Importance: The historically reported rates of subclinical cervical nodal metastases in oropharyngeal squamous cell carcinoma (OPSCC) predate the emergence of human papillomavirus as the predominant causative agent. The rate of occult nodal disease with changing etiology of OPSCC is not known, and it is challenging to anticipate which patients will be upstaged postoperatively and will require adjuvant therapy. Objective: To assess the rate of nodal upstaging and occult extranodal extension (ENE) in a multi-institutional population of patients with pathologic (p)T1-2 OPSCC treated by transoral robotic surgery and neck dissection. Design, Setting and Participants: This retrospective, multicenter cohort study of 92 participants at 2 US institutions (Albert Einstein College of Medicine, Bronx, New York [n = 38], and Icahn School of Medicine at Mount Sinai, New York, New York [n = 39]) and 1 Canadian institution (Princess Margaret Hospital, Toronto [n = 15]) examined the rate of postoperative pathologic upstaging for 92 patients with pT1-2 OPSCC undergoing transoral robotic surgery with neck dissection from August 2007 to December 2016. A neuroradiologist at each site blinded to final pathologic diagnosis reviewed preoperative imaging; these findings were compared with operative pathology and applied for tumor staging using the eighth edition of the American Joint Committee on Cancer Cancer Staging Manual. The statistical analysis was performed on December 18, 2018. Main Outcomes and Measures: Occult pathologic nodal disease and change in nodal category postoperatively. Results: Of 92 patients who met the inclusion criteria, 76 (83%) were male, and they had a mean (SD) age at surgery of 59.5 (10.5) years; 70 patients (84%) with available p16 status were positive. Five of 18 patients (28%) who had no evidence of nodal disease on imaging had occult pathologic nodal disease. Seven of 32 patients (22%) presenting with no nodal disease or with a single metastatic node on imaging received pathologic upstaging because of multiple positive nodes, indicating implementation of additional adjuvant treatment not anticipated after a priori imaging. Changes included 12 patients (13%) who had pathologic nodal upstaging and 12 (13%) with pathologic nodal downstaging in the eighth edition of staging. In the cohort, 24 patients (27%) had pathologic ENE, and 5 of 39 patients (13%) had occult ENE in the absence of radiographic evidence. Conclusions and Relevance: Predicting pathologic staging preoperatively for patients with OPSCC undergoing transoral robotic surgery and neck dissection remains a challenge. Although nodal size, tumor size, and location do not help predict ENE, the presence of nodes on imaging and nodal category may help predict ENE. Our findings suggest a small proportion of patients might benefit from further adjuvant therapies not predicted by preoperative imaging based on occult nodal upstaging and ENE.