Lower cyst fluid carcinoembryonic antigen cutoff is helpful in the differential diagnosis of mucinous versus non-mucinous pancreatic cysts

Ravi Kumar Sharma,Nikhil Bush,Surinder Singh Rana,Radhika Srinivasan,Ritambhra Nada,Rajesh Gupta,Satyavati Rana,Tarundeep Singh
DOI: https://doi.org/10.1007/s12664-022-01269-w
Abstract:Background and aim: Pancreatic cystic lesions (PCLs) are being diagnosed with increased frequency and have varying neoplastic potential. We conducted this multimodal, prospective study to evaluate the role of tumor cytology and molecular markers to differentiate PCL subtypes. Methods: Consecutive undiagnosed patients with PCLs (n = 100, mean age: 50.37 years; 41% males) were prospectively studied. Cyst fluid carcinoembryonic antigen (CEA), CA19.9, CA125, CA72.4, and vascular endothelial growth factor-alpha (VEGF-α) levels were measured by quantitative enzyme-linked immunosorbent assay (ELISA) method. Mutational analysis of the KRAS gene (exon 2, Codon 12 and 13) and GNAS gene (Exon 8, Codon 201) were performed by Sanger's sequencing. Results: The mean cyst size was 4.32 ± 2.4 cm. Fluid cytology revealed definitive diagnosis in 21 (22.3%) patients. All malignant PCLs could be identified on cytology whereas 10/14 (71%) non-malignant mucinous PCLs could also be identified on cytology based on mucin staining. Among the tested tumor markers, cyst fluid CEA had the best diagnostic performance for differentiation between mucinous and non-mucinous PCLs (AUC 0.933 [95% CI 0.86-0.91]). At a cyst fluid CEA cutoff level of 45.0 ng/mL, the sensitivity, specificity, positive predictive value, and negative predictive value for differentiation between mucinous and non-mucinous cysts were 88.5%, 96.8%, 92.0%, and 95.3%, respectively (p < 0.05). KRAS and GNAS mutation had no significant diagnostic benefit in comparison to fluid cytology and CEA levels. Conclusions: Fluid CEA at a lower cutoff of 45 ng/mL is the most accurate marker to differentiate between mucinous and non-mucinous PCL. The KRAS and GNAS mutational analysis does not improve upon the diagnostic performance of fluid cytology and tumor markers.
What problem does this paper attempt to address?