J Gastrointest Surg (2014) 18:447–456 DOI 10.1007/s11605-013-2444-6
2013 SSAT PLENARY PRESENTATION
Does Preoperative Cross-Sectional Imaging Accurately Predict Main Duct Involvement in Intraductal Papillary Mucinous Neoplasm? M. R. Barron & A. M. Roch & J. A. Waters & J. A. Parikh & J. M. DeWitt & M. A. Al-Haddad & E. P. Ceppa & M. G. House & N. J. Zyromski & A. Nakeeb & H. A. Pitt & C. Max Schmidt
Received: 2 July 2013 / Accepted: 16 December 2013 / Published online: 9 January 2014 # 2014 The Society for Surgery of the Alimentary Tract
Abstract Main pancreatic duct (MPD) involvement is a well-demonstrated risk factor for malignancy in intraductal papillary mucinous neoplasm (IPMN). Preoperative radiographic determination of IPMN type is heavily relied upon in oncologic risk stratification. We hypothesized that radiographic assessment of MPD involvement in IPMN is an accurate predictor of pathological MPD involvement. Data regarding all patients undergoing resection for IPMN at a single academic institution between 1992 and 2012 were gathered prospectively. Retrospective analysis of imaging and pathologic data was undertaken. Preoperative classification of IPMN type was based on cross-sectional imaging (MRI/magnetic resonance cholangiopancreatography (MRCP) and/or CT). Three hundred sixty-two patients underwent resection for IPMN. Of these, 334 had complete data for analysis. Of 164 suspected branch duct (BD) IPMN, 34 (20.7 %) demonstrated MPD involvement on final pathology. Of 170 patients with suspicion of MPD involvement, 50 (29.4 %) demonstrated no MPD involvement. Of 34 patients with suspected BD-IPMN who were found to have MPD involvement on pathology, 10 (29.4 %) had invasive carcinoma. Alternatively, 2/50 (4 %) of the patients with suspected MPD involvement who ultimately had isolated BD-IPMN demonstrated invasive carcinoma. Preoperative radiographic IPMN type did not correlate with final pathology in 25 % of the patients. In addition, risk of invasive carcinoma correlates with pathologic presence of MPD involvement. Keywords Intraductal papillary mucinous neoplasm (IPMN) . Pathology . Cross-sectional imaging studies
This paper was presented on a plenary session at the 2013 Digestive Disease Week/Society for Surgery of the Alimentary Tract meeting in Orlando, FL, on 21 May 2013. M. R. Barron and A. M. Roch contributed equally to this paper. M. R. Barron : A. M. Roch : J. A. Waters : J. A. Parikh : E. P. Ceppa : M. G. House : N. J. Zyromski : A. Nakeeb : H. A. Pitt : C. M. Schmidt Department of Surgery, Indiana University School of Medicine, Indianapolis, IN, USA J. M. DeWitt : M. A. Al-Haddad Division of Gastroenterology/Hepatology, Indiana University Health Hospital, Indianapolis, IN, USA C. M. Schmidt (*) IU Health Pancreatic Cyst and Cancer Early Detection Center, 980 West Walnut Street C522, Indianapolis, IN 46202, USA e-mail: [email protected]
Introduction Intraductal papillary mucinous neoplasm (IPMN) represents 30–40 % of all cystic lesions of the pancreas, 1,2 with increasing identification since its early description by Ohashi in 1982.3 The widespread use of imaging, improved spatial and contrast resolution of cross-sectional imaging techniques, and better awareness of this disease among physicians are all major contributors to the increased appreciation of IPMN today. Given the poor prognosis of pancreatic cancer (5-year survival of 4–6 %), 4,5 considerable attention has focused on its noninvasive precursors, especially IPMN, as they offer a window of opportunity for prevention and cancer early detection. The recent differentiation between “branch duct” (BDIPMN) and “main duct” (MD-IPMN) 6,7 has added further knowledge to the pathophysiologic aspects and natural history of IPMN. A stepwise sequence from low-grade to high-grade dysplasia and finally invasive carcinoma (comparable to the malignant progression of colonic polyps) is now well recognized.7–9 MD-IPMN harbors a malignant transformation
448
risk of about 40–95 %, 10,11 which makes it an indication for resection, as underlined in the current international consensus guidelines.8 Conversely, with an overall risk of malignant transformation estimated between 6 and 40 %, 11–13 close follow-up for BD-IPMN seems reasonable.8,12 In patients with mixed-type IPMN, management is identical to that of MD-IPMN, as main pancreatic duct (MPD) involvement is the leading determinant of malignant transformation. In the 2012 recommendations,8 the management of IPMN is driven mainly by radiologic characteristics and presence of related symptoms to devise an accurate preoperative classification with appropriate treatment. In IPMN, the underlying mucinous epithelial transformation process indirectly creates the radiologic findings seen in IPMN. The abnormal production of high-viscosity mucin by this epithelium causes inefficient drainage within the pancreatic ductal system, which subsequently leads to ductal obstruction and passive ductal dilation within the main duct, side-branch ducts, or both. This pathological mechanism creates the typical and characteristic computed tomography (CT) and magnetic resonance imaging (MRI) features that lead to the diagnosis of IPMN. Two recent studies from Verona 14 and Boston 15 have analyzed the discrepancies between radiological and pathological findings in establishing the correct diagnosis among all pancreatic cystic lesions. Several teams have studied the ability of radiology to stratify the malignant potential of pancreatic mucinous lesions.10,11,16–18 However, to our knowledge, no study has focused on the ability of imaging studies to determine IPMN type. This is of particular concern, as correct characterization of IPMN type ensures accurate stratification into resection and surveillance treatment groups. Appropriate stratification will identify patients at high risk of malignancy, but also safeguard against unnecessary resections with their associated complications and mortality.19 In an attempt to measure the accuracy of radiology to classify the IPMN type, the present study compares the preoperative diagnosis of IPMN type, as evaluated by various cross-sectional imaging modalities (CT and MRI/magnetic resonance cholangiopancreatography (MRCP)), and final pathology analysis as the reference standard in a large cohort of patients.
Patients and methods Patient selection From July 1992 to July 2012, data from all patients who underwent surgical resection of IPMN at Indiana University Hospital were reviewed. A retrospective analysis of a prospectively collected database, supplemented by chart review, was performed for this particular study. Patients were selected
J Gastrointest Surg (2014) 18:447–456
based upon having at least one pancreatic cross-sectional imaging study within 3 months prior to surgery. Data was collected and reported in strict compliance with patient confidentiality guidelines defined by the Indiana University Institutional Review Board. Exclusion criteria Patients were excluded from this study if they had incomplete radiological and pathological data for IPMN classification, imaging studies performed exclusively at another institution, or only remote imaging studies (>3 months before intervention). Parameters assessed Qualitative imaging analysis included assessment of the IPMN type, communication with the MPD, and presence of mural nodules. Quantitative measurements included branch duct size, MPD maximal diameter, and the presence of multifocality. MD-IPMN was radiologically defined as diffusely or focally dilated MPD with a maximal diameter equal to or greater than 5 mm.6–8 BD-IPMN was radiographically defined as a characteristic unifocal/multifocal ductal dilation confined to branch ducts with no MPD dilation (
2013 SSAT PLENARY PRESENTATION
Does Preoperative Cross-Sectional Imaging Accurately Predict Main Duct Involvement in Intraductal Papillary Mucinous Neoplasm? M. R. Barron & A. M. Roch & J. A. Waters & J. A. Parikh & J. M. DeWitt & M. A. Al-Haddad & E. P. Ceppa & M. G. House & N. J. Zyromski & A. Nakeeb & H. A. Pitt & C. Max Schmidt
Received: 2 July 2013 / Accepted: 16 December 2013 / Published online: 9 January 2014 # 2014 The Society for Surgery of the Alimentary Tract
Abstract Main pancreatic duct (MPD) involvement is a well-demonstrated risk factor for malignancy in intraductal papillary mucinous neoplasm (IPMN). Preoperative radiographic determination of IPMN type is heavily relied upon in oncologic risk stratification. We hypothesized that radiographic assessment of MPD involvement in IPMN is an accurate predictor of pathological MPD involvement. Data regarding all patients undergoing resection for IPMN at a single academic institution between 1992 and 2012 were gathered prospectively. Retrospective analysis of imaging and pathologic data was undertaken. Preoperative classification of IPMN type was based on cross-sectional imaging (MRI/magnetic resonance cholangiopancreatography (MRCP) and/or CT). Three hundred sixty-two patients underwent resection for IPMN. Of these, 334 had complete data for analysis. Of 164 suspected branch duct (BD) IPMN, 34 (20.7 %) demonstrated MPD involvement on final pathology. Of 170 patients with suspicion of MPD involvement, 50 (29.4 %) demonstrated no MPD involvement. Of 34 patients with suspected BD-IPMN who were found to have MPD involvement on pathology, 10 (29.4 %) had invasive carcinoma. Alternatively, 2/50 (4 %) of the patients with suspected MPD involvement who ultimately had isolated BD-IPMN demonstrated invasive carcinoma. Preoperative radiographic IPMN type did not correlate with final pathology in 25 % of the patients. In addition, risk of invasive carcinoma correlates with pathologic presence of MPD involvement. Keywords Intraductal papillary mucinous neoplasm (IPMN) . Pathology . Cross-sectional imaging studies
This paper was presented on a plenary session at the 2013 Digestive Disease Week/Society for Surgery of the Alimentary Tract meeting in Orlando, FL, on 21 May 2013. M. R. Barron and A. M. Roch contributed equally to this paper. M. R. Barron : A. M. Roch : J. A. Waters : J. A. Parikh : E. P. Ceppa : M. G. House : N. J. Zyromski : A. Nakeeb : H. A. Pitt : C. M. Schmidt Department of Surgery, Indiana University School of Medicine, Indianapolis, IN, USA J. M. DeWitt : M. A. Al-Haddad Division of Gastroenterology/Hepatology, Indiana University Health Hospital, Indianapolis, IN, USA C. M. Schmidt (*) IU Health Pancreatic Cyst and Cancer Early Detection Center, 980 West Walnut Street C522, Indianapolis, IN 46202, USA e-mail: [email protected]
Introduction Intraductal papillary mucinous neoplasm (IPMN) represents 30–40 % of all cystic lesions of the pancreas, 1,2 with increasing identification since its early description by Ohashi in 1982.3 The widespread use of imaging, improved spatial and contrast resolution of cross-sectional imaging techniques, and better awareness of this disease among physicians are all major contributors to the increased appreciation of IPMN today. Given the poor prognosis of pancreatic cancer (5-year survival of 4–6 %), 4,5 considerable attention has focused on its noninvasive precursors, especially IPMN, as they offer a window of opportunity for prevention and cancer early detection. The recent differentiation between “branch duct” (BDIPMN) and “main duct” (MD-IPMN) 6,7 has added further knowledge to the pathophysiologic aspects and natural history of IPMN. A stepwise sequence from low-grade to high-grade dysplasia and finally invasive carcinoma (comparable to the malignant progression of colonic polyps) is now well recognized.7–9 MD-IPMN harbors a malignant transformation
448
risk of about 40–95 %, 10,11 which makes it an indication for resection, as underlined in the current international consensus guidelines.8 Conversely, with an overall risk of malignant transformation estimated between 6 and 40 %, 11–13 close follow-up for BD-IPMN seems reasonable.8,12 In patients with mixed-type IPMN, management is identical to that of MD-IPMN, as main pancreatic duct (MPD) involvement is the leading determinant of malignant transformation. In the 2012 recommendations,8 the management of IPMN is driven mainly by radiologic characteristics and presence of related symptoms to devise an accurate preoperative classification with appropriate treatment. In IPMN, the underlying mucinous epithelial transformation process indirectly creates the radiologic findings seen in IPMN. The abnormal production of high-viscosity mucin by this epithelium causes inefficient drainage within the pancreatic ductal system, which subsequently leads to ductal obstruction and passive ductal dilation within the main duct, side-branch ducts, or both. This pathological mechanism creates the typical and characteristic computed tomography (CT) and magnetic resonance imaging (MRI) features that lead to the diagnosis of IPMN. Two recent studies from Verona 14 and Boston 15 have analyzed the discrepancies between radiological and pathological findings in establishing the correct diagnosis among all pancreatic cystic lesions. Several teams have studied the ability of radiology to stratify the malignant potential of pancreatic mucinous lesions.10,11,16–18 However, to our knowledge, no study has focused on the ability of imaging studies to determine IPMN type. This is of particular concern, as correct characterization of IPMN type ensures accurate stratification into resection and surveillance treatment groups. Appropriate stratification will identify patients at high risk of malignancy, but also safeguard against unnecessary resections with their associated complications and mortality.19 In an attempt to measure the accuracy of radiology to classify the IPMN type, the present study compares the preoperative diagnosis of IPMN type, as evaluated by various cross-sectional imaging modalities (CT and MRI/magnetic resonance cholangiopancreatography (MRCP)), and final pathology analysis as the reference standard in a large cohort of patients.
Patients and methods Patient selection From July 1992 to July 2012, data from all patients who underwent surgical resection of IPMN at Indiana University Hospital were reviewed. A retrospective analysis of a prospectively collected database, supplemented by chart review, was performed for this particular study. Patients were selected
J Gastrointest Surg (2014) 18:447–456
based upon having at least one pancreatic cross-sectional imaging study within 3 months prior to surgery. Data was collected and reported in strict compliance with patient confidentiality guidelines defined by the Indiana University Institutional Review Board. Exclusion criteria Patients were excluded from this study if they had incomplete radiological and pathological data for IPMN classification, imaging studies performed exclusively at another institution, or only remote imaging studies (>3 months before intervention). Parameters assessed Qualitative imaging analysis included assessment of the IPMN type, communication with the MPD, and presence of mural nodules. Quantitative measurements included branch duct size, MPD maximal diameter, and the presence of multifocality. MD-IPMN was radiologically defined as diffusely or focally dilated MPD with a maximal diameter equal to or greater than 5 mm.6–8 BD-IPMN was radiographically defined as a characteristic unifocal/multifocal ductal dilation confined to branch ducts with no MPD dilation (
