2018-19 Section 7-Neoplastic and Inflammatory Diseases of the Head and Neck eBook

Oral Oncology 73 (2017) 152–159

K.Y. Zhan et al.

treatment, creating two unique clinical and pathologic staging systems [3,4,8] . This is testament to ongoing controversy surrounding optimal HPV+ OPSCC management, as current National Cancer Center Net- work (NCCN) treatment guidelines do not distinguish a superior treat- ment modality (nor do they stratify treatment by HPV status) [9] . For surgically treated HPV+ OPSCC, nodal (N) staging now requires pathologic evaluation for the number of regional lymph nodes involved (pN0 = 0 nodes, pN1 = 1 – 4 nodes, pN2 = ≥ 5 nodes; pN3 was re- moved as studies show it behaves similar to pN1 in surgically-managed patients), regardless of laterality, nodal size, or negative histologic features such as extranodal extension (ENE). T4a and T4b tumors were consolidated to T4, as they failed to show di ff erent survival outcomes [3] . Lastly, pStage IV disease was reserved for distant metastatic or M1 disease. These proposals were studied in a retrospective, fi ve-center, international study of 704 HPV+ OPSCC patients managed with fi rst- line surgery +/ − adjuvant treatment [8] . Unlike clinical staging changes that have been validated with the National Cancer Data Base (NCDB), pathological staging modi fi cations have yet to be appraised in similar manner. The objective of our study is to validate these changes with the NCDB. We characterize the pathologic staging changes from the AJCC 7th to 8th editions and appraise survival outcomes in the 8th edition ’ s pathologic nodal and overall staging. The NCDB is a hospital-based cancer registry that collects high quality, de-identi fi ed, and internally appraised cancer data from over 1500 US hospitals approved by the American College of Surgeons and Commission on Cancer (CoC). It captures approximately 70% of all US cancer diagnoses – over a million unique cancer cases annually [10] . We performed a retrospective cohort study with the NCDB for adults with previously untreated, HPV+ OPSCC without distant metastatic disease, diagnosed in 2010 – 2014 (HPV status was added to NCDB coding in 2010). The Ohio State University Institutional Review Board deemed this exempt from review. Squamous cell carcinoma was identi fi ed with International Classi fi cation of Diseases for Oncology, 3rd edition (ICD-O-3) histologic codes 8070 (squamous cell carcinoma, NOS), 8071 (squamous cell carcinoma, keratinizing, NOS), and 8083 (basaloid squamous cell car- cinoma). ICD-O-3 subsite codes in our cohort include C019, C024, C051, C052, C090, C091, C098, C099, C102, C103, and C109. C108 (Overlapping lesion of oropharynx, NOS) was excluded due to concerns for site misclassi fi cation. Patients with high risk HPV documented (HPV16, 18, or high risk NOS) were studied. Patients under the age of 18, cases missing pathologic staging or lymph nodes analysis, and those being treated with palliative intent were removed. Since the AJCC 8th edition reserves pStage IV for distant metastatic disease, we excluded patients with M1 disease. Of note, extent of neck dissection and method of ascertaining HPV status are not recorded in the NCDB. A fl ow dia- gram of these inclusion/exclusion criteria is shown in Fig. 1 . Covariates for analysis include: age, race, sex, Charlson/Deyo co- morbidity score, insurance status, number of lymph nodes pathologi- cally involved and examined, treating facility type, pathologic ENE, extent of primary site surgery, receipt of chemotherapy and/or radio- therapy, and pathologic staging information. All hospital types included in this study are CoC-accredited and have the full host of diagnostic or treatment services either in-house or by referral. Facilities are further Inclusion and exclusion criteria Variables Methods Data source

Fig. 1. Flow diagram of inclusion & exclusion criteria.

classi fi ed by number of new cancer cases a year: 100 – 500 (community cancer program, CCP); > 500 (comprehensive CCP), > 500 with man- datory residency training in ≥ 4 programs (academic) [11] .

Statistical analysis

Data were imported and analyzed using SPSS statistical software version 23 (IBM SPSS Inc., Chicago). The Kaplan-Meier method was used to calculate and visualize overall survival (OS) curves; log-rank testing was performed for pair-wise survival comparisons. Four-year OS is reported as our data is from 2010 to 2014 and is consistent with previous HPV+ OPSCC studies using the NCDB [12] . Survival was calculated with the life tables function. Information regarding cause of death is not recorded in the NCDB, preventing disease speci fi c survival analysis. All statistical tests were two-sided. An alpha of < 0.01 was utilized for all statistical tests based on the Bonferroni correction for multiple comparisons. 3745 patients were included in our fi nal cohort. All patients un- derwent neck dissection and pathologic lymph node examination. Median age was 57 (standard deviation 9). Table 1 shows the dis- tribution of selected demographics, clinical, treatment, and disease- speci fi c variables in this cohort. The majority of cases were Caucasian (95%) males (84.6%) with tonsillar (66.7%) SCC, treated at academic hospitals (57.7%) and with private insurance (68.8%). A minority (5.8%) did not receive primary site surgery; 69.7% were coded as re- ceiving “ limited resection/partial pharyngectomy, ” which can include bilateral tonsillectomy. Most patients received radiotherapy (80.4%) or chemotherapy (53%). Table 2 shows the distribution changes from AJCC 7th to 8th edi- tions. Notably, proportion of pN1 disease shifted from 17.3% to 75.9% in AJCC 8th; Overall stage I disease for AJCC 8th increased to 80.2% from just 3.7% (all remaining staging will be presented in AJCC 8th edition unless speci fi ed). In our cohort, median number of lymph nodes positive for cancer was 2, interquartile range (IQR) = 1 – 3. Median number of lymph nodes pathologically examined was 27, IQR 15 – 39. Treatment patterns varied signi fi cantly across nodal and overall pathologic staging. From pN0 to pN2, surgery alone trended down and triple modality therapy (surgery + chemoradiation) became the pre- dominant management (77.8% triple modality for pN2 vs. 15.4% in Results Staging distribution, treatment patterns, extranodal extension

177