Intraoperative ultrasonography by high-resolution annular array transducer for intraductal papillary mucinous tumors of the pancreas

Intraoperative ultrasonography by high-resolution annular array transducer for intraductal papillary mucinous tumors of the pancreas Tetsuya Kaneko, MD, Akimasa Nakao, MD, Soichiro Inoue, MD, Hiroyuki Sugimoto, MD, Tsuyoshi Hatsuno, MD, Akihiro Ito, MD, Yoshiki Hirooka, MD, Tetsuro Nagasaka, MD, and Nobuo Nakashima, MD, Nagoya, Japan

Background. Intraductal papillary mucinous tumors of the pancreas (IPMT) spread along the main pancreatic duct (MPD) or branch duct, or both. It is important to intraoperatively determine the extent of IPMT lesions in the MPD to perform a successful operation. Methods. This study included 27 consecutive patients with IPMT who were referred for operation and who had undergone endoscopic ultrasonography, endoscopic retrograde pancreatography, and computed tomography as a diagnostic examination. Intraoperative ultrasonography with a 7.5-MHz annular array transducer and a 7.5-MHz conventional transducer was performed. All patients underwent operation, and the pathology was confirmed. Endoscopic ultrasonography, endoscopic retrograde pancreatography, intraoperative conventional ultrasonography, and intraoperative annular array ultrasonography results were correlated with those from a pathologic examination of the resected specimens. The diagnostic accuracy of these 4 modalities in the detection of IPMT lesions in the MPD and the branch duct was confirmed. Results. The diagnostic criterion of IPMT lesions by intraoperative annular array ultrasonography was the presence of echogenic masses with irregular margins in the MPD or branch duct, or both. An IPMT lesion in the MPD was found in 22 of the 27 patients. Intraoperative annular array ultrasonography detected 5 cases of IPMT lesions in the MPD that could not be detected by endoscopic ultrasonography or endoscopic retrograde pancreatography. In 3 of 5 patients, pancreatic resection was extended to remove IPMT lesions. Intraoperative annular array ultrasonography detected 3 multifocal lesions that could not be detected by preoperative examinations. For the diagnosis of IPMT lesions in the MPD, the respective sensitivity, specificity, and overall accuracy of intraoperative annular array ultrasonography were 86%, 100%, and 89%; the respective values were 50%, 100%, and 59% for endoscopic retrograde pancreatography; 59%, 100%, and 67% for endoscopic ultrasonography, and 32%, 100%, and 44% for intraoperative conventional ultrasonography. For the diagnosis of IPMT lesions in the branch duct, the respective sensitivity, specificity, and overall accuracy of intraoperative annular array ultrasonography were 95%, 100%, and 96%. Conclusions. Intraoperative annular array ultrasonography is suitable for longitudinal scanning of the MPD, useful for accurate diagnosis of the extent of IPMT in the pancreas, and valuable for planning surgical strategy and successful operations for IPMT. (Surgery 2001;129:55-65.) From the Department of Surgery II, the Department of Internal Medicine II, and the Department of Laboratory Medicine, Faculty of Medicine, University of Nagoya, Nagoya, Japan

Accepted for publication June 2, 2000. Reprint requests: Tetsuya Kaneko, Department of Surgery II, Faculty of Medicine, University of Nagoya, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan. Copyright © 2001 by Mosby, Inc. 0039-6060/2001/$35.00 + 0 doi:10.1067/msy.2001.109118

11/56/109118

A MUCIN-PRODUCING TUMOR of the pancreas is a relatively new clinical entity that is characterized by dilatation of the main pancreatic duct (MPD) or pancreatic duct branch, or both, with copious mucin secretion.1,2 Although many names have been given to this disease, the term intraductal papillary mucinous tumor (IPMT) of the pancreas is used in accorSURGERY 55

56 Kaneko et al

Surgery January 2001

technology is the excellent lateral resolution provided by continuous focusing (Fig 1).10 Thus, it can produce good images of longitudinal scanning of MPD. The annular array transducer is an adequate imaging technique for depicting the extent of IPMT in the intrapancreatic duct. The aim of this prospective study is to evaluate the diagnostic accuracy of the intraoperative US with annular array transducer for intraductal and intracystic mural nodules of IPMT and to compare it with other conventional imaging modalities.

Fig 1. Schema of annular array ultrasound beam. Bold lines indicate focusing zones. The width of the ultrasound beam is approximately equal over the entire depth, and lateral resolution is consequently improved, in particular in the near and far range.

dance with the classification of the World Health Organization.3 With recent advances in our understanding of IPMT, the number of reported cases has increased.4,5 IPMT includes a wide range of pathologic conditions such as adenocarcinoma, adenoma, and hyperplasia.6 It is obvious that adenocarcinoma means malignant, adenoma means premalignant, and hyperplasia means benign. Nevertheless, it is difficult to perform a differential diagnosis among these conditions preoperatively except for an obvious invasion into the extrapancreatic tissues or portal vein.7 Even when the tumor is malignant, IPMT is often resectable and has a favorable prognosis, as compared with duct cell carcinoma. With the recent progress in imaging technology, a preoperative diagnosis of IPMT can now be performed accurately to some extent.8 But there are some problems to be solved in the diagnosis of IPMT, including longitudinal spread in the MPD. From a surgical point of view, it is important to confirm the tumor location and its extent into the intrapancreatic duct intraoperatively to obtain surgical curability. To determine the surgical resection line of the pancreas intraoperatively, the intrapancreatic ductal extent and intrapancreatic location of IPMT are indispensable, and “a traditional evaluation of the pancreas by palpation and inspection” is insufficient. Moreover, conventional intraoperative ultrasonography (US) lacks sufficient resolution for visualization of the intrapancreatic ductal spread of IPMT.9 Among recent advances in US technology, a high-frequency annular array US transducer was developed to provide superior, high-resolution images. A major advantage of this annular array

PATIENTS AND METHODS From February 1995 to March 1999, twentyseven consecutive patients with IPMT were prospectively studied with the intraoperative US annular array transducer. The study group consisted of 19 men and 8 women with a mean age of 61 years (range, 42 -77 years). IPMT was defined as a dilatation of the MPT or branches, or both, with mucin excretion through a patulous orifice of the enlarged papilla. According to the classification of the Japan Pancreatic Society, the 27 IPMTs were classified into either MPD type or branch-duct type in their principal location. There were 4 MPD types and 23 branch-duct types. The IPMT was located at the head of the pancreas in 15 cases, at the body in 10 cases, at the tail in 1 case, and at the body to tail in 1 case. In all patients, pancreatic resection and pathologic examination were performed. Preoperative examination. Endoscopic US (EUS), endoscopic retrograde pancreatography (ERP), and computed tomography (CT) including dynamic study were performed in all patients. ERP was performed by an expert endoscopist (A. I.). A well-defined filling defect in the MPD or its branches, or both, was the diagnostic criterion for IPMT. ERP images were reviewed without knowledge of other examination results. EUS was carried out with a 7.5-MHz rotating transducer (GF-UM 2; Olympus Optical, Tokyo, Japan) by an expert endoscopist (Y. H.). EUS images were also reviewed without knowledge of the results of other examinations. The diagnostic criterion for IPMT was hyperechoic vegetation in the MPD or its branches, or both. EUS was performed 1 week before ERCP to avoid interference with EUS imaging. Intraoperative US examination. Intraoperative annular array US was performed with a 7.5-MHz US transducer with an incorporated standoff pad (SMA-736 SA; Toshiba, Tokyo, Japan). Conventional intraoperative US was performed with a 7.5-MHz US transducer (SSD 680; Aloka, Tokyo, Japan). In this series, all patients underwent intra-

Surgery Volume 129, Number 1

Kaneko et al 57

Fig 2. Intraoperative annular array ultrasonography of 7.5-MHz probe. Longitudinal scanning of the main pancreatic duct (arrow).

operative US with a conventional transducer. The scanhead was sterilized by cold gas and hand-held in the operative field. The scanhead was positioned over the pancreas. First, longitudinal scanning of the pancreas was done to visualize the longitudinal section of the MPD and its branches for the detection of mural nodules of IPMT. Next, transverse scanning of the pancreas was performed to confirm the findings of longitudinal scanning. Both intraoperative conventional and annular array US were performed by a single examiner (T. K.), a board-certified US expert of the Japan US Society. The reader (T. K.) of intraoperative US studies was blinded to the results of other imaging procedures. The mural nodules in the MPD were confirmed by an ultrathin pancreatoscope and a pathologic examination of the resected specimens. The mural nodules in the branch duct were confirmed by pathologic examination of the resected specimens. Pancreatoscopic examination. Pancreatoscopy was performed intraoperatively with an ultrathin pancreatoscope (Clinical Supply Co, Ltd, Gifu, Japan) to confirm the IPMT lesion in the MPD. The pancreatic resection line was decided according to the intraoperative US findings. The pancreatoscope was inserted into the MPD from the cut surface of the pancreas. First, the MPD of the resection side was examined, followed by the MPD on the nonresection side. This examination was performed by an expert (H. S.). The diagnostic criterion for IPMT in the MPD was papillary mural nodules with an appearance similar to that of salmon roe. Pathologic examination. The resected pancreatic specimens were set in 10%-buffered formalin for

routine histopathologic examination. The prepared slices were then stained with hematoxylineosin. Histologic evaluation of the specimens was performed by 2 pathologists (T. N. and N. N.). Evaluation of diagnostic accuracy. Images provided by EUS, ERP, intraoperative conventional US, and annular array US were each prospectively evaluated and correlated with pancreatoscopic and histologic findings of the resected specimens. The diagnostic criteria for IPMT were mural nodules with cystic dilatation of the MPD or branch duct. The accuracy of the differential diagnosis between IPMT and other cystic lesions was confirmed by pathologic examination of resected specimens. Serous cystadenoma contains numerous small cysts with a spongy or honeycomb pattern. A pseudocyst contains a unilocular, isolated cyst without mural nodules. The gold standard we established to confirm the presence of the nodules was pathology of the resected specimen or biopsy. Detectability of mural nodules of IPMT in the MPD and the branch duct was evaluated separately. Results of intraoperative annular array US were compared with other imaging techniques (ERP, EUS, and intraoperative conventional US). A comparison of statistical significance was analyzed by means of a Fisher exact probability test. Differences were considered significant at a value of less than .05. RESULTS Diagnosis of IPMT lesions in the MPT. Intraoperative annular array US was performed safely in all cases, providing a clear, sharp longitudinal image of the MPD (Fig 2). IPMT lesions in the

58 Kaneko et al

Surgery January 2001

Fig 3. Intraoperative annular array ultrasonography of 7.5-MHz probe. Mural nodule in the main pancreatic duct (MPD).

Fig 4. Intraoperative annular array ultrasonography of 7.5-MHz probe. Mural nodule in the branch duct (arrow). Internal structure of the cystic dilatation of the branch duct was clearly visualized.

MPD were visualized as dilated MPD with mural nodules. The nodules were demonstrated as homogeneously echogenic masses with irregular margins in the MPD (Fig 3). Intraoperative annular array US could differentiate between mural nodules and clumps of mucin. These clumps were visualized as hypoechoic amorphous shapes with slightly echogenic layers. Of 27 resected cases, hyperplasia accounted for 8, adenoma for 12, and carcinoma for 7. IPMT lesions of mural nodules in the MPD were confirmed pathologically in 22 of 27 cases.

A breakdown of mural nodules in the MPD was as follows: In 4 MPD type cases, mural nodules in the MPD were confirmed in all cases. In 18 of 23 branch-duct types, the IPMT lesion extended to the MPD from the branch-duct lesion. In 5 of 23 branch-duct types, IPMT lesions were located within the branch. Intraoperative annular array US detected 19 of 22 nodules in the MPD, and the remaining 3 nodules were detected by intraoperative pancreatoscopy only. All but 1 of 22 MPD lesions were resected and confirmed to be present by the pathologist.

Kaneko et al 59

Surgery Volume 129, Number 1

Table I. Diagnostic values for detection of mural nodules in the main pancreatic duct and comparative accuracies of ERP, EUS, conventional IOUS, and annular array IOUS ERP True-positive True-negative False-positive False-negative Sensitivity (%) Specificity (%) Overall accuracy (%)

11 5 0 11 50* 100 59.2*

EUS

Conventional IOUS

13 5 0 9 59 100 66.7

7 5 0 15 32† 100 44.4†

Annular array IOUS 19 5 0 3 86.4*† 100 88.9*†

IOUS, Intraoperative ultrasonography. *P < .05, Annular array IOUS versus ERP. †P < .01, Annular array IOUS versus conventional IOUS.

Table II. Diagnostic values for detection of mural nodules in the branch duct and comparative accuracies of ERP, EUS, conventional IOUS, and annular array IOUS True-positive True-negative False-positive False-negative Sensitivity (%) Specificity (%) Overall accuracy (%)

ERP

EUS

Conventional IOUS

15 1 0 7 68.2 100 69.6*

21 1 0 1 95.4 100 95.6

16 1 0 6 72.7 100 74

Annular array IOUS 21 1 0 1 95.4 100 95.6*

IOUS, Intraoperative ultrasonography. *P < .05, Annular array IOUS versus ERP.

The remaining multifocal lesion was not resected, transductal biopsy from the cut surface was performed, and the pathologic diagnosis was made. Mural nodules of IPMT lesions in the MPD all ranged in size from 1.5 mm to 21 mm with an average size of 6.7 mm. The detectability of IPMT lesions by imaging modalities depended upon the size of the lesion. IPMT lesions of mural nodules in the MPD were confirmed in 4 of 8 hyperplasia cases. ERP detected 1 of 4 hyperplasia cases of IPMT lesions in the MPD. EUS detected 1 of 4, intraoperative conventional US detected 0 of 4, and intraoperative annular array US detected 2 of 4 such cases. IPMT lesions of mural nodules in the MPD were confirmed in 11 of 12 adenoma cases. ERP detected 6 of 11 adenoma cases of IPMT lesions in the MPD. EUS detected 6 of 11, intraoperative conventional US detected 4 of 11, and intraoperative annular array US detected 10 of 11. IPMT lesions of mural nodules in the MPD were confirmed in all 7 carcinoma cases. ERP detected 4 of 7 carcinoma cases. EUS detected 5 of 7, intraoperative conventional US detected 3 of 7, and intraoperative annular array US detected 7 of 7.

In all cases, the diagnostic values for the detection of mural nodules in the MPD were as follows: In ERP, sensitivity, specificity, and overall accuracy were 50%, 100%, and 59%, respectively. In EUS, the respective diagnostic values were 59%, 100%, and 67%. In intraoperative conventional US, they were 32%, 100%, and 44.4%, and in intraoperative annular array US, they were 86%, 100%, and 89%, respectively. The sensitivity of the intraoperative annular array US in detection of mural nodules in the MPD was statistically superior to that of ERP (P = .0217) and conventional intraoperative US (P = .0005). The overall accuracy of intraoperative annular array US in the detection of mural nodules in the MPD was statistically superior to that of ERP (P = .0276) and conventional intraoperative US (P = .0011) (Table I). Diagnosis of IPMT lesions in the branch duct. IPMT lesions in the branch duct were visualized as clustered cysts with mural nodules shown as echogenic masses with irregular margins (Fig 4). In 23 resected cases of the branch-duct type, hyperplasia accounted for 7, adenoma for 11, and carcinoma for 5. IPMT lesions of mural nodules in the branch duct were confirmed in 21 of 23 cases by pathologic examination.

60 Kaneko et al

Surgery January 2001

Fig 5. Endoscopic retrograde pancreatography showed cystic dilatation of the branch duct in the pancreatic head.

Twelve cystic lesions in the branch duct were located in the pancreatic head, 9 in the body, and 1 in the body to tail. The average size of cystic lesions was 30.4 mm × 20.9 mm. In 18 benign lesions, mural nodules were located within the cystic lesions. IPMT lesions of mural nodules in the branch duct were confirmed in 6 of 7 hyperplasia cases. ERP detected 3 of 6 hyperplasia cases of IPMT lesions in the branch duct. EUS detected 5 of 6, intraoperative conventional US detected 3 of 6, and intraoperative annular array US detected 5 of 6. IPMT lesions of mural nodules in the branch duct were confirmed in 11 of 11 adenoma cases. ERP detected 7 of 11 hyperplasia cases of IPMT lesions in the branch duct. EUS detected 11 of 11, intraoperative conventional US detected 8 of 11, and intraoperative annular array US detected 11 of 11. IPMT lesions of mural nodules in the branch duct were confirmed in 5 of 5 carcinoma cases. ERP detected 3 of 5 carcinoma cases of IPMT lesions in the branch duct. EUS detected 5 of 5, intraoperative conventional US 4 of 5, and intraoperative annular array US 5 of 5. In all cases, diagnostic values for the detection of mural nodules in the branch duct were as follows: In ERP, sensitivity, specificity, and overall accuracy were 68%, 100%, and 70%, respectively. In EUS, the respective diagnostic values were 95%, 100%, and 96%. In intraoperative conventional US, they were 73%, 100%, and 70%, and in intraoperative annular array US, they were 95%, 100%, and 96%. The overall accuracy of the intraoperative annular array US in detection of mural

nodules in the branch duct was statistically superior to ERP (P = .047) and intraoperative conventional US (P = .047) (Table II). Surgical results. The operation was conducted with the help of annular array US in terms of branch-duct type IPMT spread into the MPD, multifocal lesions in the MPD, longitudinal extent of IPMT lesion in the MPD, and intrapancreatic or extrapancreatic infiltration. First, intraoperative annular array US detected 5 cases of IPMT lesions in the MPD that could not be found by ERP or EUS. These 5 cases were branchduct type. In 2 of 5 cases, IPMT lesions in the MPD were within 1 cm of the orifice of the dilated branch duct, so resection lines were not changed. In 3 of 5 cases, IPMT lesions extended more than 1 cm along the MPD at the dilated branch duct. In these 3 cases, resection lines were changed to remove the lesions. A breakdown of the 3 cases was as follows: The first case was a carcinoma in the body, for which a resection line was extended to the proximal (head) side. A distal pancreatectomy was performed. The second case was an adenoma in the head, for which a resection line was extended to the distal side. A pylorus-preserving pancreaticoduodenectomy was performed. The third case was an adenoma in the body for which a resection line was extended to the proximal side. A segmental resection was performed. In all 5 cases, an intraoperative frozen-section analysis of the pancreatic transection margin showed negative results. Secondly, there were 6 cases of multifocal IPMT lesions. Three of 6 cases were detected by intraoperative annular array US. There was 1 multifocal lesion in the MPD and 2 in the branch duct. In the

Kaneko et al 61

Surgery Volume 129, Number 1

A

B Fig 6. A, Intraoperative annular array ultrasonography showing the septum and mural nodules (arrowheads) in the cystic dilatation of the branch duct of the uncus. B, Intraoperative annular array ultrasonography showing the mural nodule in the main pancreatic duct of the body (arrow).

former, a preoperatively diagnosed IPMT lesion was found at the pancreatic head, and the operation proposed was pancreaticoduodenectomy (Fig 5). Intraoperative annular array US detected another lesion in the pancreatic body (Fig 6, A and B). In this case, pancreaticoduodenectomy was performed, and 2 mL of absolute ethanol was injected into the remaining pancreatic duct to ablate the multifocal IPMT lesion in the body, thus avoiding a total pancreatectomy.11 Transient hyperamylasemia was observed in this patient without any complication. The safety and the therapeutic effectiveness of this procedure were confirmed in an experimental dog model (unpublished data) by the author (T. K.). This method was unconventional, however, and was used only in this 1 patient. This procedure was considered to be the last viable option to avoid total pancreatectomy. The patient has remained

well 4 years after the operation with a good postoperative course. In 1 of 2 cases of multifocal lesions in the branch duct, the resection line was extended to remove the multifocal lesions, and a segmental resection was performed. In another case, IPMT lesions in the uncus and in the pancreatic body were detected by intraoperative annular array US. In this patient, a partial resection of the uncus and a segmental resection of the pancreatic body were performed. The third group of cases involved the longitudinal extent of IPMT in the MPD. In 2 cases of IPMT in the pancreatic body, IPMT lesions in the MPD extended more to the head side of the pancreas than at the point diagnosed preoperatively by EUS and ERP. Intraoperative annular array US showed that the IPMT lesion extended to the pancreas head beyond the right side of the portal vein (Fig 7).

62 Kaneko et al

Surgery January 2001

Fig 7. Intraoperative annular array ultrasonography showing the tumor in the main pancreatic duct (between arrows) extended beyond the right side of the portal vein.

Thus, the pancreas was amputated at the pancreas head, and complete removal was confirmed by a frozen section of the resected specimen and intraoperative pancreatoscopy. The fourth group had either invasion to the pancreatic parenchyma or extrapancreatic tissue in malignant cases. There were 7 malignant cases. Two were MPD type and 5 were branch-duct type. In 2 MPD type carcinomas, the average tumor size, including the cystic dilated portion, was 25 mm × 15 mm. In both cases, invasion to the pancreatic parenchyma was observed, but it remained within the pancreas. No lymph node metastasis was found. In terms of invasion to the pancreatic parenchyma, intraoperative annular array US detected both cases, whereas EUS detected 1, and ERCP could not detect any. Therefore, resectability was not a problem. In 5 branch-duct type carcinomas, the average tumor size, including the cystic portion, was 41 mm × 36 mm. Invasion to the pancreatic parenchyma was detected in 3 of 5 cases. In 2 of 3 cases, invasion remained within the pancreas, and in the remaining case extrapancreatic invasion was detected. In terms of invasion to the pancreatic parenchyma, intraoperative annular array US diagnosed all cases, and EUS diagnosed 2. ERCP could not diagnose invasion to the pancreatic parenchyma because it reflected only the luminal profiles of the MPD and branch. In one case of extrapancreatic invasion, retroperitoneal invasion of the IPMT lesion at the pancreatic tail was diagnosed by intraoperative annular array US. This could not be found by other imaging modalities. Extended distal

pancreatectomy, including the retroperitoneal tissue, was performed in this patient. Both intraoperative annular array US and EUS detected 1 case of lymph node metastasis. In the staging of malignant IPMT, intraoperative annular array US is more accurate than EUS for local invasion of an IPMT lesion. However, it is difficult to discriminate malignant infiltration from focal pancreatitis in the pancreatic parenchyma by annular array US. In this study, we did not experience focal pancreatitis with IPMT lesion. In terms of the differential diagnosis between benign and malignant lesions, mural nodule size was a differential point, but it proved difficult to discriminate adenocarcinoma, especially in situ carcinomas, from adenoma, the exception being adenocarcinoma with intrapancreatic or extrapancreatic infiltration. All patients underwent surgical resection, including pylorus-preserving pancreaticoduodenectomy in 11, segmental resection of the pancreas in 7 (in 1 case with partial resection of the pancreatic head), distal pancreatectomy in 6, and pancreaticoduodenectomy in 3. The mean diameter of the intraductal mural nodule was 2.9 mm in the 8 hyperplasias, 6.1 mm in the 12 adenomas, and 9.8 mm in the 7 adenocarcinomas. Follow-up information was obtained on all 27 patients. One patient with carcinoma died of lung metastasis 2.5 years after resection with curative intent. Right pleural metastasis was found in 1 patient with carcinoma 3 years after operation, but she remains alive 4 years after resection. All but 1 patient survived within the 5-year follow-up period.

Surgery Volume 129, Number 1 DISCUSSION IPMT is a slow growing and low malignancy tumor. Complete removal of the tumor results in a good prognosis.12-14 Mural nodules denote a tumorous growth component and reflect a proliferative condition, thus constituting a surgical indication.15 In terms of surgical exploration, it is important to determine the extent of IPMT lesions in the MPD.16 IPMT spreads along the pancreatic duct with a replacement of normal epithelium, and mural nodules are an imaging marker of the intraductal spread.17,18 Main-duct type IPMT extends along the MPD, and branch-duct type IPMT extends even into the MPD. Yamada et al 19 reported that malignant cases of the branch-duct type extended into the MPD in 67% of the patients and benign cases extended into the MPD in 29% of the patients. In the current study, malignant IPMT of the branch-duct type extended into the MPD, and benign IPMT of the branch-duct type extended into the MPD in 72% of the cases. Cellier et al20 found that CT, ERP, and EUS failed to accurately determine the extent of tumors in more than 30% of patients. In the current study, 7 of 27 patients had carcinomas that included both invasive and in situ carcinoma. This is a rather low incidence of carcinoma. Gastroenterologists in our university hospital take into consideration whether the surgical indication for IPMT is a mural nodule in the MPD or cystic dilatation of the branch duct. These diagnostic criteria for surgical excision include a wide range of pathologic entities, from benign hyperplasia and premalignant adenoma to malignant adenocarcinoma. This is the reason for the rather low incidence of carcinoma in this study. In our surgical department, IPMT lesions were confirmed by intraoperative US, and the resection line of the pancreas was determined. Longitudinal scanning of the MPD for detecting IPMT lesions by annular array US provided important information in determining the resection line of the pancreas because the excellent lateral resolution of annular array US provided a clear longitudinal MPD image. Thus, the intraoperative US by annular array US was a first-step examination in the operative strategy for IPMT. After amputation of the pancreas according to the findings of the intraoperative annular array US, an intraoperative frozen-section analysis of the pancreatic transection margin was performed to confirm the intraoperative US findings. In terms of the determination of the longitudinal extent of IPMT, intraoperative annular array US is superior to conventional preoperative imaging. Consequently, accurate intraoperative diagno-

Kaneko et al 63

sis of IPMT saves operation time. The role of intraoperative ultrathin pancreatoscopy was to check for small IPMT mural nodules that might be overlooked by intraoperative annular array US in the remnant pancreas.21 In this study, we used the annular array US probe intraoperatively. In an annular array transducer, the width of the US beam is approximately equal over the entire depth, and lateral resolution is improved, particularly in the near and far range.22 The 7.5-MHz annular array transducer provided longitudinal resolution of 250 µ and lateral resolution of 500 µ. Moreover, an acoustic coupler was integrated into the transducer, providing an excellent spatial resolution image by direct contact with the pancreas.23 Thus, it is suitable for longitudinal scanning of the MPD to detect intraductal mural nodules that reflect the intraductal spread of IPMT. In this study, we could determine the appropriate resection line in all cases. In terms of diagnostic accuracy, conventional intraoperative US was not as accurate as EUS because the lateral resolution of the former was inferior. Moreover, acoustic coupling of conventional intraoperative US was unstable in some cases. These factors reduced the diagnostic accuracy of conventional intraoperative US. The intraoperative annular array US could not detect mural nodules of the MPD in 3 cases (ie, 2 hyperplasias and 1 adenoma). In these 3 cases, the mural nodules were small (less than 2 mm) and benign. No malignant IPMT lesions were overlooked by the intraoperative annular array US. Multicentric IPMT lesions were another problem.24 Fujii et al25 reported that the incidence of IPMT was 32%. In our study, the incidence was 22% (6/27). Multicentric IPMT lesions were difficult to diagnose preoperatively by conventional imaging techniques. Failure to identify a multicentric lesion in the remnant pancreas leads to a postoperative recurrence of IPMT. Thus, the intraoperative survey of an IPMT lesion by annular array US was important to avoid recurrence. ERP has been reported to be useful for the diagnosis of IPMT.26,27 However, complete opacification of the pancreatic ducts and cystic tumors is often difficult because of the presence of abundant mucin and an enlarged papillary orifice. Aspiration of the mucin and use of a balloon catheter to avoid the reflux of contrast material into the duodenum can make it possible to differentiate between mural nodules and mucin in filling defects imaged by ERCP. But differentiation between intraductal mural nodules and mucin was difficult in some cases.

64 Kaneko et al

Basically, ERCP reflected the luminal profile of the pancreatic duct by contrast medium. Full visualization of the pancreatic duct by ERCP was sometimes difficult. Thus, determination of the resection line by ERCP mapping plus intraoperative frozen section margins depends on the operator’s experience. Intraoperative US with annular array probe guided the operator objectively to detect IPMT spread in the MPD, multifocal lesions, and intrapancreatic or extrapancreatic infiltration, all of which were sometimes overlooked by ERCP finding. CT scan and magnetic resonance imaging are noninvasive techniques used to detect dilatation of the MPD and branch duct,28-30 but their sensitivity in the detection of mural nodules was reported to be low.6,15 Thus, they do not provide accurate information on the intraductal extent of IPMT lesions. In contrast, EUS is reported to visualize cystic tumors and detect mural nodules more accurately than ERP.31 At present, CT and magnetic resonance imaging are screening modalities used to detect IPMT lesions. EUS is available for subsequent examination to determine operative indications and the extent of IPMT lesions. In preoperative diagnosis, EUS detected IPMT lesions accurately and gave satisfactory results. But the IPMT lesions visualized by EUS were difficult to confirm with the naked eye and with palpation by an operating surgeon. Consequently, during operation, the intraductal extent of IPMT lesions was confirmed by the intraoperative annular array US, and the resection line could be accurately determined. This is another reason to perform intraoperative annular array US in addition to EUS. Finally, intraoperative pancreatoscopy was performed to check the residual intraductal IPMT lesions. In our department, the abovementioned procedures are considered to be the best options for the management of IPMT. When mural nodules were not detected in a dilated MPD or branch duct by intraoperative annular array US, such lesions were not indicated for resection. According to this study, the mean diameter of mural nodules in benign hyperplasia was 2.9 mm against 5 mm in another study.15 These data lend support to the idea that mural nodules less than 3 mm are benign and safe without surgical intervention. Therefore, when mural nodules were detected unexpectedly during operation by annular array US, if the sizes were less than 3 mm, these nodules including multicentric IPMT lesions could be monitored to avoid total pancreatectomy. On the basis of our results, the percentage of hyperplasia less than 3 mm was 75%. The percentage of adenoma or carcinoma mural nodules more than 3 mm was

Surgery January 2001

100%. Consequently, if more than 3 mm, multicentric IPMT lesions should be resected because of their increased malignant potential.

REFERENCES 1. Ohhashi K, Murakami F, Takekoshi T. Four cases of “mucinproducing” cancer of the pancreas on specific findings of the papilla of Vater. Prog Dig Endosc 1982;20:348-51. 2. Itai Y, Ohhashi K, Nagai H, Murakami Y, Kokubo T, Makita K, et al. “Ductectatic” mucinous cystadenoma and cystadenocarcinoma of the pancreas. Radiology 1986;161:697-700. 3. Klöppel H, Solcia E, Longnecker DS, Capella C, Sobin LH. Histological typing of tumors of the exocrine pancreas. International histological classification of tumors. 2nd ed. Berlin: Springer; 1996. 4. Rickaert F, Cremer M, Deviere J, Tavares L, Lambilliotte JP, Schröder S, et al. Intraductal mucin-hypersecreting neoplasms of the pancreas. A clinicopathologic study of eight patients. Gastroenterology 1991;101:512-9. 5. Warshaw AL. Mucinous cystic tumors and mucinous ductal ectasia of the pancreas. Gastrointest Endosc 1991;37:199201. 6. Itoh S, Ishiguchi T, Ishigaki T, Sakuma S, Maruyama K, Senda K. Mucin-producing pancreatic tumor: CT finding and histopathologic correlation. Radiology 1992;183:81-6. 7. Calan LD, Levard H, Hennet H, Fingerhut A. Pancreatic cystadenoma and cystadenocarcinoma: diagnostic value of preoperative morphological investigations. Eur J Surg 1995;161:35-40. 8. Furukawa T, Tsukamoto Y, Naitoh Y, Mitake M, Hirooka Y, Hayakawa T. Differential diagnosis of pancreatic diseases with intraductal ultrasound system. Gastrointest Endosc 1994;40:213-9. 9. Kubota K, Noie T, Sano K, Abe H, Bandai Y, Makuuchi M. Impact of intraoperative ultrasonography on surgery for cystic lesions of the pancreas. World J Surg 1997;21:72-7. 10. Schwaighofer B, Stiglbauer R, Frühwald F, Tratting S. A new transducer technology the annular phased array: comparison with a conventional mechanically rotating sector transducer: the upper abdomen. Eur J Radiol 1987;7:285-7. 11. Kaneko T, Nakao A, Nomoto S, Takagi H. Two cases of mucin-producing cystic tumor of the pancreas: usefulness of intraoperative pancreatoscopy. Gastroenterol Endosc 1997;39:633. 12. Warshaw AL, Compton CC, Lewandrowski K, Cardenosa G, Mueller PR. Cystic tumors of the pancreas. New clinical, radiologic, and pathologic observations in 67 patients. Ann Surg 1990;212:432-45. 13. Shyr YM, Su CH, Tsay SH, Lui WY. Mucin-producing neoplasms of the pancreas: intraductal and mucinous cystic neoplasms. Ann Surg 1996;223:141-6. 14. Sugiyama M, Atomi Y, Kuroda A. Two types of mucin-producing cystic tumors of the pancreas: diagnosis and treatment. Surgery 997;122:617-25. 15. Yamaguchi K, Ogawa Y, Chijiiwa K, Tanaka M. Mucin-hypersecreting tumors of the pancreas: assessing the grade of malignancy preoperatively. Am J Surg 1996;171:427-31. 16. Rivera JA, Castillo CF, Pins M, Compton CC, Lewandroski KB, Ratter DW, et al. Pancreatic mucinous ductal ectasia and intraductal papillary neoplasms. A single malignant clinicopathologic entity. Ann Surg 1997;225:637-46.

Kaneko et al 65

Surgery Volume 129, Number 1 17. Morohoshi T, Kanda M, Asanuma K, Klöppel G. Intraductal papillary neoplasms of the pancreas. A clinicopathologic study of six patients. Cancer 1989;64:1329-35. 18. Bastid C, Sarles BH, Payan MJ, Sahel J. Mucinous ductal ectasia of the pancreas: a premalignant disease and a cause of obstructive pancreatitis. Pancreas 1991;6:15-22. 19. Yamada M, Kozuka S, Yamao K, Nakazawa S, Naitoh Y, Tsukamoto Y. Mucin-producing tumors of the pancreas. Cancer 1991; 68:159-68. 20. Cellier C, Cuillerier E, Palazzo L, Rickaert F, Flejou J-F, Napoleon B, et al. Intraductal papillary and mucinous tumors of the pancreas: accuracy of preoperative computed tomography, endoscopic retrograde pancreatography and endoscopic ultrasonography, and long-term outcome in a large surgical series. Gastrointest Endosc 1998;47:42-9. 21. Kaneko T, Nakao A, Nomoto S, Furukawa T, Hirooka Y, Nakashima N, et al. Intraoperative pancreatoscopy with ultrathin pancreatoscope for mucin-producing tumors of the pancreas. Arch Surg 1998;133:263-7. 22. Fishell EK, Foster FS, Connor T, Khodai M, Harasciewics K, Hunt JW. Clinical performance of a cone/annular array ultrasound breast scanner. Ultrasound Med Biol 1990;16: 361-74. 23. Kubota M. Study on ultrasonically guided fine-needle aspiration cytology for breast disorders with high-resolution transducer. Tokai J Exp Clin Med 1996;21:7-12. 24. Obara T, Saitoh Y, Maguchi H, Ura H, Kitazawa S, Koike Y, et al. Multicentric development of pancreatic intraductal

25.

26.

27.

28.

29.

30.

31.

carcinoma through atypical papillary hyperplasia. Hum Pathol 1992;23:82-5. Fujii T, Obara T, Maguchi H, Tanno S, Ura H, Kohgo Y. Clinicopathological study of multiple mucin-producing tumors of the pancreas: multicentric development of carcinoma through atypical hyperplasia. J Jpn Pancreas Soc 1996;198:249-57. Obara T, Maguchi H, Saitoh Y, Ura H, Koike Y, Kitazawa S, et al. Mucin-producing tumor of the pancreas: a unique clinical entity. Am J Gastroenterol 1991;86:1619-25. Ragiman I, Kortan P, Walden D, Kandel G, Marcon NE, Haber GB. Mucinous ductal ectasia: cholangiopancreatographic and endoscopic findings. Endoscopy 1994;26:303-7. Itai Y, Kokubo T, Atomi Y, Kuroda A, Haraguchi Y, Terano A. Mucin-hypersecreting carcinoma of the pancreas. Radiology 1987;165:51-5. Procacci C, Graziani R, Bicego E, Bergamo-Andreis IA, Mainardi P, Zamboni G, et al. Intraductal mucin-producing tumors of the pancreas: imaging findings. Radiology 1996;198:249-57. Irie H, Honda H, Tajima T, Kuroiwa T, Yoshimitsu K, Makisumi K, et al. Optimal MR cholangiopancreatographic sequence and its clinical application. Radiology 1998;206: 379-87. Sugiyama M, Atomi Y, Saitoh M. Intraductal papillary tumors of the pancreas: evaluation with endoscopic ultrasonography. Gastrointest Endosc 1998;48:164-71.

Surgery is abstracted and/or indexed in Index Medicus, Science Citation Index, Current Contents/ Clinical Medicine, Current Contents/Life Sciences, and MEDLINE. This Journal has been registered with Copyright Clearance Center, Inc, 222 Rosewood Dr, Danvers, MA 01923. Consent is given for the copying of articles for personal or internal use of specific clients. This consent is given on the condition that the copier pay directly to the Center the per-copy fee stated on the first page of each article for copying beyond that permitted by US Copyright Law. This consent does not extend to other kinds of copying, such as for general distribution, resale, advertising and promotional purposes, or for creating new collective works. All inquiries regarding copyrighted material from this publication other than those that can be handled through Copyright Clearance Center should be directed to Journals Permission Department, Mosby, Inc, 6277 Sea Harbor Dr, Orlando, FL 32887; Fax (407) 345-4058.