NESS Main Site  |  Past & Future Meetings
New England Surgical Society

Back to 2022 Abstracts


Identification of High-Risk Features in Mucinous Cystic Neoplasms of the Pancreas
Maximiliano Servin-Rojas, Zhi Ven Fong, Carlos Fernandez-del Castillo, Cristina R. Ferrone, Jorge A. Roldan, Dario Rocha-Castellanos, Piotr Zelga, Keith D. Lillemoe, Motaz Qadan
Department of Surgery, Massachusetts General Hospital, Boston, Massachusetts, United States

Objective: Mucinous cystic neoplasms of the pancreas (MCN) are uncommon tumors seen almost exclusively in middle-aged women that have the potential of becoming invasive cancer. The objective of the study was to describe our institutional experience with MCN and to identify preoperative demographic, clinical, or radiologic factors that can predict the presence of high-grade dysplasia or invasive carcinoma (HGD/IC) in MCN. Design: Retrospective cohort study. Setting: A large academic tertiary referral hospital in the United States. Patients or Other Participants: Patients aged 18 years or older who underwent pancreatic resection for pancreatic cysts at the Massachusetts General Hospital from 1990 to 2020. We identified patients who fulfilled pathologic criteria for MCN. There were a total of 157 patients identified. Interventions: Not applicable. Main outcome measures: The main outcome of interest was the presence of HGD/IC in the final pathologic analysis. We collected demographic, clinical, radiologic, and pathologic variables for each patient. A bivariate analysis was performed to select variables associated with HGD/IC. Multivariable logistic regression was used to identify predictors of HGD/IC in MCN. Results: The rate of HGD/IC was 11%. In both the benign and HGD/IC groups, most patients were female (91% vs. 88%; p = 0.7) and tumors were more likely to be in the body or tail of the pancreas (84% vs. 95%; p = 0.6). In the bivariate analysis, tumor size ≥ 4 cm (p < 0.001), the presence of an enhancing mural nodule (p= 0.04), and a serum CA 19-9 level ≥ 37 U/mL (p = < 0.001) were associated with HGD/IC. In the multivariable analysis, tumor size > 4 cm (OR 10.9, 95% confidence interval 1.82-208, p = 0.029) and a CA 19-9 level ≥ 37 U/mL (OR 1.33, 95% confidence interval 1.23-22.8, p = 0.026) were significant predictors of HGD/IC. There were no cysts with HGD/IC in patients with tumor size < 4 cm in the absence of an elevated CA 19-9 or mural nodularity. Conclusions: The rate of HGD/IC in resected MCNs was 11%. Tumors with a size above 4 cm and/or a high CA 19-9 level should be considered for prompt surgical resection. Conversely, tumors below 4 cm with no other high-risk features have a negligible risk for HGD/IC and may benefit from non-operative surveillance. Mural nodularity is an additional suspicious feature. These findings may contribute to future guidelines.


Table 1. Demographic, clinical, and radiologic characteristics of patients with a confirmed diagnosis of MCN stratified by pathologic status


 Benign, N = 140HGD/IC, N = 17p-value*
Age at diagnosis49 (36, 59)53 (46, 61)0.2
Sex  0.7
Female127 (91%)15 (88%) 
Smoking Status  0.8
Positive30 (23%)4 (27%) 
Family History of Pancreatic Cancer  0.3
Yes7 (5%)2 (12%) 
Abdominal pain  0.081
Yes64 (46%)4 (24%) 
Jaundice  0.11
Yes0 (0%)1 (6%) 
Acute Pancreatitis  0.7
Yes22 (16%)3 (18%) 
Size in CT (cms)3.2 (2.3, 4.9)5.8 (4.5, 8.9)<0.001
Tumor size  <0.001
Over 4 cm51 (36%)16 (94%) 
Enhancing Mural Nodule  0.042
Yes20 (14%)6 (35%) 
Dilated Main Pancreatic Duct  0.3
Yes2 (1%)1 (6%) 
Serum CA 19-9 (U/mL)11 (6,26)496 (19, 3823)0.002
Serum CA 19-9 (High/Low)  <0.001
High (≥ 37 U/mL)13 (12%)7 (58%) 
Tumor Location  0.6
Tail71 (51%)12 (71%) 
Body46 (33%)4 (24%) 
Head19 (14%)1 (5%) 
Neck4 (2%)0 (0%) 
Median (IQR: interquartile range); n (%); * Mann Whitney U Test; Fisher

Table 2. Multivariable logistic regression of predictors of HGD/IC among resected patients with MCN


 OR95% CIp-value
Tumor size   
Over 4 cm10.91.82, 2080.029
Serum CA 19-9   
High5.151.23, 22.80.026
Enhancing Mural Nodule   
Yes1.330.25, 6.140.7

OR: Odds Ratio; 95% CI: 95% Confidence Intervals.
Back to 2022 Abstracts