GLUCOSE LEVELS IN EUS-ASPIRATED CYST FLUID HAVE A HIGH ACCURACY FOR THE DIAGNOSIS OF MUCINOUS PANCREATIC CYSTIC LESIONS

Background and aim: The differential diagnosis between mucinous (M) and non mucinous (NM) pancreatic cystic lesions (PCLs) is of paramount importance to drive the appropriate patients’ management. Several intracystic biomarkers have been investigated, and Carcinoembryonic antigen (CEA) levels >192 ng/ml are considered suggestive of mucinous cysts (MC), yet with limited accuracy. Recently, some studies have suggested that low intracystic glucose levels are associated with MC, but these data are limited to small series or to operated patients. The aim of the study was to assess the diagnostic accuracy of intracystic glucose as compared to CEA levels in EUS obtained cyst fluid to differentiate MPCLs vs NMPCLs. Material and methods: We prospectively enrolled patients with PCLs 20 mm and indication to analyze cyst fluid who underwent EUS + FNA with a 22G needle for PCLs from July 2018 to October 2019. Glucose and CEA levels were examined in the hospital laboratory with same methods as for circulating levels. A cytological analysis of all the sample was conducted in all the cases, with rapid on site evaluation (ROSE) by expert cytopathologists. The diagnostic accuracy of glucose and CEA was evaluated considering as “gold standard” the surgical pathology report in case of subsequent surgery, the cytological report of FNA and/or imaging features (MRI and EUS) in other cases as agreed by two expert pancreatologists blind to PCLs fluid analysis. Results: 48 patients with PCLs were enrolled (mean age 60 years, range 26-83; 25 fe males-52%). Mean size of cystic lesions was 46 mm (range20-140 mm). The final diagnosis, based on surgical pathology in 6 cases, on EUS cytology in 15 cases and on endoscopic and imaging data in the remaining 27 cases, was of MPCLs in 24 cases (50%) (all BD- or mixed type-IPMNs), of serous cystic adenomas in 16 (33,3%), pseudocyst in 2 (4,2%), cystic neuroendocrine tumors in 2 (4.2%) (Ki 67: 2% and 3% respectively), of mesenterial cyst, simple cyst in chronic pancreatitis and Schwan noma in 1 case each. The mean intracystic glucose level was 13 mg/dl (95% 2.2-24) in MPCLs and 93.6 mg/dl (95% CI 84.2-102.9) in NMPCLs (p<0.0001). At a ROC curve analysis the best cut-off of glucose to distinguish M and NM PCLs was 30 mg/dl. An intracystic glucose value 30 mg/dL showed an area under the curve (AUC) of 0.95 (CI 0.85-0.99), with a sensitivity of 91.3% and specificity 100%. AUC for CEA 192 ng/ml was 0.69 (CI 0.54-0.81), with a sensitivity of 37.5% and specificity of 100%. There was no correlation between glucose levels and cyst size or CEA levels. Conclusions: Intracystic glucose dosage of cyst fluid obtained during EUS with FNA represents a valid and simple tool for the differential diagnosis of M vs NM PCLs being more accurate than CEA levels. The mechanisms leading to different glucose levels in different cyst types remain unexplored.