Background: Patients with HPV associated squamous cell carcinoma of the oropharynx have a favorable outcome with respect to tumor control and overall survival (OS) after treatment with chemoradiotherapy (CRT). The leading cause of death in these patients is reported to be peripheral metastases. Because of the favorable outcome for these patients, the medical community tries to find methods to ameliorate treatment intensity, mainly in order to reduce long term morbidity. However, between 10 and 15 % of patients relapse locally in the primary tumor as first site of disease failure. Methods: In this multicentre, randomized, controlled, phase II trial 152 patients with locoregionally advanced oropharyngeal cancer were randomized in a 1:1 ratio to either radiotherapy with cetuximab (arm B) versus the same regimen preceded by 2 cycles of induction chemotherapy (IC) with taxotere/cisplatin/5-FU (arm A). To decrease risk of local failure, an escalated radiation (RT) dose of 74.8 Gy was delivered to T3/T4 tumors and to T2 tumors > 3 cm in the base of tongue. Eligibility criteria included patients 18-75 years, ECOG performance status 0-1 and adequate organ functions. Primary endpoint of the study was to compare PFS between the treatment arms, secondary objectives were recurrence pattern, locoregional control, OS and toxicity. Results: PFS at 2 years was 84.2% (95%CI 76.4-92.8) in arm A and 78.4 (95%CI 69.5-88.3) in arm B (p = 0.20). At the time of analysis there were 26 disease failures, 9 in arm A and 17 in arm B. In arm A there were 3 local, 2 regional and 4 distant relapses as first site of recurrence, and in arm B 4, 4 and 9 relapses in corresponding sites. Local relapse rate as first site of failure was low, i.e. in 4.7%. Local failures constituted 7 relapses out of all 26 (27%), and in 4 of these local failures (57%), the escalated dose was given. In arm B, patients who had no IC, there were more than twice as many patients who had distant metastases as first site of relapse compared to arm A (n.s.). No patient who had a response to IC in the primary tumor or in regional lymph nodes, respectively, measured with CT or MRI, of either CR+CR or CR+PR, had any recurrence, local, regional or distant. So, IC could identify 22 out of 73 evaluable patients (30%) in arm A, who never had any recurrence during follow up. Conclusions: A high incidence of distant metastases as first site of failure is recognized while local failure rate is low. However, despite a high RT dose of 74.8 Gy, delivered to the primary tumor, relapses can appear in this site. Radiological response of tumor to IC could identify patients with no tumor relapse, at least within the first years of follow up, possible candidates for de-escalation treatment protocols. Clinical trial information: 2009-013438-26.