Background: In a group of 74 patients where computed tomography (CT) showed locally extending but not metastatic pancreatic cancer, we have demonstrated (Surg Endosc 2005;19:638-42) that 1/3 will have occult metastatic disease using diagnostic laparoscopy and peritoneal lavage for cytology (DL-PLC). Using a larger cohort, we asked if there are predictive factors that increase the likelihood that occult metastatic disease would be discovered by DL-PLC.
Methods: Between 2000 and 2008, 202 consecutive patients underwent DL-PLC when thin-cut, contrast-enhanced pancreas protocol CT showed locally-advanced pancreatic cancer without evidence of metastatic disease. Univariate and logistic regression analyses were used to identify factors that predicted which patients would have occult metastatic disease discovered by DL-PLC.
Results: DL-PLC upstaged 55 of 202 patients (27.2%) to Stage IV. Of these, 40 (19.2%) patients had positive peritoneal lavage cytology; 25 (12.4%) patients had hepatic metastases, and only 4 (2%) had metastatic peritoneal deposits. Univariate analysis demonstrated that increasing tumor size by CT and CA 19-9 > 500 U/mL predicted positive DL-PLC. Not predictive were location of tumor (head vs non-head of pancreas), involvement of multiple mesenteric vessels, weight loss, elevated bilirubin, back pain, or low serum albumin. Multivariate logistic regression analysis demonstrated that increasing tumor size was the only predictive factor of positive DL-PLC. In our cohort, when tumor size was greater than 4.5 cm, DL-PLC was positive in 45% of patients.
Conclusion: Despite advancements in the sensitivity of high-resolution CT, the use of diagnostic laparoscopy and peritoneal lavage cytology in selected patients provides for additional accuracy by upstaging 27% of patients to Stage IV. Compared with previous staging studies, modern pancreas protocol CT may have reduced the incidence of unanticipated peritoneal deposits to only 2% of patients. Only tumor size was able to predict the results of DL-PLC justifying its use in patients with locally-advanced pancreatic cancer that have no evidence of metastatic disease by CT.
Session: Podium Presentation
Program Number: S101