Development of a Novel Composite Pathologic Risk-Stratification for Surgically Resected HPV+ Oropharyngeal Cancer

Presentation: AHNS-011
Topic: Mucosal - HPV Positive
Type: Oral
Date: Wednesday, May 1, 2019
Session: 1:00 PM - 1:10 PM Best of Mucosal HPV Positive Abstracts
Authors: John D Cramer, MD, Yusuf Dundar, MD, Jeffrey Hotaling, MD, Naweed Raza, MD, George Yoo, MD, Ho-Sheng Lin, MD
Institution(s): Wayne State University School of Medicine

IMPORTANCE: Human papillomavirus-associated (HPV+) oropharyngeal squamous cell carcinoma (OPSCC) is a distinct form of head and neck squamous cell carcinoma (HNSCC) requiring a unique staging system to accurately predict survival. However pathologic risk-stratification for HPV+ OPSCC largely remains based on the experience with HPV-unassociated (HPV-) HNSCC.  

OBJECTIVES: We sought to compare the survival discrimination of the traditional pathologic risk-stratification system for HPV+ OPSCC and HPV- HNSCC. We then derived a novel pathologic risk-stratification system for HPV+ OPSCC to improve survival discrimination.

DESIGN, SETTING AND PARTICIPANTS: We identified 15,324 patients with nonmetastatic HNSCC treated with upfront primary surgery and neck dissection in the National Cancer Data Base from 2010-2013. We compared traditional pathologic risk-stratification for HPV+ OPSCC and HPV- HNSCC. We derived a novel pathologic risk-stratification system from Cox models to improve performance for HPV+ OPSCC that incorporated the composite score of pathologic adverse features.

MAIN OUTCOMES AND MEASURES: Survival discrimination of pathologic risk-stratification systems were measured with concordance indices.

RESULTS: Traditional pathologic risk-stratification results in wide separation of survival curves for HPV- HNSCC (5-year overall survival 76.2% for low-risk, 54.5% for intermediate-risk and 40.9% for high-risk) but not for HPV+ OPSCC (5-year overall survival 93.2% for low-risk, 88.9% for intermediate-risk and 83.7% for high-risk).

Survival discrimination with traditional pathologic risk-stratification was good for HPV- HNSCC with a concordance index of 0.68 but poor for HPV+ OPSCC with a concordance index of 0.58. We empirically derived a novel composite risk-stratification system for HPV+ OPSCC. Traditional pathologic features that were prognostic included T-stage, lymphovascular invasion, positive margins, number of pathologic lymph nodes, and degree of extranodal extension and these were assigned 1-3 points based on their empiric prognostic importance to derive a composite risk score. Composite risk-stratificaiton for HPV+ OPSCC improved survival discrimination to a concordance index of 0.73 for total score and 0.67 when scores were grouped into three categories (5-year overall survival 91.8% for 0-3 points, 83.1% for 4-5 points and 53.2% for 6+ points). The new composite risk-stratification system resulted in substantial decrease in the proportion of HPV+ OPSCC classified as high-risk patients from 47.2% to 6.6%. Confirming the potential clinical applicability of this composite pathologic risk-stratification system, adjuvant treatment with radiation did not improve survival for patients with a low-risk composite risk-score (0-3) but did improve survival for patients with an intermediate- (4-5) or high-risk (6+) score. Adjuvant chemotherapy did not impact survival regardless of composite risk-score.

CONCLUSIONS AND RELEVANCE: Current pathologic risk-stratification suffers from poor survival discrimination in HPV+ OPSCC and misclassifies many patients with a favorable prognosis as high-risk. We derived a novel composite pathologic risk-stratification system for HPV+ OPSCC that improves survival discrimination.

Figure 1: Comparison of Overall Survival with Traditional Pathologic Risk-Stratification versus Composite Risk Stratification

Cox proportional hazard model of A: Traditional pathologic risk-stratification in HPV- HNSCC, B: Traditional pathologic risk-stratification in HPV+ OPSCC and C: Composite pathologic risk-stratification score in HPV+ OPSCC. Models adjusted for age, sex, race, comorbidities, adjuvant treatment, hospital and path risk-stratification.