P218

ASSOCIATION FOR ACADEMIC SURGERY AND SOCIETY OF UNIVERSITY SURGEONS—ABSTRACTS 80 cases identified, 64 were found eligible for review, 14 did not have surgical resection and 2 had neoadjuvant therapy. Results: Adequate nodal resection was achieved in 69% of patients. The median number of lymph nodes sampled was 17 (mean ⫽ 20). Factors associated with sub-optimal lymph node retrieval was laparoscopic resection and gross pathology performed by the attending rather than the physicians assistant. Conclusions: Lymph node status is a strong predictor of outcome in colorectal cancer with regard to overall survival and recurrence rate. In order to recommend appropriate treatment, adequate nodal sampling is required to accurately stage colorectal cancer patients. Although our rate of sufficient nodal sampling is better than that reported for the national population, this still leaves a large proportion of our patients potentially not being staged properly. The number of nodes retrieved during colon resection for cancer needs to be examined by individual institutions. P216. WITHDRAWN P217. COMPLIANCE WITH SURGICAL TREATMENT GUIDELINES FOR MELANOMA IN A COMMUNITY TEACHING HOSPITAL. J. Erickson 1, J. M. Velasco 2, T. J. Hieken 2; 1Rush Medical College, Chicago, IL, 2Rush North Shore Medical Center, Skokie, IL Objective: Wide variation in the surgical treatment of melanoma occurs, despite efforts to standardize care with the publication of clinical practice guidelines. Failure to apply these standards may lead to inaccurate staging, a poor outcome, increased morbidity or be cost-ineffective. The purpose of our study was to evaluate the treatment of melanoma patients in our institution and identify factors associated with variance from the National Comprehensive Cancer network (NCCN) recommendations for margins of excision and application of sentinel lymph node biopsy with selective lymph node dissection. Methods: We retrospectively accessed Cancer Registry data on 245 clinical stage I and II melanoma patients who formed the basis of this study. Treatment data was confirmed by individual review of all pathology reports and operative notes. Statistical analysis was performed using an SAS statistical software package. Results: Based on tumor stage, margins of excision conformed to NCCN guidelines in 85% of T0, 89% of T1 and 68% of T2,3 and 4 tumors. Over-treatment was more common than under-treatment for T0 and T1 melanomas (3% vs. 9%) whereas under-treatment was more common than over-treatment for thicker melanomas (T2, 3 and 4, 26% vs. 4%). Lymph node staging was performed in 10% of T1a patients, 71% of T1b patients, 68% of T2 and T3 patients and 46% of T4 patients. Completion lymph node dissection was performed after a positive sentinel lymph node biopsy in 83% of cases. Treatment by a surgical oncologist and age ⬍ 80 years was associated with greater adherence to NCCN guidelines (p⫽0.01). Conclusions: Despite widespread dissemination of oncology treatment algorithms for the management of melanoma patients, we found that treatment conformed to NCCN guidelines in only approximately two-thirds of our cases. Further investigation is needed to determine how best to promote more uniform, high-quality care, to determine the effect of deviation from such standards on patient outcome and to assess the associated costs of under- or over-treatment of melanoma patients. P218. TREATMENT STRATEGY FOR BRANCH DUCT TYPE INTRADUCTAL PAPILLARY MUCINOUS NEOPLASMS OF THE PANCREAS. K. Nagai, A. Kida, K. Kami, Y. Kawaguchi, S. Uemoto, R. Doi; Kyoto University, Kyoto, Japan Introduction: Surgical resection is recommended for main duct type of intraductal papillary mucinous neoplasms (IPMNs) of the pancreas because the majority of this type of tumors is malignant; however, the indication of surgical resection for the branch duct type of IPMNs still remains controversy. The aim of this study was to 319 describe clinico-pathological characteristics of surgically resected branch duct type of IPMNs and to extract factors that correlate with malignancy. Methods: Fifty-seven patients with branch duct IPMNs consecutively resected between January 1984 to June 2006 in our hospital were reviewed. They were classified into benign tumor group (adenoma and borderline) and malignant tumor group (carcinoma-in-situ [CIS] and invasive carcinoma). We compared the clinical characteristics and the survival time of the patients with invasive IPMN, non-invasive IPMNs (adenoma, borderline and CIS) and stage-matched patients with pancreatic ductal carcinoma at the same era. Results: There were 32 males (56.1%) and 25 females (43.9%) with a mean age at diagnosis of 63 years (range, 41-85 years). Eighteen patients (31.6%) had adenoma, 10 (17.5%) borderline, 9 (15.8%) CIS and 20 (35.1%) invasive carcinoma. Thirty-one patients (54.4%) were symptomatic. Most common symptoms were onset or worsening of diabetes (26.3%), abdominal pain (22.8%) and loss of body weight (19.3%). Four clinical factors were significantly correlated with malignancy by univariate analysis, i.e., the presence of abdominal pain (p⫽0.033), loss of body weight (p⫽0.003), size of cystic lesions (p⬍0.0001) and the presence of mural nodules (p⫽0.012). Moreover, size of cystic lesion was an independent predictive factor for malignancy by multivariate analysis (95% C.I. 0.87-0.99). However, it is of note that 4 cases with cystic lesions in diameter less than 30 mm and no mural nodules were malignant. Moreover, 2 cases of them were asymptomatic. No patients with non-invasive IPMN died of their disease following surgical resection. The survival time of invasive IPMNs were significantly shorter than that of non-invasive IPMNs (p⫽0.0003). The overall 5-year survival rate for patients with invasive IPMNs was 58.4%, which is much higher than pancreatic ductal carcinoma (p⫽0.02). Conclusions: While non-invasive IPMNs showed excellent survival rate following resection, the survival rate of invasive IPMNs were low. Precise preoperative determination of malignancy was difficult in branch duct type of IPMNs. Therefore, they should be managed with aggressive policy about resection before progression to invasive IPMNs. P219. TRENDS IN MAMMOGRAPHY AND CLINICAL BREAST EXAMINATION: RESULTS OF A NATIONAL POPULATION-BASED SURVEY. A. B. Chagpar, C. R. Scoggins, R. C. Martin, II, K. M. McMasters; University of Louisville, Louisville, KY Introduction: Breast cancer remains the leading malignancy affecting women in the United States. There are well-accepted clinical guidelines for early detection of breast cancer through mammography and clinical breast examination. The purpose of this study was to determine trends in the utilization of these techniques over the past 5 years. Methods: The National Health Interview Survey is a population-based computer-assisted personal household interview survey of non-institutionalized civilian Americans conducted annually by the National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention. Women over 40 years of age who completed the cancer control module of this survey in 2000 and 2005 formed the cohort of interest for this study. Data were evaluated using SAS and SAS-callable SUDAAN software. Results: 10,994 and 11,128 women over age 40 were surveyed in 2000 and 2005, respectively. Although there was a slight decrease in the proportion of women reporting that they had ever had a mammogram (79.6% vs. 78.8%, p⬍0.001), there was no change in the fraction of women who reported having a mammogram in the year preceding the survey between 2000 and 2005 (59.7% vs. 59.8%). The reasons reported for not having a mammogram changed somewhat between 2000 and 2005. While the main reason for not getting a mammogram remained that people “didn’t think about it” (41.3% vs. 39.0%), the second most common reason reported in both years was that their doctor didn’t order it (12.1% vs. 13.7%). In addition, there was a decrease in the proportion of individuals reporting ever having had a clinical breast exam (80.4% vs. 74.1%, p⬍0.001) and in the propor-