A Novel Scoring System to Improve the Detection Efficiency of Pancreatic Cystic Lesions in the General Population.

ABSTRACT

Objective Pancreatic cystic lesions (PCLs) are known risk factors for pancreatic cancer. Therefore, this study explored the predictors identifying PCLs in a general population and developed a scoring system to help more efficiently diagnose these entities during medical checkups. Methods We reviewed 9,369 examinees of abdominal ultrasound (AUS) during medical checkups between January 2013 and November 2019. Predictors of PCLs were identified using a multivariate logistic regression analysis, and we constructed a scoring system based on these predictors. Results PCLs were detected in 118 (1.3%). Age 50-59 years old [odds ratio (OR) 2.52, 95% confidence interval (CI) 1.18-5.35], 60-69 years old (OR 3.91, 95% CI 1.86-8.26), and ≥70 years old (OR 10.5, 95% CI 5.03-21.7) as well as abdominal pain (OR 1.85, 95% CI 1.14-3.00), alcohol consumption (OR 1.72, 95% CI 1.03-2.89), a family history of pancreatic cancer (OR 2.41, 95% CI 1.09-5.34), and pre-diabetes or diabetes (OR 1.78, 95% CI 1.05-3.00) were predictors of PCLs. The following scores were assigned according to regression coefficients age (50-59 years old, 1 point; 60-69 years old, 1.5 points; ≥70 years old, 2.5 points); abdominal pain, 1 point, alcohol consumption, 1 point; a family history of pancreatic cancer, 1 point; and pre-diabetes, 1 point. The PCL detection rate increased with the total score 0.2% for total score 0 point, 5.4% for ≥4.0 points. The area under the curve of the scoring system was 0.75 (95% CI 0.70-0.79). Conclusion Our scoring system allows the risk of PCLs to be determined and may help more efficiently diagnose these entities.