size, laterality, and number, with extranodal extension (ENE) incorporated as a novel factor. By contrast, HPV(+) OPC pN-classification was based entirely on LN number, ignoring ENE. Given that LN number and ENE are dominant prognostic factors in nearly every head and neck cancer, we questioned the widely held belief that HPV(+) and HPV(-) OPC require widely divergent pN-classification systems. Patients with OPC undergoing surgery with or without adjuvant therapy diagnosed from 2010 to 2015 with complete LN and HPV data were identified in the National Cancer Database. Multivariable Cox regression models were constructed to determine the association between survival and nodal factors. The non-linear association between LN number and survival was modeled with restricted cubic splines. Recursive partitioning analysis (RPA) was used to derive a novel pN-classification system based on LN(+) number, size, laterality, lower LN(+), and ENE. Among the 7090 patients that met inclusion criteria, 73% (N = 5199) were HPV(+) and 27% (N = 1891) were HPV(-). Using multivariable non-linear modeling, mortality risk increased continuously with each additional LN(+) in both HPV(+) and HPV(-) OPC with nearly identical slope. Mortality risk increased by 19.1% (hazard ratio [HR], 1.19 [1.17-1.21], P<0.001) and 18.5% (HR 1.19; [1.16-1.21], P<0.001) with each additional LN for HPV(+) and HPV(-) OPC, respectively up to an identical change point of 5 LN. ENE was also an independent predictor of mortality in both HPV(+) (HR 1.73 [1.36-2.20], P<0.001) and HPV(-) OPC (HR 1.55 [1.20-1.99], P<0.001). In RPA for both diseases, LN number was the dominant nodal factor driving survival, with ENE a secondary factor, producing nearly identical classification systems for HPV(+) OPC [N1: 1-4 LN(+); ENE(-), N2: 1-4 LN(+), ENE(+); N3: >4LN(+)] and HPV(-) OPC [N1: 1-6 LN(+), ENE(-); N2: 1-6 LN(+), ENE(+);N3: >6 LN(+)]. Given these similarities, and the identical change point of 5 LN(+) identified in non-linear modeling, we propose a simple and unifying pN-classification system for OPC: N1: 1-5 LN(+), ENE(-); N2: 1-5 LN(+), ENE(+); and N3: >5 LN(+). HPV(+) and HPV(-) OPC exhibit remarkably similar relationships between nodal characteristics and relative mortality, despite absolute mortality being higher in HPV(-) OPC. Number of LN(+) is the dominant nodal factor driving survival, with ENE an important secondary factor, in both diseases. A simplified, accurate, and unified pathologic nodal classification for OPC is possible using these two factors.