Background: Pancreatic cancer has an extremely poor prognosis. At the time of diagnosis, over 80% of patients have locally advanced or invasive disease. The carcinogenesis of pancreatic ductal adenocarcinoma (PDAC) involves stepwise progression from pancreatic intraepithelial neoplasia (PanIN) to invasive carcinoma. PanINs measure less than 5 mm and are not detectable on current imaging modalities. Precursor lesions, such as high-grade PanIN (PanIN-3) represent a target for early intervention. Low grade PanIN-1 and -2 are widely prevalent but their significance is unclear. Aims: To evaluate prevalence, significance and implications of PanIN-3 in PDAC. Methods: A retrospective review of the pathology database (1/2000 to 7/2013) at a tertiary care center. Demographic information, patient history, imaging studies, EUS findings and surgical pathology were reviewed. Results: A total of 607 pancreatic resections for cystic and solid lesions were reviewed. Neoadjuvant chemotherapy was given to 48 patients. In the remaining 559 patients, 198 (35.4%) had primary surgical resection for PDAC. Among these patients, 74.5% (of 185) showed the presence of any subtype of PanIN. High grade or PanIN-3 was present in 36.2% (of 185) resections for PDAC. Further, PanIN-3 was mostly found in PDAC (90.5%, p-value <0.001) compared to all other lesion types. Among patients with PDACs, who did not receive neoadjuvant therapy, univariate analysis (tables 1 and 2), demonstrated that Caucasian race and the presence of perineural invasion were significantly associated with PanIN-3. Pathological evidence of chronic pancreatitis and higher number of positive lymph nodes trended towards significant. Using these variables in a binomial logistic regression model demonstrated that Caucasian race (OR 10.9, p-value 0.02, 95% confidence interval 1.4, 86.6) and the presence of perineural invasion (OR 6.4, p-value 0.02, 95% confidence interval 1.4, 29.2) were significantly associated with PanIN-3. Interestingly, a comparison of PanIN3 among PDACs who underwent primary surgical resection versus post-neoadjuvant chemotherapy resection revealed that there was an absence of significant reduction in PanIN-3 among post-neoadjuvant resected PDACs (66 of 168 (39.3%) vs. 12 of 36 (33.3%), p = 0.57). Conclusion: High grade PanIN lesions are particularly prevalent in PDACs. In patients undergoing primary surgical resection for PDACs, high grade PanIN-3 lesions were notably associated with Caucasian race and perineural invasion. Supporting prior literature that PanIN-3 lesions may harbor cells with invasive features, the finding of PanIN-3 association with perineural invasion needs to be validated by larger studies. Table 1. Univariate analysis of clinicopathologic categorical variables of PanIN-1 and -2 vs. PanIN-3