Introduction: The relevance of clinical trial data to real-world results is often questioned. Recent technological and analytical advancements in “big data” research allow for reliable abstraction of patient-level information from electronic health records (EHRs). A retrospective analysis of data extracted from the Flatiron Health database consisting of > 1.5 M patients treated at > 255 cancer clinics across the United States was undertaken to evaluate real-world outcomes of patients with metastatic pancreatic cancer (mPCa) who were treated with first-line (1L) nab-paclitaxel plus gemcitabine (nab-P/G). Methods: Data for patients with mPCa who were diagnosed between 01 March 2015 and 31 October 2015 and treated with 1L nab-P/G (initiated within 60 days of metastatic diagnosis) were electronically and manually abstracted from the EHR via technology-enabled chart abstraction. Patients were followed until death, loss to follow up, or data cutoff (31 July 2016). Median progression-free survival (PFS) and overall survival (OS) were calculated for all patients and those who received a subsequent 5-fluorouracil (5-FU)– or capecitabine-based regimen using Kaplan Meier methods. Outcomes were measured from the start of 1L nab-P/G treatment. Patient and disease characteristics, laboratory values, treatment patterns, and other medical information for the predefined cohort are reported. Results: 251 patients met the inclusion criteria within the study timeframe. The study population was drawn from 84 US community (92% of patients) and academic sites (8% of patients). The median follow-up was 7.8 months (Q1, Q3: 4.4, 10.9) and the number of deaths was 154 (61%) at the data cutoff. The median age was 68 years (range, 34 - 84 years), 50% were men, and 65% were white (13% unknown). At the initiation of nab-P/G, 50% had an Eastern Cooperative Oncology Group performance status (ECOG PS) of 0 - 1 (8% had ECOG PS 2 - 3; 42%, missing); 73% and 22% had liver and lung metastases, respectively (nonmutually exclusive categories). Carbohydrate antigen 19-9 levels were available for 171 patients (68%), and the median level at the start of 1L treatment was 1186.0 U/ml. Median duration of 1L nab-P/G treatment was 4.4 months. Among those who discontinued 1L treatment prior to the data cutoff (n = 172), the most common reasons for treatment discontinuation were disease progression (66%) and toxicity (14%). PFS and OS for all patients and for those who received a subsequent regimen containing either 5-FU or capecitabine (102 patients [40.6%]) are reported in the table. Conclusion: Results from this preliminary real-world analysis are consistent with those from the Phase III MPACT trial and provide evidence for the effectiveness of nab-P/G in 1L and among patients who receive subsequent therapy.P-216 Table Open table in a new tab