Solitary fibrous tumors (SFT) of the retroperitoneum are rare spindle cell neoplasms, with a paucity of data on treatment outcomes. We hypothesized that surgical excision offered acceptable outcomes in SFTs. The National Cancer Database (NCDB) was used to identify patients with SFT from 2004 to 2011. Primary outcome measures were 30 day mortality and overall survival. Descriptive analyses were performed. Furthermore, a systematic review of published literature was conducted after creating a pre-specified search strategy. Of 51 patients in the NCDB, 58.8 % (n = 30) were males, with a median age 60 years (IQR 49–72 years). Median tumor size was 16 cm (IQR 11–21 cm). Surgical resection was performed in 92.2 % (n = 47) with 63.8 % (n = 30) having a margin negative resection. Peri-operative mortality was 2.1 % (n = 1). Of survival outcomes available for 18 patients, the median OS was 51.1 months. From the systematic review, we identified 8 studies, with 24 patients. Median age and tumor size was similar to the NCDB [47.5 years (IQR 39–66.5 years), 12 cm (IQR 7–17 cm)]. Majority [91.7 % (n = 22)] underwent surgical excision alone while one received adjuvant chemotherapy and none received radiation. After median follow up of 54 months (IQR 28–144 months), 79.2 % (n = 19) were alive without disease. Three patients (12.5 %) died of disease, one was alive with disease and one was lost to follow up. Recurrence was reported in 16.7 % (n = 4) of patients. Complete surgical excision is a viable treatment modality for retroperitoneal SFT leading to long term survival. Low recurrence rates would argue against the need for routine adjuvant radiation or chemotherapy.

379 Background: The role of radiotherapy (RT) for surgically resected intrahepatic cholangiocarcinoma (ICC) remains poorly defined. Radiotherapy is often considered when a positive resection margin exists. The present study sought to examine the impact of radiotherapy following liver resection. Methods: Patients with early stage ICC, who underwent surgical resection, were identified from the National Cancer Database (1998-2011). Patients were stratified by resection margin status and receipt of RT. Survival was analyzed by Kaplan-Meier method and a multivariate regression model was used to identify predictors of survival. Results: A total of 2,182 patients were identified. R0 status was obtained in 1,624 patients (74.4%). RT was delivered to 405 patients (R0=209, R1/R2=196). In the R1/R2 group, 196 patients received RT vs. 362 R1/R2 patients that did not receive RT. Survival for R0 vs. R1/R2 was 32m vs. 16.5m (p65years (Hazards Ratio [HR]: 1.20 (95%CI: 1.04-1.39), p=0.013), and positive resection margins (HR: 1.95 (95%CI: 1.65-2.30), p

718 Background: Observation alone after initial treatment of unresectable metastatic colorectal cancer is often anxiety provoking for physician and patient alike. The CAIRO 3 trial suggested the efficacy of maintenance capecitabine and bevacizumab in improving overall survival for such patients. We hypothesized that the cost effectiveness for maintenance capecitabine and bevacizumab would be worse than accepted population thresholds. Methods: Data from the CAIRO-3 trial was used to populate a semi-markov model, in which patients transitioned between different disease and complication based states. Transition probabilities were extratcted from the trial. Costs were determined from a thirs payer persective from the Medicare part B ASP drug pricing file. Utility was converted from the global quality of life scale. Incremental cost effectiveness ratios were calculated. Results: Cost of the maintenance arm after 10 cycles was $108,848 with a gain in 14.93 quality adjusted life months, while the quality adjusted life months gained at no cost in the observation arm was 13.67. This yielded an ICER of $1,036,648/QALY. Two way sensitivity analyses demonstrated dominance of observation across a wide range of parameters unless the cost per cycle was < $6250. Conclusions: Maintenance capecitabine and bevacizumab is not cost effective and is higher than the willingness to pay threshold for any developed nation. Reducing drug pricing is the only way to financially support the argument for this treatment strategy.

765 Background: Curative cancer surgery remains the cornerstone of treatment of localized malignancies. Extended post operative recovery (eLOS) is associated with debility and immunosuppression. We hypothesized that eLOS is associated with worse long term oncological outcome. Methods: The NCDB participant user file from 2004-13 was utilized to identify patients with localized gastrointestinal malignancies undergoing curative cancer resections. Quartiles (Qx) of length of stay (LOS) were calculated based on histology and site specific indicators. Survival modeling was performed using cox-proportional hazards and conditional models were created. Alpha of 0.05 was set for statistical significance. Results: Of 1,616,215 patients in the cohort, 577,860 patients were included who met inclusion criteria. Majority of patients were ≥65 years (56%) with a predominance of colon cancer (70%). The median length of stay (LOS) varied from 5 days (partial hepatectomy) to 10 days (esophagectomy). After adjusting for age, gender, stage, and site of disease, increasing length of stay was significantly associated with worse oncological survival with a dose response effect (Q1 HR 1.0, Q2 1.12 (1.10-1.14) p