Background: Recent years have brought important developments in preoperative imaging and use of laparoscopic staging of patients with pancreatic adenocarcinoma (PAC). There are few data about the optimal combination of preoperative studies to accurately identify resectable patients. Study Design: We conducted a statewide review of all patients with surgically managed PAC from 1996 to 2003 using data from the Oregon State Cancer Registry, augmented with clinical information from primary medical record review. We documented the use of all staging modalities, including CT, endoscopic ultrasonography, and laparoscopy. Primary outcomes included resection with curative intent. The association between staging modalities, clinical features, and resection was measured using a multivariate logistic regression model. Results: There were 298 patients from 24 hospitals who met the eligibility criteria. Patients were staged using a combination of CT (98%), laparoscopy (29%), and endoscopic ultrasonography (32%). The overall proportion of patients who went to surgical exploration and were resected was 87%. Of patients undergoing diagnostic laparoscopy, metastatic disease that precluded resection was discovered in 24 (27.6%). For patients who underwent diagnostic laparoscopy and were not resected, vascular invasion was the most common determinant of unresectability (56.6%). In multivariate analysis, preoperative weight loss and surgeon decision to use laparoscopy predicted unresectability at laparotomy. Conclusions: This population-based study demonstrates that surgeons appear to use laparoscopy in a subset of patients at high risk for metastatic disease. The combination of current staging techniques is associated with a high proportion of resectability for patients taken to surgical exploration. With current imaging modalities, selective application of laparoscopy with a dual-phase CT scan as the cornerstone of staging is a sound clinical approach to evaluate pancreatic cancer patients for potential resectability.
ASJC Scopus subject areas