Abstract
<title>Abstract</title> <p> Background Differentiation of pancreatic cysts is essential for management. Intracystic glucose analysis is currently the most sensitive and available biomarker for this purpose; however, sending aspirate to the laboratory for glucometry consumes volume that could be used for additional biomarkers. Glucose on-site evaluation (GOSE) could eliminate this limitation. Aims To evaluate the analytical interchangeability between GOSE and laboratory glucometry in pancreatic cyst fluid aspirated by EUS-FNA/B. Methods Pancreatic cysts that underwent EUS-FNA/B for diagnostic evaluation were retrospectively enrolled. A drop of aspirate was used for GOSE; the remainder was used for laboratory glucometry. Cysts were classified as mucinous (definitive, probable, or compatible) or non-mucinous. Interchangeability was assessed using the Bland-Altman analysis, intraclass correlation coefficient (ICC), and linear regression. Diagnostic performance was compared using ROC curve analysis with DeLong test. Results 117 patients were included. Median glucose values were 44 mg/dl for laboratory glucometry and 43 mg/dl for GOSE; 58 pancreatic cysts were categorized as mucinous, and 59 as non-mucinous. Pearson’s R <sup>2</sup> was 0.95 (p < 0.001). The Bland-Altman analysis demonstrated a mean bias of + 1.3 mg/dl (95%CI -0.5 to + 3.1), with LoA of -18.2 to + 20.7 mg/dl and 98.3% of pairs within the LoA. Agreement was consistent across mucinous and non-mucinous subtypes. ICC was 0.98 (95%CI 0.97–0.98). Conclusion GOSE demonstrates excellent analytical interchangeability with laboratory glucometry (ICC 0.98, bias + 1.3 mg/dl) and equivalent diagnostic performance at the standard 50 mg/dl threshold. These findings support using GOSE as the sole measurement method while preserving cyst fluid for additional biomarker analysis. </p>