Back to 2014 Annual Meeting Abstracts
Preoperative Identification of Main Duct IPMN: A Clinicopathologic Correlate
*Lindsay Hollander1,2,3, *Kisha Mitchell1, *Harry Aslanian1, *James Farrell1, *Joanna Gibson1, *Suzan Abu-Abed1, Charles Cha1, Ronald Salem1
1Yale University School of Medicine, New Haven, CT;2University of Connecticut School of Medicine, Farmington, CT;3VA Connecticut Healthcare System, West Haven, CT
Objective:
Intraductal papillary mucinous neoplasms (IPMNs) are increasingly diagnosed with the current widespread use of cross-sectional imaging. Main duct IPMN (MD-IPMN) carries a high incidence of malignancy and resection is indicated. The accurate diagnosis of this entity is vitally important. Preoperative imaging with EUS, MRI and CT is often used to make the diagnosis but as yet correlation between preoperative imaging and final pathology has not been clearly evaluated. This study seeks to evaluate this correlation.
Design:
77 patients underwent resection for IPMN between 2003 and 2013. Seventy-three underwent preoperative EUS; 65 had CT performed; and 39 had MRI. Specimens were reevaluated by a single gastrointestinal pathologist blinded to the preoperative diagnosis. Receiver operating characteristic (ROC) analysis was performed to assess the accuracy of preoperative identification (mainly by EUS) of MD-IPMN using the pathologic reevaluations of the surgical specimens as a reference standard.
Results:
Of the 77 lesions, 39 were branch duct IPMN (BD-IPMN) and 38 had MD involvement, either as MD-IPMN or mixed type IPMN. Accuracy in predicting MD-IPMN was 73%, positive predictive value 75%, negative predictive value 71%. A dilated main pancreatic duct (>5mm) was indicative of MD involvement (p=0.02). No other measurable parameter correlated with the diagnosis of MD-IPMN. The weighted κ values indicated good agreement between the initial pathologic diagnoses and subsequent reevaluation (κ=0.67).
Conclusions:
The preoperative diagnosis of MD- IPMN is not perfect and the consequences of an incorrect diagnosis are substantial. Imaging criteria need to be refined to optimize the diagnosis of this condition.
Back to 2014 Annual Meeting Abstracts
|
|
