Redefining Candidates for Deintensification in Locoregionally Advanced P16+ Oropharyngeal Cancer Based on Relative Risk.

PURPOSE

Randomized trials have found that patients with locoregionally advanced p16+ oropharyngeal squamous cell carcinoma (OPSCC) do not benefit from treatment deintensification, even among favorable risk groups. Although various methods have been used to identify candidates for treatment deintensification, the optimal approach is unknown.

METHODS AND MATERIALS

We conducted a multi-institutional cohort study of 444 patients with previously untreated p16+ OPSCC undergoing definitive radiation therapy with or without systemic therapy between 2009 and 2022. We compared the following 2 approaches for identifying candidates for deintensification: (1) favorable versus unfavorable risk, using NRG-HN005 eligibility criteria, and (2) low versus high relative risk of cancer events, using the Head and Neck Cancer Intergroup predictive classifier ("omega score"). We tested differences in outcomes and systemic therapy allocation by risk group using multivariable Cox models, competing event models, and logistic regression, and compared characteristics of hypothetical deintensification trials using the 2 approaches. Progression-free survival events were defined as cancer recurrence (locoregional or distant) or death from any cause.

RESULTS

Median follow-up time was 52 months; 120 patients (27.0%) were favorable risk; a different 120 patients had low omega score; 28 patients (6.3%) met both criteria; 184 patients (41.4%) had discordant classification. On ordinal logistic regression, decreasing omega score was associated with a statistically significantly lower odds of receiving intensive therapy (normalized odds ratio, 0.37 per SD; 95% CI, 0.24-0.57), with a greater magnitude than favorable risk group (odds ratio, 0.66; 95% CI, 0.44-0.99). Among patients receiving cisplatin and/or platinum-based induction (n = 374), favorable risk was associated with significantly improved progression-free survival (hazard ratio, 0.59; 95% CI, 0.36-0.99), whereas lower omega score was associated with a significantly decreased relative hazard for cancer events (relative hazard ratio, 0.18; 95% CI, 0.070-0.46). In simulations, selecting patients with low omega scores increased the efficiency of hypothetical noninferiority trials.

CONCLUSIONS

Considering patients' relative risk of cancer events can help define optimal populations for treatment deintensification in p16+ OPSCC.