e16737 Background: Brain metastases (BM) occur extremely rare in pancreatic adenocarcinoma (PDAC) and few data are available regarding those patients‘ care and outcomes. Methods: We performed a retrospective monocentric analysis of our database to identify patients (pts) diagnosed with PDAC and BM from July 1997 to December 2019. Results: 16 pts were eligible among 4900 pts diagnosed with PDAC in our institution (0.3%). Median age was 64 years (38.2-74.6) with 50% female. At diagnosis, 68.8% were metastatic including 27.3% with synchronous BM. About 1/3 pts had ≥ 2 lines of chemotherapy before BM. BM were discovered from neurological symptoms in 62% of cases and either with systematic brain CT/MRI or unknown in 19% for both. BM treatment was: surgery (25%), whole brain radiation therapy (RT) (43.8%), stereotactic RT (6.3%), radiosurgery (6.3%), best supportive care (BSC) or unknown (6.3%). At follow-up cutoff date (01/01/2020), most of pts were dead (75%), 2 were alive and 2 lost of follow-up. Mean interval between initial diagnosis and BM was 9.9 months (mths) (0-73). Median time to develop BM was different between pts with non-metastatic or metastatic disease at diagnosis: 16.3 mths (6.5-44) and 4.2 mths (0-36.1), respectively (HR = 0.43 (CI95: 0.14-1.09; p = 0.09)). Median overall survival (mOS) was 14.5 mths (1.6-80.2). Definitely, the non-metastatic group at diagnosis had better survival with mOS of 40.2 mths (24.7-80.2) compared to 6.5 mths (1.7-49.8) for the metastatic group (HR = 0.24 (CI95: 0,06-0,63; p = 0.012)). The median survival period after diagnosis of BM was only 3.4 mths (0.5-13.7). Pts who underwent BM surgery had better survival with a median survival from surgery of 5.5 months (4-13.7) compared to RT (0.8 mths (0.4-2)) or BSC (0.6 mths (0.5-5.9)). HR for surgery versus RT was 0.12 (CI95: 0.02-0.31; p = 0.003). After BM diagnosis, 43.8% of patients had a systemic chemotherapy, without objective response on BM. One interesting metastatic pt with BRCA1 mutation achieved a complete response (CR) after FOLFIRINOX. One BM occurred 2 years after diagnosis, was treated with surgery + RT but relapsed 4.5 mths later with new BM. Extra-cranial CR was persistent. This pt, still alive, had the longest survival period after diagnosis of BM (13.7 mths) and OS (49.8 months). Conclusions: To our knowledge, this is one of the largest cohort reported of BM in PDAC. Very few cases exist to guide therapy. Surgery appears to be the best treatment for BM, when feasible. Further investigations are still needed to understand the pathogenicity of BM in pancreatic cancer.

e14045 Background: First-line treatment of PCNSL consists in a polychemotherapy based on high-dose MTX with hyperhydratation until MTX elimination. Renal toxicity remains frequent with pejorative impact on renal function and patient survival. Our objective was to determine urinary predictive factors of MTX elimination delay. Methods: Patients with PCNSL followed in our institution were prospectively recruited from December 2020 to May 2021. Daily serum and urinary creatinine and ionogram were collected from the day before MTX administration until elimination. Patients characteristics, biological data and MTX concentration were recorded. Normal elimination time was defined by MTX ≤ 0.05 at 72 hours. Results: We collected data from 64 MTX cycles corresponding to 20 patients. Median age was 71 years (range, 33 - 86) and median Karnofsky Performance Status (KPS) was 70. Forty-four percent of cycles were full dose (3500 mg/m2). Median urinary osmolarities one day before, the day and 24 hours after MTX administration were 325 mOsm (range, 62 - 542), 316 mOsm (range, 58 - 671) and 291 mOsm (range, 56 - 554), respectively. Median time for MTX elimination was 94.7 hours (range, 47 - 205). Only one patient needed to receive MTX antidote (glucarpidase). The factors significantly associated with MTX elimination time were age (p < 0.001), KPS (p < 0.001) and urinary K+ concentration at 24 hours post MTX administration (u-K+) (p = 0.008). Using ROC curve analyses, we determine optimal cutoffs for age (80 years, p < 0.001, AUC = 0.825), KPS (70, p < 0.001, AUC = 0.808) and u-K+ (17.2 mmol/L, p = 0.008 AUC = 0.712). This u-K+ cutoff presented with a specificity of 81% and a sensibility of 60%. Focusing on patients aged < 80 years (52 cycles / 64), we confirmed that both KPS (p = 0.022) and u-K+ (p = 0.027) were significantly correlated with MTX elimination time by multivariate analysis. Then, we derived a 3-factors score including 5 subgroups: 1) aged ≥ 80 (N = 12); 2) age < 80, KPS < 70 and high u-K+ (N = 11); 3) age < 80, KPS < 70 and low u-K+ (N = 6); 4) age < 80, KPS ≥ 70 and high u-K+ (N = 8); 5) age < 80, KPS≥70 and low u-K+ (N = 21). The median times for MTX elimination were 112, 96, 95, 82 and 72 hours respectively. The normal elimination rates in each subgroup were 17%, 0%, 17%, 37% and 71% respectively. Conclusions: Age, KPS and u-K+ were correlated with MTX elimination time. We derived a predictive score based on these items that could be used in routine to anticipate MTX elimination delay and to potentially prevent the risk of renal toxicity. These results need to be prospectively confirmed in a larger cohort.