The American Journal of Surgery (2015) 209, 478-482
Midwest Surgical Association
Should all branch-duct intraductal papillary mucinous neoplasms be resected? Jennifer K. Plichta, M.D., M.S.a, Kristen Ban, M.D.a, Zachary Fridirici, M.D.a, Anjali S. Godambe, D.O.b, Sherri Yong, M.D.b, Sam Pappas, M.D.a, Gerard J. Abood, M.D., M.S.a, Gerard V. Aranha, M.D., F.R.C.S.C., F.A.C.S.a,* a
Department of Surgery, bDepartment of Pathology, Loyola University Health Systems, 2160 South First Avenue, Maywood, IL 60153, USA
KEYWORDS: IPMN; Intraductal papillary mucinous neoplasm; Pancreas
Abstract BACKGROUND: The relationship between branch-duct intraductal papillary mucinous neoplasms (IPMNs) and malignancy remains controversial and difficult to assess. METHODS: Between January 1, 1999 and January 1, 2013, we identified 84 patients with IPMN who underwent resection. RESULTS: Preoperatively, 55 patients underwent endoscopic ultrasounds and 58 underwent biopsy. Only 7 lesions were specified preoperatively as branch-duct, which inconsistently correlated with the surgical specimen. Of the 82 patients where the duct was specified, there were 33 malignant lesions. There was no correlation between branch-duct origin and invasive carcinoma. Malignant tumor size did not significantly differ by the duct of origin. Of the 28 patients with invasive carcinoma, branchduct lesions were significantly associated with the presence of positive lymph nodes, perineural invasion, and lymphovascular invasion. CONCLUSIONS: Our study supports the resection criteria for branch-duct IPMN based on size and symptoms. However, it also questions the reliability of our preoperative testing to rule out malignant branch-duct IPMN lesions. Ó 2015 Elsevier Inc. All rights reserved.
Intraductal papillary mucinous neoplasms (IPMNs) are mucin-producing neoplasms that arise within the main pancreatic duct and/or major side branches. IPMNs can be There were no relevant financial relationships or any sources of support in the form of grants, equipment, or drugs. Funding was provided by the Department of Surgery, Loyola University Health Systems. * Corresponding author. Tel.: 11-708-327-2391; fax: 11-708-3273565. E-mail address: [email protected] Manuscript received July 22, 2014; revised manuscript October 28, 2014 0002-9610/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2014.10.010
classified according to their location as either main-duct IPMN, branch-duct IPMN, or mixed-type IPMN (which involves both main and branch ducts). Neoplasms involving the main pancreatic duct (main and mixed subtypes) have a well-demonstrated risk of malignancy, with 60% to 92% of these lesions harboring high-grade dysplasia (or carcinoma in situ) or invasive cancer on resection.1–3 The malignant risk of branch-duct IPMN remains more controversial and difficult to assess, but ranges from 15% to 25%.4 Specific features in branch-duct IPMN, including larger size, mural nodules, multiplicity, and epithelial subtype, are associated with an increased risk of malignancy.4 Numerous modalities
J.K. Plichta et al.
Potential branch-duct IPMN resection
Table 1 Select preoperative characteristics of the study population and lesions, stratified by the duct of origin specified in the final pathology report
Age (years), mean Male sex Size on preoperative imaging (cm), mean Mural nodule Asymptomatic Biopsy with dysplasia or malignant cells
Branch duct (n 5 28)
Main duct or mixed type (n 5 54)
P value
68 16 (57%) 3.2
71 26 (48%) 2.4
.18 .44 .03
1 (4%) 15 (54%) 16 (57%)
0 (0%) 16 (30%) 27 (50%)
.19 .06 .62
are available for the characterization of IPMN lesions, but the optimal test for preoperative identification of malignancy remains unclear. Consensus criteria have been developed to guide the evaluation and management of IPMN, which were most recently updated in 2012.2 Recommendations include surgical excision of all main-duct and mixed-type IPMNs versus observation for select branch-duct IPMNs (size ,3 cm, absence of symptoms and mural nodules, negative cytology). When to resect branch-duct IPMNs remains an area of particular controversy as a recent systematic review of these guidelines highlighted their poor positive predictive value (PPV) with regard to the detection of malignancy.5 The aim of this study was to review the presentation, preoperative evaluation, and final pathology from a consecutive series of IPMN patients who underwent resection at a single, large-volume academic center.
Methods Through a retrospective chart review, patients with IPMN lesions who underwent surgical resection from January 1, 1999 to January 5, 2013 at Loyola University Hospital Systems were identified. Clinicopathologic factors were assessed from the medical record, including clinical presentation, preoperative evaluation/imaging, surgeries performed, and pathologic results. Statistical analyses were conducted using Stata 10.0 (StataCorp, College Station, TX). Categorical variables were analyzed using chi-square tests, and continuous variables were analyzed using Mann–Whitney U tests or Student t tests. Multivariate analysis using logistic regression and reporting odds ratios was also performed. Statistical significance was defined as a P value less than .05 (2-sided). This study was approved by the Loyola University Health Systems Institutional Review Board.
Results There were 84 consecutive patients identified, including 43 men and 41 women. The median age of the patient
479 population was 71.5 years (range 49 to 88 years). Although 32 lesions were found incidentally, the majority of patients were symptomatic (Table 1). Most patients (62%) underwent pancreaticoduodenectomy, and the remainder had either a distal pancreatectomy (36%) or a central pancreatectomy (2%). Pathologic evaluation identified 28 branch-duct lesions and 54 main-duct lesions, of which 38 had concurrent branch-duct lesions; duct of origin data were unavailable for 2 specimens. Furthermore, 33 lesions had associated invasive (n 5 28) or in situ (n 5 5) carcinoma. Of the 38 lesions with dysplasia only, 11 were designated severe, 23 moderate, and 3 low-grade dysplasia. Branch-duct lesions with dysplasia (n 5 16) included the following: 3 lowgrade dysplasia, 10 moderate dysplasia, and 3 severe dysplasia. The presence of symptoms preoperatively was significantly associated with malignancy (51% vs 22%, P 5 .01). The most common symptoms were abdominal pain (54%) and weight loss (38%). At presentation, 12% of patients reported jaundice. Weight loss or jaundice on initial presentation was associated with a subsequent diagnosis of malignancy among patients with branch-duct lesions (both P , .05). Of the 31 incidentally noted lesions, only 7 were found to harbor malignancy, which was not associated with the duct of origin (P 5 .74). Before surgery, 55 patients underwent endoscopic ultrasounds (EUSs), 29 underwent endoscopic retrograde cholangiopancreatography (ERCP), and 10 underwent magnetic resonance cholangiopancreatography (MRCP). Preoperative imaging reported the duct of origin in 46 cases, including 6 lesions with mixed-type, 7 branch-duct only, and 33 mainduct only (Table 2). Branch-duct specification on imaging (n 5 7) inconsistently correlated with the surgical specimen (P 5 .67). Although 2 lesions were accurately identified as branch-duct (29%), the other 5 were ultimately diagnosed as mixed-type (71%). Dysplasia was present in 6 of the 7 lesions specified as branch-duct preoperatively, including 1 of the 2 lesions that were ultimately identified as branch-duct on final pathology (no invasive or in situ carcinoma was noted). Of the 58 patients who underwent preoperative biopsy (fine-needle aspiration, FNA), 55 reports were available for review. Pathology included 11 benign biopsies, 3 with dysplasia, and 41 concerning for malignancy. Positive cytology (n 5 41) on preoperative FNA was significantly associated with malignancy in the final specimen (P 5 .048); this yields a PPV of 51% and a negative predictive value of 86%. Of the 3 biopsies with dysplasia, final pathology revealed an invasive malignancy in 1 lesion and the other 2 accurately represented dysplasia alone. Presence or absence of mural nodules and size of the main pancreatic duct was not consistently documented. Only 1 mural nodule was noted on preoperative evaluation, and that lesion was noted to harbor dysplasia. On pathologic review, invasive carcinoma was reported in 9 of the 28 branch-duct lesions and 24 of the 54 mainduct/mixed-type lesions. There was no correlation between
480
The American Journal of Surgery, Vol 209, No 3, March 2015
Table 2 The duct of origin on preoperative imaging compared with the duct specified on final pathologic review, including the presence of malignancy Duct of origin on final pathologic review Duct of origin on preoperative imaging
Main duct Benign Malignant Branch duct Benign Malignant Mixed type Benign Malignant
branch-duct origin and invasive carcinoma (odds ratio .59, confidence interval .23 to 1.54). Multivariate analysis including tumor size did not influence this outcome (odds ratio .7, confidence interval .23 to 2). Malignant tumor size did not significantly differ by duct of origin (main-duct 3.3 cm, branch-duct 3 cm; P 5 .2). Of the 69 patients where the lesion size was documented preoperatively, size greater than 3 cm was significantly associated with malignancy (72% vs 23%, P , .001). Furthermore, 2 of the 9 malignant branch-duct lesions were less than 2 cm in size on preoperative imaging. Of the 6 patients with branch-duct lesions greater than or equal to 3 cm on preoperative imaging, only 50% harbored invasive malignancy. Of the 28 patients with invasive carcinoma, positive lymph nodes were present in 46% of the patients (n 5 13). Lymphovascular invasion was reported in 36% of the cases (n 5 10), and 61% of malignant lesions demonstrated perineural invasion (n 5 17). Malignant branch-duct lesions (n 5 9) were significantly associated with the presence of positive lymph nodes (n 5 8, P 5 .002), perineural invasion (n 5 9, P 5 .003), and lymphovascular invasion (n 5 6, P 5 .019).
Comments Because of increasing clinical awareness and widespread use of improved cross-sectional imaging, the incidence of patients presenting with cystic neoplasms of the pancreas has increased dramatically. Of the 4 common cystic tumors of the pancreas, a recent dramatic increase has been seen in IPMN.4 There is general agreement that main-duct IPMNs should be resected, because the frequency of malignancy in these patients varies between 60% and 92% and twothirds of those malignant neoplasms are invasive.1–3 Therefore, the guidelines for resection of main-duct IPMN are those patients who present with symptoms (ie, jaundice and/or worsening diabetes), main-duct diameter greater than 15 mm, and presence of a mural nodule, because all these are significant predictors of malignancy. The same can be said for mixed-type IPMN.2,3 Only recently has there been a suggestion that nonoperative management of
Main duct
Branch duct
Mixed type
2 7
4 7
6 6
0 0
2 0
5 0
0 0
1 0
3 2
main-duct IPMN may be indicated in select patients.6 Schmidt et al analyzed 334 patients with IPMN, and 171 had main-duct IPMN. Final pathology included 20% dysplasia and 27% invasive IPMN. They found that a normal preoperative carbohydrate antigen 19-9 and negative cytology were predictive of a lower malignancy rate. They concluded that patients who are at a higher risk because of age and comorbidity, especially those with negative cytology and a normal serum carbohydrate antigen 19-9, may be managed nonoperatively. However, this does not mitigate the fact that all fit patients with main-duct IPMN and mixed-type IPMN should undergo appropriate resection. The delineation in branch-duct IPMN is quite different and is the focus of this study. With branch-duct lesions, the frequency of malignancy varies between 6% and 46%, and the frequency of invasive cancer varies from 0% to 31%.3 Several factors have been found to be predictive of malignancy, including size of the cyst greater than 30 mm, the presence of a mural nodule, and symptoms.2,3 If these criteria are not present, the patients in one study showed no radiographic progression of the neoplasms with an average follow-up of 33 months.3 Based on this, the consensus of the working group on the International Association of Pancreatology suggested that the indications for surgery in branch-duct IPMN should be symptoms, a cyst size greater than 30 mm, and the presence of a mural nodule. However, the increased use of EUS and FNA, and improvement in the technology for computed tomography (CT) and magnetic resonance imaging in differentiating lesions that are malignant from those that are benign,7 led the consensus group to issue new guidelines for the management of branch-duct IPMN in 2012.2 The new guidelines accounted for the studies that showed an association with rapid increase in cyst size, high-grade atypia, and wall thickening. They also showed that cyst size greater than 30 mm was a weaker indication of malignancy. Based on these findings, the working group of the International Association of Pancreatology issued new guidelines for resection of branch-duct IPMN. These were that younger patients (,65 years of age) with a cyst size greater than 2 cm could be resected, because of the cumulative risk of
J.K. Plichta et al.
Potential branch-duct IPMN resection
malignancy. Presence of a mural nodule, positive cytology, and symptoms continued to be indications for surgery. These new guidelines by the consensus group have been supported by several studies.8,9 It would appear that the controversy would end there; however, several other authors have suggested caution in using the consensus guidelines. Fernandez-del Castillo suggested that the use of the consensus guidelines only would miss 8.8% of patients with hyperplasia and 6.5% of cases with invasive carcinoma.10 In a study of 138 patients with branch-duct IPMN, Lee et al found that a tumor size greater than 20 mm was the most valuable predictor of malignancy, and that tumors less than 20 mm with no mural nodule had an incidence of invasive cancer of 9.2%, but in a tumor less than 20 mm with the presence of a mural nodule, the incidence of invasive cancer was 25%.11 They suggested that many branch-duct IPMNs carry malignancy and surgery is recommended strongly, especially in the low-risk patient. Fritz et al analyzed a total of 287 consecutively resected IPMNs; of these lesions, 123 were branchduct IPMNs, 69 were less than 3 cm without mural nodules, thickening of the wall, or other characteristics of malignancy, the so-called ‘‘Sendai negative.’’12 Of all the Sendai-negative branch-duct IPMNs, 24.6% showed invasive carcinoma or high-grade dysplasia. Because of this, they suggested that the Sendai criteria may need to be adjusted and that resection should be used more liberally. Tanaka et al used 4 predictive factors to help determine malignancy in branch-duct IPMN, including (1) the presence of a cyst greater than or equal to 30 mm in diameter; (2) the presence of mural nodules; (3) a history of acute pancreatitis; and (4) atypia on pancreatic fluid cytology.13 When there are 2 or more predictive factors present, the sensitivity and specificity for malignancy was 96% and 71%, respectively. Ooi et al performed a systematic review of 12 studies on branch-duct IPMN, including 690 surgically treated branch-duct IPMNs, of which 24% were malignant.5 They concluded that the PPV of the Sendai Consensus Guidelines in predicting malignant branch-duct IPMN was low, and that some malignant lesions may be missed based on these criteria. Our study supports the resection criteria for branch-duct IPMN based on size and symptoms. However, tumor size did not clearly identify potentially malignant branch-duct lesions, and mural nodules were not common in our patient population. For the 15 asymptomatic branch-duct IPMN lesions, preoperative imaging suggested 2 branch-duct lesions, 6 main-duct lesions, and 7 indeterminate. Of the 2 predicted branch-duct lesions, preoperative biopsy results were suspicious for malignancy. Final pathology of the 15 asymptomatic branch-duct lesions yielded 3 invasive cancers (20%) and 10 lesions with dysplasia (67%). Utilization of EUS with biopsy has become increasingly common (43% 1999 to 2007 vs 83% 2008 to 2013) and it appears to be the most reliable predictor of malignancy in this study. Although it yielded a significant number of false positives, it will likely remain an important tool in the preoperative assessment of patients with IPMN lesions.
481 The use of tumor markers and cytokine analysis of pancreatic cyst fluid to differentiate malignant versus benign branch-duct IPMN is still in its infancy, but is gaining more recognition and may become a more useful tool in selecting patients for surgery. Finally, the question of follow-up for those patients with branch-duct IPMN who show a low probability for malignancy needs to be addressed. In one study, 49 patients with branch-duct IPMN who displayed a low probability of malignancy were followed for a mean follow-up of 77 months, and 77.5% of the patients remained free of symptoms.14 The study protocol included an initial MRCP or CT scan at 6 months, followed by a CT scan or EUS at 1 year. If there was no change, an MRCP at 2 years was performed, and if no change, an MRCP every 2 years was performed thereafter. This appears reasonable for those patients who have very low risk of harboring a malignancy in a branch-duct IPMN. However, the accuracy of these tests in diagnosing main versus mixed versus isolated branch-duct lesions likely remains a barrier to gaining widespread support. Although our study highlights several important considerations for the surgical management of branch-duct IPMN, it also has a few notable limitations. The study was performed as a retrospective chart review and thus limited some of the data collection and review. Although the pathologists are always thorough and meticulous with their specimen dissection and analysis, it is possible that the duct of origin could have been misreported or other IPMN lesions may have been misdiagnosed. In the earlier patients, the preoperative evaluation and diagnostic imaging utilized in our population was highly variable. Detailed radiologic review has always been routine, but the specific algorithm and modalities employed has since become more standardized. Finally, the actual number of patients with a diagnosed IPMN at our institution is largely unknown, because of the variety of providers that follow these patients and those who were likely lost to follow-up. As such, it is difficult to ascertain if our study population represents the expected proportion of patients with IPMN lesions. In addition, there are many hospitals in the surrounding area that also manage this particular patient population, and the data from these institutions were not available for our review. In summary, the malignant potential of IPMN remains clinically elusive based on current clinical strategies, particularly in regards to invasive characteristics of branch-duct IPMN. Furthermore, our study questions the reliability of our preoperative testing to rule out malignant branch-duct IPMN lesions. Until improved diagnostic imaging and preoperative risk stratification scores are established and validated, surgical resection should be considered for all IPMN lesions to reliably assess for concurrent malignancy in low-risk surgical patients.
References 1. Marchegiani G, Mino-Kenudson M, Sahora K, et al. IPMN involving the main pancreatic duct: biology, epidemiology, and long-term outcomes following resection. Ann Surg; 2014; [Epub ahead of print].
482 2. Tanaka M, Fernandez-del Castillo C, Adsay V, et al. International consensus guidelines 2012 for the management of IPMN and MCN of the pancreas. Pancreatology 2012;12:183–97. 3. Tanaka M, Chari S, Adsay V, et al. International consensus guidelines for management of intraductal papillary mucinous neoplasms and mucinous cystic neoplasms of the pancreas. Pancreatology 2006;6:17–32. 4. Farrell JJ, Fernandez-del Castillo C. Pancreatic cystic neoplasms: management and unanswered questions. Gastroenterology 2013;144: 1303–15. 5. Goh BK, Tan DM, Ho MM, et al. Utility of the sendai consensus guidelines for branch-duct intraductal papillary mucinous neoplasms: a systematic review. J Gastrointest Surg 2014;18:1350–7. 6. Roch AM, DeWitt JM, Al-Haddad MA, et al. Nonoperative management of main pancreatic duct-involved intraductal papillary mucinous neoplasm might be indicated in select patients. J Am Coll Surg 2014; 219:122–9. 7. Kim KW, Park SH, Pyo J, et al. Imaging features to distinguish malignant and benign branch-duct type intraductal papillary mucinous neoplasms of the pancreas: a meta-analysis. Ann Surg 2014;259:72–81. 8. Kobayashi G, Fujita N, Maguchi H, et al. Natural history of branch duct intraductal papillary mucinous neoplasm with mural nodules: a Japan Pancreas Society multicenter study. Pancreas 2014;43:532–8. 9. Jang JY, Park T, Lee S, et al. Validation of international consensus guidelines for the resection of branch duct-type intraductal papillary mucinous neoplasms. Br J Surg 2014;101:686–92. 10. Sahora K, Mino-Kenudson M, Brugge W, et al. Branch duct intraductal papillary mucinous neoplasms: does cyst size change the tip of the scale? A critical analysis of the revised international consensus guidelines in a large single-institutional series. Ann Surg 2013;258:466–75. 11. Jang JY, Kim SW, Lee SE, et al. Treatment guidelines for branch duct type intraductal papillary mucinous neoplasms of the pancreas: when can we operate or observe? Ann Surg Oncol 2008;15:199–205. 12. Fritz S, Klauss M, Bergmann F, et al. Small (sendai negative) branchduct IPMNs: not harmless. Ann Surg 2012;256:313–20. 13. Ohtsuka T, Kono H, Nagayoshi Y, et al. An increase in the number of predictive factors augments the likelihood of malignancy in branch duct intraductal papillary mucinous neoplasm of the pancreas. Surgery 2012;151:76–83. 14. Arlix A, Bournet B, Otal P, et al. Long-term clinical and imaging follow-up of nonoperated branch duct form of intraductal papillary mucinous neoplasms of the pancreas. Pancreas 2012;41:295–301.
Discussion Dr C. Max Schmidt (Indianapolis, IN). Of the 32 patients without symptoms, how many had branch duct IPMN? What were their indications for surgery? What percent ultimately had malignancy? And in light of these
The American Journal of Surgery, Vol 209, No 3, March 2015 data, do you recommend that all asymptomatic IPMN undergo pancreatic resection? The second question is, How many cases of branch duct IPMN were seen during this same time period at Loyola Medical Center and followed to determine their natural history without surgical resection. And the third question is, the minority branch duct involved. How much do you think the size of your study plays into how strong your conclusions should be about the management of IPMN going forward? Dr Plichta: To address your first question, of the patients that were without symptoms, there were actually 15 that were branch-duct, and of those, three of those patients ended up having malignant lesions, so about 20%. For the second question, we also agree that it is very difficult to interpret our data without having a denominator, knowing those patients that would be followed long-term with serial imaging. And then to your last question, obviously the size of our study is rather limiting. The next thing I think that we would need to do is to collaborate with another institution to maybe combine the data and that might be a better way to assess some of the things that we’ve looked at. Dr Margo C. Shoup (Warrenville, IL). Did you use CEA of the cystic fluid at all to predict malignancy? And what about PET scan? There is some data showing that it might be helpful for IPMN. Dr Plichta: So I haven’t looked particularly at our use of PET scan in this patient population. However, we do have data on both CA 19-9, CEA and some other factors that may be helpful in the future. We have yet to analyze those particular data parameters. Dr Aaron Sasson (Omaha, NE). If you can comment on the low utilization of your MR use, I would appreciate it. And, secondly, can you comment on how many of your branch duct IPMN’s were multifocal? Dr Plichta: MRCP is actually not utilized much at our institution, and I would say only about 10 patients, actually, of our 84 underwent MRCP. I would also have to go back and check to see how many were multifocal.
©2015 Elsevier
Midwest Surgical Association
Should all branch-duct intraductal papillary mucinous neoplasms be resected? Jennifer K. Plichta, M.D., M.S.a, Kristen Ban, M.D.a, Zachary Fridirici, M.D.a, Anjali S. Godambe, D.O.b, Sherri Yong, M.D.b, Sam Pappas, M.D.a, Gerard J. Abood, M.D., M.S.a, Gerard V. Aranha, M.D., F.R.C.S.C., F.A.C.S.a,* a
Department of Surgery, bDepartment of Pathology, Loyola University Health Systems, 2160 South First Avenue, Maywood, IL 60153, USA
KEYWORDS: IPMN; Intraductal papillary mucinous neoplasm; Pancreas
Abstract BACKGROUND: The relationship between branch-duct intraductal papillary mucinous neoplasms (IPMNs) and malignancy remains controversial and difficult to assess. METHODS: Between January 1, 1999 and January 1, 2013, we identified 84 patients with IPMN who underwent resection. RESULTS: Preoperatively, 55 patients underwent endoscopic ultrasounds and 58 underwent biopsy. Only 7 lesions were specified preoperatively as branch-duct, which inconsistently correlated with the surgical specimen. Of the 82 patients where the duct was specified, there were 33 malignant lesions. There was no correlation between branch-duct origin and invasive carcinoma. Malignant tumor size did not significantly differ by the duct of origin. Of the 28 patients with invasive carcinoma, branchduct lesions were significantly associated with the presence of positive lymph nodes, perineural invasion, and lymphovascular invasion. CONCLUSIONS: Our study supports the resection criteria for branch-duct IPMN based on size and symptoms. However, it also questions the reliability of our preoperative testing to rule out malignant branch-duct IPMN lesions. Ó 2015 Elsevier Inc. All rights reserved.
Intraductal papillary mucinous neoplasms (IPMNs) are mucin-producing neoplasms that arise within the main pancreatic duct and/or major side branches. IPMNs can be There were no relevant financial relationships or any sources of support in the form of grants, equipment, or drugs. Funding was provided by the Department of Surgery, Loyola University Health Systems. * Corresponding author. Tel.: 11-708-327-2391; fax: 11-708-3273565. E-mail address: [email protected] Manuscript received July 22, 2014; revised manuscript October 28, 2014 0002-9610/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2014.10.010
classified according to their location as either main-duct IPMN, branch-duct IPMN, or mixed-type IPMN (which involves both main and branch ducts). Neoplasms involving the main pancreatic duct (main and mixed subtypes) have a well-demonstrated risk of malignancy, with 60% to 92% of these lesions harboring high-grade dysplasia (or carcinoma in situ) or invasive cancer on resection.1–3 The malignant risk of branch-duct IPMN remains more controversial and difficult to assess, but ranges from 15% to 25%.4 Specific features in branch-duct IPMN, including larger size, mural nodules, multiplicity, and epithelial subtype, are associated with an increased risk of malignancy.4 Numerous modalities
J.K. Plichta et al.
Potential branch-duct IPMN resection
Table 1 Select preoperative characteristics of the study population and lesions, stratified by the duct of origin specified in the final pathology report
Age (years), mean Male sex Size on preoperative imaging (cm), mean Mural nodule Asymptomatic Biopsy with dysplasia or malignant cells
Branch duct (n 5 28)
Main duct or mixed type (n 5 54)
P value
68 16 (57%) 3.2
71 26 (48%) 2.4
.18 .44 .03
1 (4%) 15 (54%) 16 (57%)
0 (0%) 16 (30%) 27 (50%)
.19 .06 .62
are available for the characterization of IPMN lesions, but the optimal test for preoperative identification of malignancy remains unclear. Consensus criteria have been developed to guide the evaluation and management of IPMN, which were most recently updated in 2012.2 Recommendations include surgical excision of all main-duct and mixed-type IPMNs versus observation for select branch-duct IPMNs (size ,3 cm, absence of symptoms and mural nodules, negative cytology). When to resect branch-duct IPMNs remains an area of particular controversy as a recent systematic review of these guidelines highlighted their poor positive predictive value (PPV) with regard to the detection of malignancy.5 The aim of this study was to review the presentation, preoperative evaluation, and final pathology from a consecutive series of IPMN patients who underwent resection at a single, large-volume academic center.
Methods Through a retrospective chart review, patients with IPMN lesions who underwent surgical resection from January 1, 1999 to January 5, 2013 at Loyola University Hospital Systems were identified. Clinicopathologic factors were assessed from the medical record, including clinical presentation, preoperative evaluation/imaging, surgeries performed, and pathologic results. Statistical analyses were conducted using Stata 10.0 (StataCorp, College Station, TX). Categorical variables were analyzed using chi-square tests, and continuous variables were analyzed using Mann–Whitney U tests or Student t tests. Multivariate analysis using logistic regression and reporting odds ratios was also performed. Statistical significance was defined as a P value less than .05 (2-sided). This study was approved by the Loyola University Health Systems Institutional Review Board.
Results There were 84 consecutive patients identified, including 43 men and 41 women. The median age of the patient
479 population was 71.5 years (range 49 to 88 years). Although 32 lesions were found incidentally, the majority of patients were symptomatic (Table 1). Most patients (62%) underwent pancreaticoduodenectomy, and the remainder had either a distal pancreatectomy (36%) or a central pancreatectomy (2%). Pathologic evaluation identified 28 branch-duct lesions and 54 main-duct lesions, of which 38 had concurrent branch-duct lesions; duct of origin data were unavailable for 2 specimens. Furthermore, 33 lesions had associated invasive (n 5 28) or in situ (n 5 5) carcinoma. Of the 38 lesions with dysplasia only, 11 were designated severe, 23 moderate, and 3 low-grade dysplasia. Branch-duct lesions with dysplasia (n 5 16) included the following: 3 lowgrade dysplasia, 10 moderate dysplasia, and 3 severe dysplasia. The presence of symptoms preoperatively was significantly associated with malignancy (51% vs 22%, P 5 .01). The most common symptoms were abdominal pain (54%) and weight loss (38%). At presentation, 12% of patients reported jaundice. Weight loss or jaundice on initial presentation was associated with a subsequent diagnosis of malignancy among patients with branch-duct lesions (both P , .05). Of the 31 incidentally noted lesions, only 7 were found to harbor malignancy, which was not associated with the duct of origin (P 5 .74). Before surgery, 55 patients underwent endoscopic ultrasounds (EUSs), 29 underwent endoscopic retrograde cholangiopancreatography (ERCP), and 10 underwent magnetic resonance cholangiopancreatography (MRCP). Preoperative imaging reported the duct of origin in 46 cases, including 6 lesions with mixed-type, 7 branch-duct only, and 33 mainduct only (Table 2). Branch-duct specification on imaging (n 5 7) inconsistently correlated with the surgical specimen (P 5 .67). Although 2 lesions were accurately identified as branch-duct (29%), the other 5 were ultimately diagnosed as mixed-type (71%). Dysplasia was present in 6 of the 7 lesions specified as branch-duct preoperatively, including 1 of the 2 lesions that were ultimately identified as branch-duct on final pathology (no invasive or in situ carcinoma was noted). Of the 58 patients who underwent preoperative biopsy (fine-needle aspiration, FNA), 55 reports were available for review. Pathology included 11 benign biopsies, 3 with dysplasia, and 41 concerning for malignancy. Positive cytology (n 5 41) on preoperative FNA was significantly associated with malignancy in the final specimen (P 5 .048); this yields a PPV of 51% and a negative predictive value of 86%. Of the 3 biopsies with dysplasia, final pathology revealed an invasive malignancy in 1 lesion and the other 2 accurately represented dysplasia alone. Presence or absence of mural nodules and size of the main pancreatic duct was not consistently documented. Only 1 mural nodule was noted on preoperative evaluation, and that lesion was noted to harbor dysplasia. On pathologic review, invasive carcinoma was reported in 9 of the 28 branch-duct lesions and 24 of the 54 mainduct/mixed-type lesions. There was no correlation between
480
The American Journal of Surgery, Vol 209, No 3, March 2015
Table 2 The duct of origin on preoperative imaging compared with the duct specified on final pathologic review, including the presence of malignancy Duct of origin on final pathologic review Duct of origin on preoperative imaging
Main duct Benign Malignant Branch duct Benign Malignant Mixed type Benign Malignant
branch-duct origin and invasive carcinoma (odds ratio .59, confidence interval .23 to 1.54). Multivariate analysis including tumor size did not influence this outcome (odds ratio .7, confidence interval .23 to 2). Malignant tumor size did not significantly differ by duct of origin (main-duct 3.3 cm, branch-duct 3 cm; P 5 .2). Of the 69 patients where the lesion size was documented preoperatively, size greater than 3 cm was significantly associated with malignancy (72% vs 23%, P , .001). Furthermore, 2 of the 9 malignant branch-duct lesions were less than 2 cm in size on preoperative imaging. Of the 6 patients with branch-duct lesions greater than or equal to 3 cm on preoperative imaging, only 50% harbored invasive malignancy. Of the 28 patients with invasive carcinoma, positive lymph nodes were present in 46% of the patients (n 5 13). Lymphovascular invasion was reported in 36% of the cases (n 5 10), and 61% of malignant lesions demonstrated perineural invasion (n 5 17). Malignant branch-duct lesions (n 5 9) were significantly associated with the presence of positive lymph nodes (n 5 8, P 5 .002), perineural invasion (n 5 9, P 5 .003), and lymphovascular invasion (n 5 6, P 5 .019).
Comments Because of increasing clinical awareness and widespread use of improved cross-sectional imaging, the incidence of patients presenting with cystic neoplasms of the pancreas has increased dramatically. Of the 4 common cystic tumors of the pancreas, a recent dramatic increase has been seen in IPMN.4 There is general agreement that main-duct IPMNs should be resected, because the frequency of malignancy in these patients varies between 60% and 92% and twothirds of those malignant neoplasms are invasive.1–3 Therefore, the guidelines for resection of main-duct IPMN are those patients who present with symptoms (ie, jaundice and/or worsening diabetes), main-duct diameter greater than 15 mm, and presence of a mural nodule, because all these are significant predictors of malignancy. The same can be said for mixed-type IPMN.2,3 Only recently has there been a suggestion that nonoperative management of
Main duct
Branch duct
Mixed type
2 7
4 7
6 6
0 0
2 0
5 0
0 0
1 0
3 2
main-duct IPMN may be indicated in select patients.6 Schmidt et al analyzed 334 patients with IPMN, and 171 had main-duct IPMN. Final pathology included 20% dysplasia and 27% invasive IPMN. They found that a normal preoperative carbohydrate antigen 19-9 and negative cytology were predictive of a lower malignancy rate. They concluded that patients who are at a higher risk because of age and comorbidity, especially those with negative cytology and a normal serum carbohydrate antigen 19-9, may be managed nonoperatively. However, this does not mitigate the fact that all fit patients with main-duct IPMN and mixed-type IPMN should undergo appropriate resection. The delineation in branch-duct IPMN is quite different and is the focus of this study. With branch-duct lesions, the frequency of malignancy varies between 6% and 46%, and the frequency of invasive cancer varies from 0% to 31%.3 Several factors have been found to be predictive of malignancy, including size of the cyst greater than 30 mm, the presence of a mural nodule, and symptoms.2,3 If these criteria are not present, the patients in one study showed no radiographic progression of the neoplasms with an average follow-up of 33 months.3 Based on this, the consensus of the working group on the International Association of Pancreatology suggested that the indications for surgery in branch-duct IPMN should be symptoms, a cyst size greater than 30 mm, and the presence of a mural nodule. However, the increased use of EUS and FNA, and improvement in the technology for computed tomography (CT) and magnetic resonance imaging in differentiating lesions that are malignant from those that are benign,7 led the consensus group to issue new guidelines for the management of branch-duct IPMN in 2012.2 The new guidelines accounted for the studies that showed an association with rapid increase in cyst size, high-grade atypia, and wall thickening. They also showed that cyst size greater than 30 mm was a weaker indication of malignancy. Based on these findings, the working group of the International Association of Pancreatology issued new guidelines for resection of branch-duct IPMN. These were that younger patients (,65 years of age) with a cyst size greater than 2 cm could be resected, because of the cumulative risk of
J.K. Plichta et al.
Potential branch-duct IPMN resection
malignancy. Presence of a mural nodule, positive cytology, and symptoms continued to be indications for surgery. These new guidelines by the consensus group have been supported by several studies.8,9 It would appear that the controversy would end there; however, several other authors have suggested caution in using the consensus guidelines. Fernandez-del Castillo suggested that the use of the consensus guidelines only would miss 8.8% of patients with hyperplasia and 6.5% of cases with invasive carcinoma.10 In a study of 138 patients with branch-duct IPMN, Lee et al found that a tumor size greater than 20 mm was the most valuable predictor of malignancy, and that tumors less than 20 mm with no mural nodule had an incidence of invasive cancer of 9.2%, but in a tumor less than 20 mm with the presence of a mural nodule, the incidence of invasive cancer was 25%.11 They suggested that many branch-duct IPMNs carry malignancy and surgery is recommended strongly, especially in the low-risk patient. Fritz et al analyzed a total of 287 consecutively resected IPMNs; of these lesions, 123 were branchduct IPMNs, 69 were less than 3 cm without mural nodules, thickening of the wall, or other characteristics of malignancy, the so-called ‘‘Sendai negative.’’12 Of all the Sendai-negative branch-duct IPMNs, 24.6% showed invasive carcinoma or high-grade dysplasia. Because of this, they suggested that the Sendai criteria may need to be adjusted and that resection should be used more liberally. Tanaka et al used 4 predictive factors to help determine malignancy in branch-duct IPMN, including (1) the presence of a cyst greater than or equal to 30 mm in diameter; (2) the presence of mural nodules; (3) a history of acute pancreatitis; and (4) atypia on pancreatic fluid cytology.13 When there are 2 or more predictive factors present, the sensitivity and specificity for malignancy was 96% and 71%, respectively. Ooi et al performed a systematic review of 12 studies on branch-duct IPMN, including 690 surgically treated branch-duct IPMNs, of which 24% were malignant.5 They concluded that the PPV of the Sendai Consensus Guidelines in predicting malignant branch-duct IPMN was low, and that some malignant lesions may be missed based on these criteria. Our study supports the resection criteria for branch-duct IPMN based on size and symptoms. However, tumor size did not clearly identify potentially malignant branch-duct lesions, and mural nodules were not common in our patient population. For the 15 asymptomatic branch-duct IPMN lesions, preoperative imaging suggested 2 branch-duct lesions, 6 main-duct lesions, and 7 indeterminate. Of the 2 predicted branch-duct lesions, preoperative biopsy results were suspicious for malignancy. Final pathology of the 15 asymptomatic branch-duct lesions yielded 3 invasive cancers (20%) and 10 lesions with dysplasia (67%). Utilization of EUS with biopsy has become increasingly common (43% 1999 to 2007 vs 83% 2008 to 2013) and it appears to be the most reliable predictor of malignancy in this study. Although it yielded a significant number of false positives, it will likely remain an important tool in the preoperative assessment of patients with IPMN lesions.
481 The use of tumor markers and cytokine analysis of pancreatic cyst fluid to differentiate malignant versus benign branch-duct IPMN is still in its infancy, but is gaining more recognition and may become a more useful tool in selecting patients for surgery. Finally, the question of follow-up for those patients with branch-duct IPMN who show a low probability for malignancy needs to be addressed. In one study, 49 patients with branch-duct IPMN who displayed a low probability of malignancy were followed for a mean follow-up of 77 months, and 77.5% of the patients remained free of symptoms.14 The study protocol included an initial MRCP or CT scan at 6 months, followed by a CT scan or EUS at 1 year. If there was no change, an MRCP at 2 years was performed, and if no change, an MRCP every 2 years was performed thereafter. This appears reasonable for those patients who have very low risk of harboring a malignancy in a branch-duct IPMN. However, the accuracy of these tests in diagnosing main versus mixed versus isolated branch-duct lesions likely remains a barrier to gaining widespread support. Although our study highlights several important considerations for the surgical management of branch-duct IPMN, it also has a few notable limitations. The study was performed as a retrospective chart review and thus limited some of the data collection and review. Although the pathologists are always thorough and meticulous with their specimen dissection and analysis, it is possible that the duct of origin could have been misreported or other IPMN lesions may have been misdiagnosed. In the earlier patients, the preoperative evaluation and diagnostic imaging utilized in our population was highly variable. Detailed radiologic review has always been routine, but the specific algorithm and modalities employed has since become more standardized. Finally, the actual number of patients with a diagnosed IPMN at our institution is largely unknown, because of the variety of providers that follow these patients and those who were likely lost to follow-up. As such, it is difficult to ascertain if our study population represents the expected proportion of patients with IPMN lesions. In addition, there are many hospitals in the surrounding area that also manage this particular patient population, and the data from these institutions were not available for our review. In summary, the malignant potential of IPMN remains clinically elusive based on current clinical strategies, particularly in regards to invasive characteristics of branch-duct IPMN. Furthermore, our study questions the reliability of our preoperative testing to rule out malignant branch-duct IPMN lesions. Until improved diagnostic imaging and preoperative risk stratification scores are established and validated, surgical resection should be considered for all IPMN lesions to reliably assess for concurrent malignancy in low-risk surgical patients.
References 1. Marchegiani G, Mino-Kenudson M, Sahora K, et al. IPMN involving the main pancreatic duct: biology, epidemiology, and long-term outcomes following resection. Ann Surg; 2014; [Epub ahead of print].
482 2. Tanaka M, Fernandez-del Castillo C, Adsay V, et al. International consensus guidelines 2012 for the management of IPMN and MCN of the pancreas. Pancreatology 2012;12:183–97. 3. Tanaka M, Chari S, Adsay V, et al. International consensus guidelines for management of intraductal papillary mucinous neoplasms and mucinous cystic neoplasms of the pancreas. Pancreatology 2006;6:17–32. 4. Farrell JJ, Fernandez-del Castillo C. Pancreatic cystic neoplasms: management and unanswered questions. Gastroenterology 2013;144: 1303–15. 5. Goh BK, Tan DM, Ho MM, et al. Utility of the sendai consensus guidelines for branch-duct intraductal papillary mucinous neoplasms: a systematic review. J Gastrointest Surg 2014;18:1350–7. 6. Roch AM, DeWitt JM, Al-Haddad MA, et al. Nonoperative management of main pancreatic duct-involved intraductal papillary mucinous neoplasm might be indicated in select patients. J Am Coll Surg 2014; 219:122–9. 7. Kim KW, Park SH, Pyo J, et al. Imaging features to distinguish malignant and benign branch-duct type intraductal papillary mucinous neoplasms of the pancreas: a meta-analysis. Ann Surg 2014;259:72–81. 8. Kobayashi G, Fujita N, Maguchi H, et al. Natural history of branch duct intraductal papillary mucinous neoplasm with mural nodules: a Japan Pancreas Society multicenter study. Pancreas 2014;43:532–8. 9. Jang JY, Park T, Lee S, et al. Validation of international consensus guidelines for the resection of branch duct-type intraductal papillary mucinous neoplasms. Br J Surg 2014;101:686–92. 10. Sahora K, Mino-Kenudson M, Brugge W, et al. Branch duct intraductal papillary mucinous neoplasms: does cyst size change the tip of the scale? A critical analysis of the revised international consensus guidelines in a large single-institutional series. Ann Surg 2013;258:466–75. 11. Jang JY, Kim SW, Lee SE, et al. Treatment guidelines for branch duct type intraductal papillary mucinous neoplasms of the pancreas: when can we operate or observe? Ann Surg Oncol 2008;15:199–205. 12. Fritz S, Klauss M, Bergmann F, et al. Small (sendai negative) branchduct IPMNs: not harmless. Ann Surg 2012;256:313–20. 13. Ohtsuka T, Kono H, Nagayoshi Y, et al. An increase in the number of predictive factors augments the likelihood of malignancy in branch duct intraductal papillary mucinous neoplasm of the pancreas. Surgery 2012;151:76–83. 14. Arlix A, Bournet B, Otal P, et al. Long-term clinical and imaging follow-up of nonoperated branch duct form of intraductal papillary mucinous neoplasms of the pancreas. Pancreas 2012;41:295–301.
Discussion Dr C. Max Schmidt (Indianapolis, IN). Of the 32 patients without symptoms, how many had branch duct IPMN? What were their indications for surgery? What percent ultimately had malignancy? And in light of these
The American Journal of Surgery, Vol 209, No 3, March 2015 data, do you recommend that all asymptomatic IPMN undergo pancreatic resection? The second question is, How many cases of branch duct IPMN were seen during this same time period at Loyola Medical Center and followed to determine their natural history without surgical resection. And the third question is, the minority branch duct involved. How much do you think the size of your study plays into how strong your conclusions should be about the management of IPMN going forward? Dr Plichta: To address your first question, of the patients that were without symptoms, there were actually 15 that were branch-duct, and of those, three of those patients ended up having malignant lesions, so about 20%. For the second question, we also agree that it is very difficult to interpret our data without having a denominator, knowing those patients that would be followed long-term with serial imaging. And then to your last question, obviously the size of our study is rather limiting. The next thing I think that we would need to do is to collaborate with another institution to maybe combine the data and that might be a better way to assess some of the things that we’ve looked at. Dr Margo C. Shoup (Warrenville, IL). Did you use CEA of the cystic fluid at all to predict malignancy? And what about PET scan? There is some data showing that it might be helpful for IPMN. Dr Plichta: So I haven’t looked particularly at our use of PET scan in this patient population. However, we do have data on both CA 19-9, CEA and some other factors that may be helpful in the future. We have yet to analyze those particular data parameters. Dr Aaron Sasson (Omaha, NE). If you can comment on the low utilization of your MR use, I would appreciate it. And, secondly, can you comment on how many of your branch duct IPMN’s were multifocal? Dr Plichta: MRCP is actually not utilized much at our institution, and I would say only about 10 patients, actually, of our 84 underwent MRCP. I would also have to go back and check to see how many were multifocal.
©2015 Elsevier
