One Centerʼs Experience of Pancreatic Cyst Surveillance with EUS

Abstract

Purpose: Pancreatic cysts are increasing in incidence given advances in imaging techniques. Their evaluation and management poses a number of challenges. Surveillance is frequently undertaken in smaller, asymptomatic cysts or in patients deemed to be poor surgical candidates. Frequency and modality for surveillance remain uncertain. The purpose of this project was to retrospectively analyze the utility of pancreatic cyst surveillance with EUS/FNA. Methods: This was a secondary analysis from a previously created database by our group. After IRB approval, we requested the medical record numbers of all patients coded with ICD-9 code of 577.2 (pancreatic cyst) from October, 2003 through December, 2011. We included all pancreatic cysts that were evaluated for a K-Ras mutation and/or had a final pathologic diagnosis. For the purpose of the current analysis, we selected the patients who underwent more than one EUS/FNA for evaluation of the pancreatic cyst.TableResults: A total of 37 patients were identified who had at least one follow-up EUS/FNA after their initial endoscopic evaluation. K-Ras: Five patients had a repeat K-Ras analysis over an average of 230 days. In each case, the repeat K-Ras testing was unchanged from the original (two were negative, three were positive). CEA: Eleven patients had a repeat cyst fluid CEA measurement over an average of 575 days. A discordance in CEA levels on follow-up, defined as one value below 192 ng/mL (considered to be predictive of a mucinous lesion) and the other value above that level, were found in two of eleven cases. In both instances, the CEA level decreased, over 723 days and 176 days (average 450 days). Average follow-up time in non-discordant CEA cases was 614 days. Cytology: 24 patients underwent at least one follow-up cyst cytology evaluation. Several of the patients had more than one surveillance cytology analysis, for a total of 35 cytology surveillance cases. The average follow-up time between two consecutive specimens was 539 days. Cytologic change occurred in 14 out of 35 cases over an average of 683 days, as compared to 443 days in cases where cytology did not change. In 11 of the 14 cases where cytology changed, the transformation was from a non-diagnostic evaluation to a particular cytologic diagnosis. The other three cases demonstrated a more advanced cytologic diagnosis. Conclusion: Given the relative lack of information gained from repeat CEA and K-RAS analysis in the setting of adequate cellularity for cytologic analysis, it may be reasonable to space out the interval of invasive pancreatic cyst surveillance to a longer period of time or forego it altogether in favor of non-invasive surveillance.