Role of EUS-FNA-based cytology in the diagnosis of mucinous pancreatic cystic lesions: a systematic review and meta-analysis

Abstract

Background: Preoperative diagnosis of malignancy in pancreatic cystic lesions (PCLs) remains challenging. Most non-mucinous cystic lesions (NMCLs) are benign, but mucinous cystic lesions (MCLs) are more likely to be premalignant or malignant.

Aim: The aim of this study was to assess the sensitivity, specificity, and positive and negative likelihood ratios (LRs) of EUS-FNA-based cytology in differentiating MCLs from non-mucinous PCLs.

Methods: We conducted a comprehensive search of MEDLINE, SCOPUS, Cochrane, and "CINAHL Plus" databases to identify studies, in which the results of EUS-FNA-based cytology of PCLs were compared with those of surgical biopsy or surgical excision histopathology. A DerSimonian-Laird random effect model was used to estimate the pooled sensitivity, specificity, and LRs, and a summary receiver-operating characteristic (SROC) curve was constructed.

Results: We included 376 patients from 11 distinct studies who underwent EUS-FNA-based cytology and also had histopathological diagnosis. The pooled sensitivity and specificity in diagnosing MCLs were 0.63 (95% CI, 0.56-0.70) and 0.88 (95% CI, 0.83-0.93), respectively. The positive and negative LRs in diagnosing MCLs were 4.46 (95% CI, 1.21-16.43) and 0.46 (95% CI, 0.25-0.86), respectively. The area under the curve (AUC) was 0.89.

Conclusions: EUS-FNA-based cytology has overall low sensitivity but good specificity in differentiating MCLs from NMCLs. Further research is required to improve the overall sensitivity of EUS-FNA-based cytology to diagnose MCLs while evaluating PCL.

Figure 1

Flow diagram of study selection process for the Systematic Review and Meta-analysis

Figure 2

Pooled sensitivity and specificity. The size of each round is proportional to the sample size for each study, and the horizontal lines through the rounds indicate a graphical representation of the 95% CI of that study. For the combined analysis, the diamond and vertical dashed bar indicates the pooled sensitivity or specificity, with the left and right ends of the vertical bar indicating the pooled 95% CI.

Figure 3

Pooled positive and negative likelihood ratios. The size of each round is proportional to the sample size for each study, and the horizontal lines through the rounds indicate a graphical representation of the 95% CI of that study. For the combined analysis, the diamond and vertical dashed bar indicates the pooled positive or negative likelihood ratio, with the left and right ends of the vertical bar indicating the pooled 95% CI.

Figure 4

Pooled sensitivity and specificity for all prospective studies. The size of each round is proportional to the sample size for each study, and the horizontal lines through the rounds indicate a graphical representation of the 95% CI of that study. For the combined analysis, the diamond and vertical dashed bar indicates the pooled sensitivity or specificity, with the left and right ends of the vertical bar indicating the pooled 95% CI.

Figure 5

Pooled positive and negative likelihood ratios for all prospective studies. The size of each round is proportional to the sample size for each study, and the horizontal lines through the rounds indicate a graphical representation of the 95% CI of that study. For the combined analysis, the diamond and vertical dashed bar indicates the pooled positive or negative likelihood ratio, with the left and right ends of the vertical bar indicating the pooled 95% CI.

Figure 6

Summary Receiver Operating Characteristic (SROC) Curve for all 11 studies of meta-analysis and SROC Curve for 4 prospective studies