Background: TNM-8 staging separates oropharyngeal squamous cell carcinomas (OPSCC) into human papillomavirus (HPV)-mediated and-unrelated OPSCC based on p16INK4a overexpression (p16þ), as surrogate marker for HPV. However, OPSCC is histologically and clinically heterogenous including tonsillar and base of tongue squamous cell carcinomas (TSCC and BOTSCC respectively), and carcinomas of soft palate and walls (otherOPSCC). The significance of HPV is established in TSCC/BOTSCC, while its role in otherOPSCC is unclear, which is not considered in TNM-8. Here, p16þ was therefore evaluated in relation to overall survival (OS) and tumor stage per OPSCC subsite. Patients and methods: All 932 patients, treated with curative intent in Stockholm 2000e2016 with OPSCC, previously analyzed for p16 expression, were included. Clinical data, including stage and OS, was collected retrospectively. Results: Patients with p16þ otherOPSCC had significantly poorer OS compared to patients with p16þ TSCC/BOTSCC (p Z 0.005) and their survival was similar to that of patients with p16-otherOPSCC/TSCC/BOTSCC. Moreover, patients with TNM-8 stage I-II and p16þ otherOPSCC had a significant poorer OS compared to patients with p16þ TSCC/BOTSCC

e15553 Background: Phase III studies with targeted agents have shown improved survival for mRCC. There is limited outcome data for the whole mRCC patient group. We collect data in a clinical registry at Karolinska for all patients with mRCC in the Greater Stockholm Region since 2007. Here we present the relationship between ECOG performance status and treatment outcome in this population. Methods: Between 2007 and 2012, 273 patients with mRCC were included in the registry. Patients were divided in two groups: Those that were treated with targeted agents and non- treated patients (symptomatically treated). Results: ECOG PS, available for 63%, was lower for patients treated (n=127) compared to patients not treated with systemic therapy (n=46) (median 1 vs. 2. Median OS from diagnosis to death in respective group was 731 day (n=106) and 269 days (n=55). The median OS from date of first metastasis to death was 500 days and 176 days respectively. In the treated group median OS from start of systemic therapy to death was 321 days. There was a difference in median OS from start of systemic therapy to death between patients who received 1 (n=52)(A), 2 (n=32)(B), 3 (n=16)(C) or > 3 lines (n=6)(D) with a median OS of 160.5 days, 395 days, 641 days and 728.5 days respectively. Mean PS at first line treatment was 1.36 for patients who received only one line and ≤ 1 for patients who received several lines. 13/18 patients with ECOG PS 0 at 1st line received >1 line, corresponding numbers for PS 1, 2 and 3 at 1stline were 21/40, 3/14 and 0/4. The average/ medians ECOG PS at time for the last treatment line were 1.36/1 (A), 1.47/1 (B), 1.64/2 (C) and 1.67/2 (D). Median OS from start of systemic treatment to death was dependent on PS at the start of treatment, ECOG 0 (n=18) 622 days, 1 (n=40) 724 days, ECOG2 (n=14) 226 days, ECOG3 (n=4) 44 days. Irrespective of treatment line, patients with PS 0-1 at initiation of any line of therapy had a 55% chance (47/86) of receiving a subsequent line of treatment, while patients with PS 2-3 only had a 19% chance (7/37). Conclusions: Patients with a good performance status, ECOG 0-1 at treatment start, had longer OS and received more lines of treatment. Patients with ECOG PS 3 could in our hands not be treated at all with targeted agents.