was defined by stage: ≥1 visit to a surgeon for stage 1 and ≥2 visits to a medical oncologist for stages 2 & 3. Adherence to CEA was defined as ≥2 levels drawn. CT and colonoscopy adherence were defined as completing each test within the time frame. Overall adherence was defined as completing all 4 exams within 10-14 months. Kaplan-Meier estimates were used to evaluate time to each adherence outcome. Effects of clinical factors on the outcomes were assessed via log-rank test. Results: In this interim analysis, 186 patients from 3 centers were included. Patient characteristics are shown in Table 1. Proportion adherent at 14 months to individual exams are presented in Table 2. 82% completed physician visits (median time= 2 months). Adherence to CEA levels was 86.5% (median time = 7 months) which was a similar rate to receipt of CT scan (86.1%, median time = 9 months). Adherence to colonoscopy (68.7%, median time = 12 months) was less. Only 54.5% of patients received all 4 recommended surveillance exams within 14 months after colon resection. Median time to completing all exams was 14 months. Factors positively associated with overall adherence included younger age (p=0.002), advanced stage (p=0.03), center (p=0.02) and black race(p= 0.02). Gender was not associated with overall adherence (p=0.51). Conclusions: Patients with resected CRC recruited from 3 NCI-designated comprehensive cancer centers showed moderate overall rates of adherence to published surveillance guidelines. Frequency of physician visits and CEA assessment were acceptable. While CT scan adherence was accept- able, it was typically performed too soon relative to surgery, while colonoscopy was performed less frequently. Patient, provider and system factors likely account for these findings. Review of additional patient data is ongoing to confirm and extend findings. Table 1: Patient characteristics (n=186 subjects) Table 2: Actuarial rates of surveillance adherence with median months to completion CI, confidence interval; All, adherence to physician visits, CEA, CT and colonoscopy Mo1736 Impact of Mailing Time and Season on Fecal Immunochemical Test (FIT) Positivity Jason A. Dominitz, Douglas J. Robertson, Dennis Ahnen, Kathy D. Boardman, Barbara Del Curto, Peter Guarino, Thomas F. Imperiale, Tassos Kyriakides, David Lieberman, Dawn Provenzale, Aasma Shaukat, Shahnaz Sultan, Robert Wallace Background: Fecal occult blood testing (FOBT) has been proven to reduce colorectal cancer mortality. The FIT has been shown to be superior to guaiac FOBT in test performance. However, hemoglobin degrades over time, leading to concerns about the impact of delays in processing after specimen collection. The CONFIRM Study is a multicenter, randomized controlled trial comparing screening colonoscopy to FIT, now enrolling 50,000 average risk Veterans. Participating sites cover the entire US mainland, as well as Puerto Rico, Hawaii and Guam. Aim: To determine the impact of season and time between specimen collection and laboratory processing upon the FIT positivity rate. Methods: Participants randomized to FIT are asked to submit one FIT sample (OC FIT-CHEK®, Polymedco) annually. All FIT kits are mailed to a central laboratory in Albuquerque, NM for processing. Initially, samples were returned via 1 st class mail. Subsequently, US Priority Mail has been used for all centrally distributed FIT kits. Upon receipt at the lab, FIT kits are refrigerated until processing with testing performed several days per week. All specimens are analyzed on the Diana (Polymedco Inc.) and hemoglobin results ≥100ng/ml buffer are deemed positive. FIT kits received >14 days after specimen collection are not analyzed, as per manufacturer instructions. The chi- sq test was used to compare positivity rate by season. Results: 12.4% of FIT kits arrived without a collection date recorded. Average time from specimen collection to receipt in the lab was reduced with introduction of Priority Mail shipping (6.1 days vs. 7.4 days, t-test p < 0.0001). Overall, 22,957 FIT kits have been processed with a positivity rate of 6.8%. Positivity rates were significantly different by season, with the lowest rate in the summer S-767 AGA Abstracts (p=0.007, Table). There was no significant difference in positivity rates between those samples received with a specified collection date (6.7%) and those with no collection date (7.1%, p=0.42). Conclusions: Requiring individuals to record the collection date would result in a significant number of unusable FIT kits. Our finding of no difference in the positivity rate supports processing and clinical use of these kits for cancer screening as requiring a replace- ment kit may result in decreased adherence. Introduction of Priority Mail shipping shortens the shipping time. However, we found no noticeable differences in positivity rate across the 14 day transit window. Our finding of significant differences in positivity by season is consistent with prior studies, though the magnitude of the difference is small. Avoidance of screening during summer months could introduce programmatic challenges but should be considered, though the impact on sensitivity and specificity is unknown. Efforts to assure stability of FIT specimens during transit should continue. (For the CONFIRM investigators) Association of Season and Shipping Time on FIT positivity Mo1737 Metabolic Factors in Addition to Aging and Current Smoking Are Independently Associated With the Prevalence of Colorectal Neoplasia in Japan: Analyses of Comprehensive Health Checkup Data Nagamu Inoue, Michiyo Takayama, Kanako Makino, Eisuke Shiomi, Ryoko Shimizu-Hirota, Toshifumi Yoshida, Kazuhiro Kashiwagi, Hiroshi Hirose, Suketaka Momoshima, Yoshinori Sugino, Yasushi Iwao OBJECTIVE: In Japan, colorectal cancer (CRC) is the third leading cause of cancer death. The incidence and mortality rate of CRC has been increased in Japan for the past decades, which corresponds with the economic development and concomitant shifts from traditional lifestyle towards a westernized lifestyle. We investigated the risk factors such as metabolic and lifestyle factors for the occurrence of colorectal neoplasia (CRN) by using comprehensive health checkup data. METHODS: We conducted a retrospective analysis in clinical practice at a single center. Among 6698 subjects who took comprehensive health checkup at our hospital between in August 2012 and March 2015, 1511 subjects who also underwent screening colonoscopy were enrolled. In the present study, CRN was defined as adenomatous polyp ≥ 5 mm in size and adenocarcinoma. A diagnosis of metabolic syndrome (MetS) was made by Metabolic Syndrome Diagnostic Criteria Exploratory Committee in Japan. Demographic characteristics, anthropometric measurements, visceral fat area (VFA) measured at the umbilical level by computed tomography, hematological metabolic parameters, degree of liver fat evaluated by ultrasonography, and current smoking and drinking habits were assessed. Association between variables and CRN was evaluated by univariate analysis using t-test, χ 2 test, Mann-Whitney test, and then by multivariate analysis using multiple logistic regression model. Multivariate analysis was conducted separately for men and women. A p value < 0.05 was considered statistically significant. RESULTS: One hundred and seventy subjects had CRN (11.3%) and 5 had invasive colorectal cancer (0.3%). Three hundred and fifty-eight subjects were diagnosed as MetS (24.9%; 31.3% in men, 13.0% in women) and presence of MetS was significantly associated with CRN both in men ( p = 0.027) and women (p < 0.01). Univariate analysis[m4] identified significant association of age, body mass index (BMI), VFA, systolic blood pressure, low-density lipoprotein (LDL)-cholesterol, triglycer- ide[m6] s (TG), fasting plasma glucose, hemoglobin A1c (HbA1c), fatty liver, current smoking habits, current drinking habits and ≥ 10 kg weight gain compared with the body weight at the age of twenty with prevalence of CRN. Logistic regression analysis revealed that age and current smoking were independent factors associated with CRN prevalence. In addition, BMI, LDL-cholesterol and TG in men and MetS in women were also independent factors associated with CRN prevalence (Table) CONCLUSION: The present study demonstrated that metabolic factors in addition to aging and current smoking might be risk factors for CRN in Japan. Multivariate analysis of risk factors for colorectal neoplasia AGA Abstracts