Background: The incidence of OPSCC has risen rapidly, due to an epidemic of human papillomavirus (HPV) infection. Radiation therapy (RT) has historically been the standard treatment, but transoral robotic surgery (TORS) has surpassed RT in the US as the most common approach, based on assumptions of reduced toxicity or improved quality of life (QOL). No randomized trials have previously compared these treatments. Methods: The ORATOR trial (NCT01590355) enrolled patients with T1-T2 N0-2(≤4 cm) OPSCC amenable to TORS. We randomly assigned patients, stratified by p16 status, to RT (70 Gy/35 fractions, with chemotherapy if N1-2) vs. TORS (± adjuvant [chemo]RT based on pathology). The primary endpoint was a definitive comparison of swallowing QOL at 1-year using the MD Anderson Dysphagia Inventory (MDADI), powered to detect a 10-point improvement (a clinically-meaningful change [CMC]) in the TORS arm. Secondary endpoints included adverse events (AEs), other QOL outcomes [including EORTC scales, the Voice Handicap Index-10, Neck Dissection Impairment Index, and Patient Neurotoxicity Questionnaire], overall- and progression-free survival (OS, PFS). All analyses were pre-specified and intention-to-treat. Results: Between 2012 and 2017, 68 patients were randomized (n = 34 in each arm), in Canada and Australia. Median age was 59 years; 87% were male. Primary tumor sites were palatine tonsil (74%) or base of tongue (26%). Arms were well-balanced for baseline factors, including p16 status (88% in each arm). Median follow-up was 27 months. MDADI scores at 1-year were statistically superior in the RT arm (mean ± SD: 86.9 ± 11.4 vs. 80.1 ± 13.0 in the TORS arm; p = 0.042), but not meeting the definition of a CMC. For the other QOL metrics, outcomes were similar at 1-year. Feeding tube rates at 1-year were 3% (n = 1) vs. 0% respectively. Rates of treatment-related grade ≥2 AEs were similar (91% vs. 100%, p = 0.24), with more neutropenia, constipation and tinnitus in the RT arm and more trismus in the TORS arm (all p < 0.05). There was one TORS bleeding-related death. OS and PFS were similar. Conclusions: RT had superior swallowing QOL scores at 1 year compared to TORS, but the difference was not a CMC. Toxicities differed between the arms. This study provides the first level 1 evidence to inform patients of the QOL impact of both approaches. Clinical trial information: NCT01590355