Westin, Gustavo Figueiredo Marcondes

Objectives Prognostic factors for solid pseudopapillary neoplasms (SPN) of the pancreas have been incompletely defined in the literature. We aimed to use the National Cancer Database to describe survival and characterize prognostic factors in patients with pancreatic pseudopapillary neoplasms. Methods We identified 304 patients with pancreatic SPN diagnosed between 2004 and 2012 using the National Cancer Database. All patients were included in the survival analysis. Kaplan-Meier method, χ 2 , and log-rank tests were used for statistical analysis. Cox proportional hazard regression model was used for multivariate analysis. Results The median age at diagnosis was 36.5 years. Eighty-five percent of patients were female, 69% were white, and 25% were black. Metastases were observed in 5% of patients. Median tumor size was 5.1 cm. Ninety-two percent of patients underwent primary tumor resection. At 60 months, 98% of patients who underwent resection were alive, and 40% who did not undergo resection were alive. On multivariate analysis, female sex, resection of primary tumor, and absence of metastasis were correlated with improved survival. Conclusions Patients with SPN who undergo resection have an excellent survival at 5 years. Surgery should be considered in all patients with this diagnosis.

Chronic Helicobacter pylori ( H. pylori) infection is a major gastric adenocarcinoma (GA) risk factor. GA disproportionately affects U.S. Hispanics compared with non-Hispanic Whites (NHWs). Since H. pylori infection studies in Hispanics are few, infection rates in Hispanic and NHW men in Bexar County were compared, and relationships with ethnicity and obesity examined. Age- and zip code-matched participants from a community-dwelling cohort were randomly selected. Sera from 284 men were analyzed by enzyme immunoassay for H. pylori antibodies. Adjusted risk ratio estimation for matched data was conducted to identify differences. Hispanics had a markedly higher prevalence of infection (30.3%) than NHWs (9.2%). Matched risk ratio (mRR) analyses revealed a strong association between H. pylori seropositivity and Hispanic ethnicity (mRR = 3.31; 95% CI [1.91, 5.73], adjusted by BMI, smoking status, and family history of cancer (mRR range = 3.28-3.89). BMI mRRs (range = 1.19-1.22) were significant in all models. In this cohort, Hispanic men had higher H. pylori infection rates than NHWs, and parallel the disproportionately higher rates of GA; obesity contributes to this higher prevalence. Future studies should address country of origin, acculturation, and other factors influencing obesity to further elucidate risk of GA in Hispanic populations.

404 Background: Testicular sarcomas represent less than 1% of all testicular tumors in adults. Because of their rarity, little is known about the features and outcomes of these tumors. We present an analysis of the prognosis and survival predictors in patients with testicular sarcomas from the Surveillance, Epidemiology, and End Results Program (SEER). Methods: Using the SEER registry, we identified adults with testicular sarcomas between 1980 and 2013. Cases were identified using a combination of ICD-O-3 site and histology codes. The JMP and SEER*STAT 8.3.2 were used for statistical analysis. Overall survival (OS) was analyzed using the Kaplan-Meier method. The prognostic effect of variables was studied using Cox proportional hazards models. Results: We identified 264 patients with testicular sarcomas with a median follow-up time of 9.2 years. The median age at diagnosis was 55 years (range 15-94) and the median tumor size was 6.6 cm (range 0.7-45). 20% of patients with documented staging presented with de novo metastatic disease. Rhabdomyosarcomas were the most common histology (37%). The median OS was 129 months (95%; CI 85-180). OS rates at 1 year and 5 years were 85% and 62% respectively. 53% of deaths were related to cancer. 1-year OS rates for localized, locally advanced and metastatic disease were 91%, 93% and 53% (P<0.0001). Variables that were significant for survival in univariate analysis included age, histology, stage, and surgical resection. Factors that negatively affected survival in multivariate analysis were age >55 years (HR 2; 95% CI, 1.1-3.7), advanced stage (HR 4.8; 95% CI, 2.7-8.4) and surgical unresectability (HR 3.7; 95% CI, 1.1-9.3). Patients with rhabdomyosarcoma had favorable outcomes compared to other histologies (59% vs 17% alive at 10 years, P=0.01). Conclusions: Testicular sarcomas are rare entities with overall good prognosis after surgical resection. Patients presenting with distant metastases, however, have worse outcomes with only about 50% of the patients alive at 1 year. Rhabdomyosarcomas are the most common histology encountered and have relatively better survival. Younger age, earlier stage and surgical resection are factors associated with significantly better outcomes.

420 Background: Advanced biliary cancers respond poorly to chemotherapy and have a dismal survival. Currently gemcitabine-cisplatin (GC) is the standard of care for patients with advanced disease. The efficacy of second line treatment options is not well-defined. We retrospectively evaluated outcomes of patients undergoing second line treatment for advanced biliary cancers at Mayo Clinic. Methods: After obtaining approval from the institutional board review (IRB), we identified 42 patients with advanced and metastatic biliary cancers that progressed after first line (GC) and received second line treatment from 2007-2014. All patients had adequate follow up data. JMP software was used for statistical analysis. Kaplan-Meier method and log-rank tests were used for survival analysis, and multivariate cox proportional hazards model was used to evaluate the prognostic effect of pertinent clinical variables. All tests were two sided and a P value of < 0.05 was considered significant. Results: The median age at diagnosis was 61.5 (range 26-87). The most common regimens used for second line settings were 5-FU-based (69%) followed by gemcitabine-based regimens (19%). The overall response rate (ORR) was 9.1%, while the disease control rate (DCR; stable disease and partial response) was 33%. The median overall survival (mOS) for the entire cohort from first line was 16.8 months (95% CI 12.7-21.7) and the progression-free survival (PFS) was 9.16 months (95% CI 7.3-14.7). From second line start the mOS was 10.4 months (95%; CI 7.3-23.7), and PFS was 2.8 months (95% CI 2.3-4.0). 73% of patients who received second line therapy were alive at 6 months and 50% at 12 months. 25 patients received third line treatment after progression. An univariate and multivariate analysis were performed including age, sex, stage at diagnosis, regimen used, CA19-9 and total body weight loss > 5%, but no significant prognostic factors were identified. Conclusions: Patients with biliary cancers who progress after first line treatment have poor prognosis, but may benefit from second line regimens. 5-FU based regimens were the most commonly used in the second line setting, and 73% of patients who received second line therapy were alive at 6 months.

401 Background: Pancreatic neuroendocrine tumors (PNETs) are rare neoplasms comprising less than 2% of pancreatic tumors. Outcomes of PNETs vary by stage, grade, and clinical presentation. New therapies may have improved outcomes. We present an analysis of the epidemiology and prognosis of patients (pts) with PNETs evaluating a more recent period than previously reported. Methods: Using data from the SEER registry, we identified adults with PNETs between 2000 and 2013. Cases were identified using a combination of ICD-O-3 site and histology codes. The JMP and SEER*STAT 8.3.2 were used for statistical analysis. Overall survival (OS) was analyzed using the Kaplan–Meier method. The prognostic effect of variables was studied using multivariate Cox proportional hazards models. Results: We identified 5993 pts with PNETs. Over the study duration, there was an increase in the annual incidence from 0.3 to 1.2 per 100.000. This was largely explained by the increase in number of pts with localized disease. The OS for the entire cohort was 54 months (95%; CI 50-58) and 5-year OS rates in localized, locally advanced and metastatic disease were 81.7%, 62% and 26% respectively. New therapies were increasingly implemented after 2008 such as sunitinib, everolimus and capecitabine/temozolomide. OS was better for pts diagnosed after 2008 (HR 0.73; 95% CI, 0.69-0.79; P < 0.0001) and this remained significant after excluding pts with localized disease (HR 0.85; 95% CI, 0.78-0.92; P = 0.0002). Factors found to favorably affect OS included younger age, female sex, early stage, low grade, and surgery. In metastatic disease, pts who had surgery had better OS and this was significant after adjusting for possible confounders (HR 0.36; 95% CI, 0.3-0.45). Conclusions: In this large population-based study, we saw an increase in the annual incidence of PNETs mainly due to an increased number of pts with localized disease. The OS has improved over the study duration likely due to both stage migration and new therapies used in advanced disease. Younger age, female gender, low histologic grade, early stage and surgery were found to be factors that favorably affected survival. Surgery was found to be a predictor of better outcomes in metastatic disease.

179 Background: GISTs arise from interstitial cells of Cajal anywhere within the gastrointestinal tract, but those of the esophagus and GEJ are exceedingly rare ( < 1% of cases). Our aim was to characterize clinicopathological features and outcomes of esophageal and GEJ GISTs and compare survival with a population-based registry (SEER). Methods: 28 cases were identified using the Mayo Clinic Cancer Registry from 1997 to 2016, and 378 cases from the SEER 18 registry from 2000 to 2013. Mayo cases were re-reviewed by Mayo pathologists. We analyzed patient characteristics, tumor location, TNM staging, mitotic index, molecular diagnostics, IHC staining, and histomorphology. JMP software was used to calculate time to recurrence (TTR) and overall survival (OS) utilizing Kaplan-Meier and log-rank. Results: At Mayo, 60% of tumors arose in the distal esophagus. Mean tumor size was 5.73 cm. 42% of cases were low grade. None had nodal involvement. Five cases were metastatic at diagnosis, all to the liver. 86.7% had spindle cell morphology. IHC staining was positive for KIT in 92%, followed by CD34 (52%), DOG-1 (24%), and actin (16%). Mutational analysis was done on 10 cases and KIT mutation was found in 8 cases; 2 were wild-type. Two-thirds of patients underwent surgery, mostly esophagectomy. 14 patients received adjuvant imatinib. Median time to diagnosis from symptom onset was 2 months; median time to surgery from diagnosis was 1 month. Of 10 patients with complete follow-up, 2 had recurrences: 1at the anastomotic site and 1 in the liver. Median TTR from surgery was 36.5 months. Median follow-up time was 31.5 months. Median OS from diagnosis for the Mayo cohort was 129.5 months (95% CI 55.7-x). SEER cohort median OS was 81 months (95% CI 63-101), and females had superior OS compared to males (HR 0.67; 95% CI 0.48-0.93, p = 0.016). Conclusions: Patients undergoing surgical resection for esophageal and GEJ GISTs had a favorable prognosis, but recurrences occurred. The superior OS seen in the Mayo cohort suggests early resection and adjuvant imatinib may improve outcomes.

Abstract Breast cancer diagnosis and death has risen in low to middle income countries (LMIC) in the past 20 years. Early detection, diagnosis and treatment in the LMIC setting are thought to impact morbidity and survivorship. The population of South Florida is composed of a significant portion of Afro-Caribbean immigrants from LMICs. We compared a cohort of recent Haitian immigrants with residents in Haiti who had a diagnosis of breast cancer to evaluate the effects of geographical location on presentation, treatment and outcomes. Methods: Retrospective cohort study. Participants were Haitian females with breast cancer that were living in Miami and presented to the University of Miami/Jackson Memorial Hospital (UM/JMH) between 2008 and 2014 and patients living in Haiti and presented to the Breast Cancer Treatment program at the Hospital Bernard Mevs Project Medishare (HBMPM) from 2013 to 2015 with a new diagnosis or who had been diagnosed as far back as 2008 and were presenting for continuation of care. The main outcome was the relationship between cohort location and breast cancer characteristics and cancer–specific survival by log-rank. Cox proportional hazards regression was used to compare breast cancer–specific survival overall and analyzed by stage and other significant explanatory factors. Results: 102 patients from UM/JMH and 94 patients from HBMPM were included in the analysis. The cohort in Haiti presented after longer median duration of symptoms (20 vs 3 months, p&lt;0.001) and more advanced stage (31% Vs 20% with locally advance and 27% vs 19% with metastatic disease). The median age of diagnosis was 55 years in Miami compared to 50 years in Haiti (p=0.03). In Miami all patients had estrogen receptor (ER) and her-2 status ascertained at diagnosis, whereas only 52% and 27% of the Haitian cohort had them. Tumor characteristics in Miami vs Haiti in patients with reported status: ER + 74% (n=102) vs 55% (n=49), p=0.02, her-2+ 20% (n=102) vs 28% (n=25), p=0.36, triple negative 27% (n=102) vs 28% (n=25),p=0.49. Treatment modalities and censoring was available for all the patients in Miami and 72 patients in Haiti. In Haiti, radiation therapy was not widely available compared to Miami. Once patients had known ER and her-2 status, there was no difference in treatment with hormonal therapy (p=0.97) but none of the patients in Haiti received her-2 directed therapy (p&lt;0.001). 50% of the patients with unknown ER status in Haiti received Tamoxifen. Twice as many Haitian women died after a median follow-up of 23 months (12% vs 24%), with median survival of 53 months for Haitians in Haiti (not reached for Miami cohort). Breast cancer specific survival was significantly different in the two cohorts (p&lt;0.000). After exploratory analysis, the risk of death was higher for Haitians in Haiti than Haitians in Miami [HR 3.13 (CI 1.50-6.51)] and was higher for more advanced disease [HR 2.58 (CI 1.67-3.99)]. No other variable or interaction reached statistical significance in our model. Conclusion: Haitian women diagnosed with breast cancer in Haiti experience a significantly worse outcome than Haitians living in Miami that appears to be related to a more advanced stage, younger age, more ER negative tumors and lack of timely effective treatments including restricted access to trastuzumab and radiation. In addition there is a difference in the age of breast cancer onset (Caucasian 62, African American 58, Haitian American 55, Haitian 50) and ER positivity (Caucasian 70%, African American 48%, Haitian American 74%, Haitian 55%) depending on race and ethnic group that raises the question of whether there are disparities within subgroups of women of African origin who have breast cancer. Confounding epigenetic-related variables may impact survival and they need further exploration in these populations. Citation Format: Alexandra Gomez, Vincent DeGennaro, Jr., Sophia George, Estefania Santamaria, Gustavo Westin, Gabriel Dieudina, Judith Hurley. Presentation, treatment, and outcomes of Haitian women with breast cancer in Miami and Haiti: Disparities in breast cancer. A retrospective cohort study. [abstract]. In: Proceedings of the Eighth AACR Conference on The Science of Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; Nov 13-16, 2015; Atlanta, GA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2016;25(3 Suppl):Abstract nr C55.