- Email: [email protected]
Feasibility and safety of microforceps biopsy in the diagnosis of pancreatic cysts
ORIGINAL ARTICLE
Feasibility and safety of microforceps biopsy in the diagnosis of pancreatic cysts Omer Basar, MD,1 Osman Yuksel, MD,2 Dennis J. Yang, MD,3 Jason Samarasena, MD,4 David Forcione, MD,1 Christopher J. DiMaio, MD,5 Mihir S. Wagh, MD,6 Kenneth Chang, MD,4 Brenna Casey, MD,1 Carlos Fernandez-del Castillo, MD,7 Martha B. Pitman, MD,8,* William R. Brugge, MD1,* Boston, Massachusetts, USA; Ankara, Turkey; Gainesville, Florida; Aurora, California; New York, New York; Denver, Colorado, USA
Background and Aims: The tissue acquisition and diagnostic yield of cyst fluid cytology is low-to-moderate and rarely provides a specific diagnosis. The aim of this study was to compare the tissue acquisition and diagnostic tissue yield of microforceps biopsy (MFB) with cyst fluid cytology. Methods: In this multicenter study, data of 42 patients who had cysts both aspirated by EUS-guided FNA (EUS-FNA) and biopsy specimens were then obtained with an MFB device, were collected. Cytology analysis of cyst fluid and histologic analysis of biopsy specimens were done. Acquisition yield was defined as percentage of patients with tissue present in the aspirate or biopsy. Diagnostic tissue yield was evaluated at 3 levels: the ability of differentiation between mucinous and/or nonmucinous cysts, detection of high risk for malignancy, and specific cyst type diagnosis. Results: The mean patient age was 69 years. Sixteen pancreatic cysts (38.1%) were located in the head, 17 (40.5%) in the body, and 9 (21.4%) in the tail. The mean cyst size was 28.2 mm (12-60 mm); 25 of 42 (60%) were septated. The EUS-FNA tissue (fluid) acquisition yield was 88.1% (37/42). The MFB tissue acquisition yield was 90.4% (38/42). The diagnostic cytology yield to differentiate between mucinous and/or nonmucinous cysts was 47.6% (20/42), and the MFB histologic yield to differentiate between mucinous and/or nonmucinous cysts was 61.9% (26/42) (P Z .188). The percentage of cysts at high risk for malignancy by cytology was 54.7% (23/42), and MFB was 71.5% (30/42) (P Z .113). However, the ability of MFB to provide a specific cyst type diagnosis was 35.7% (15/42), and that for cytology was 4.8% (2/42) (P Z .001). Surgical histology was concordant with that of MFB in 6 of 7 patients (85%), and with that of cytology in 1 of 7 patients (15%). Conclusion: The cyst tissue acquisition yield for MFBs was 90%. Although cytology of cyst fluid and MFB were comparable in distinguishing mucinous and nonmucinous cysts and detecting cysts at high risk for malignancy, MFB was far superior to cytology for providing a specific cyst diagnosis. (Gastrointest Endosc 2018;-:1-8.)
Abbreviations: CEA, carcinoembryonic antigen; cPNET, cystic pancreatic neuroendocrine tumor; EUS-FNA, EUS-guided FNA; IPMN, intraductal papillary neoplasm; MCN, mucinous cystic neoplasm; MFB, microforceps biopsy. DISCLOSURE: O. Basar has received a grant from U.S. Endoscopy Group, Inc. D. Yang is a consultant for US Endoscopy. J. Samarasena is a consultant for US Endoscopy. C. DiMaio is a consultant and speaker for Boston Scientific and Medtronic. M. Wagh is a consultant for Boston Scientific and Medtronic. K. Chang is a consultant for Olympus and Cook Medical and a speaker for Medtronic. W. Brugge does research for Boston Scientific. All other authors disclosed no financial relationships relevant to this publication. *Authors Pitman and Brugge contributed equally to the article. Copyright ª 2018 by the American Society for Gastrointestinal Endoscopy
www.giejournal.org
0016-5107/$36.00 https://doi.org/10.1016/j.gie.2018.02.039 Received September 22, 2017. Accepted February 20, 2018. Current affiliations: Pancreas Biliary Center, Gastrointestinal Unit, Massachusetts General Hospital, Boston, Massachusetts (1), Department of Gastroenterology, Eryaman Private Hospital, Ankara, Turkey (2), Division of Gastroenterology, University of Florida Health, Gainesville, Florida (3), Division of Gastroenterology and Hepatology, University of California, Aurora, California (4), Division of Gastroenterology, Mount Sinai Medical Center, New York, New York (5), Division of Gastroenterology and Hepatology, University of Colorado, Denver, Colorado (6), Department of General and Gastrointestinal Surgery; (7), Department of Pathology, Massachusetts General Hospital, Boston, Massachusetts (8). Reprint requests: Omer Basar, MD, 3-H GI Associates, Zero Emerson Place, Blossom St, Massachusetts General Hospital, Boston, MA 02114.
Volume
-,
No.
-
: 2018 GASTROINTESTINAL ENDOSCOPY 1
Value of microforceps biopsy for pancreatic cysts
Basar et al
Primary pancreatic cysts are broadly divided into nonneoplastic and neoplastic, the latter termed pancreatic cystic neoplasms. Pancreatic cystic neoplasms are further classified into nonmucinous neoplasms (serous cystadenomas), mucinous neoplasms (intraductal papillary neoplasms [IPMN]), and mucinous cystic neoplasms [MCN]). In addition to primary pancreatic cysts, solid tumors can become cystic secondarily from degeneration, including pancreatic ductal adenocarcinoma and cystic pancreatic neuroendocrine tumors (cPNETs). The risk of malignancy in pancreatic cysts usually dictates the need for surgical resection.1,2 The most commonly used tools for both the diagnosis and differentiation of pancreatic cysts are cross-sectional imaging and EUS.3,4 Because imaging alone has some limitations, cyst fluid analysis obtained by EUS-guided FNA (EUS-FNA) aids in further evaluation.4-6 Cytology is limited because of scant cellularity and difficulty in detecting thin, watery, extracellular mucin, but cytology has been shown to be highly accurate for the diagnosis of a cyst at high risk for malignancy.7-10 However, cytology is dependent on cells being shed into the cyst fluid for analysis. Interpretation challenges arise from GI contamination, degenerative changes of the cells, heterogeneity of the cyst lining epithelium, and lack of experience and interpretive expertise.11 Recently, a U.S. Food and Drug Administration– approved single-use Moray microforceps biopsy device (U.S. Endoscopy, Mentor, Ohio) (Fig. 1) has been designed for use in EUS procedures to enable sampling from cysts that can be accessed with a 19-gauge EUS-FNA needle. The aim of this study was to compare the tissue acquisition and diagnostic yield of the microforceps biopsy (MFB) with cyst fluid cytology.
PATIENTS AND METHODS This study is a retrospective, controlled, open label, and multicenter investigation including Massachusetts General Hospital, University of Florida Health, University of California at Irvine, Mount Sinai Hospital, and University of Colorado. The patients gave their consent for EUS-FNA, cyst fluid analysis, and for use of the MFB device in the evaluation of pancreatic cysts. The institutional review boards of each center approved the study. The data of all patients (initial and subsequent) undergoing MFBs between 2015 and 2016 were collected, without a predefined study protocol. Exclusion criteria included patients with a bleeding diathesis, women with known pregnancy, patients with a history of pancreatic cancer, patients with acute pancreatitis or a high clinical suspicion of a pseudocyst or abscess, patients with a solid pancreatic mass or a clinically suspected pancreatic adenocarcinoma with cystic degeneration, and patients with cysts of extra-pancreatic origin. 2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2018
Figure 1. Moray microforceps. Permission granted by US Endoscopy.
EUS-FNA and pancreatic cyst biopsy In this multicenter study, data of 42 patients who had cysts aspirated by EUS-FNA were collected, and biopsy specimens were then obtained with an MFB device. Before the start of the procedure, all patients were given broad-spectrum antibiotics. The number, location, size of the cyst, presence or absence of septations, mural nodule and an adjacent mass on EUS were recorded. After evaluating the cyst with EUS, a 19-gauge Flex EUS needle (Boston Scientific, Marlborough, Mass) was placed into the cyst cavity, and the cyst fluid was aspirated first. Without removing the needle, an MFB device was introduced through the needle (Fig. 2), and pinch biopsy specimens were obtained from the cyst wall, septations, nodules, or adjacent masses if present (Fig. 3). The standardized order of obtaining biopsy specimens was the adjacent mass, mural nodule, cyst wall, and septations, respectively. If the biopsy specimen seemed insufficient in size for histology, additional passes were made. Aspirated cyst fluid was sent for CEA evaluation and for cytology analysis by using a cytospin preparation stained with routine Papanicolaou stain. The MFB specimens were sent in formalin and processed as a routine histology specimen. Adverse events including intra-cystic bleeding, pain, or pancreatitis related to the procedure were assessed during the procedure and for 2 hours after the procedure before discharge.
Cytology and histology evaluation Failure to obtain fluid by EUS-FNA or tissue by MFB were considered acquisition failures and were classified as nondiagnostic. The term tissue acquisition yield referred to the ability to collect fluid or tissue for analysis and was calculated as the number of patients with fluid or tissue obtained by aspiration or MFB, which was either diagnostic or nondiagnostic, divided by the total number of patients. The diagnostic tissue yield was evaluated at 3 levels (defined below) and calculated as the number of cases www.giejournal.org
Basar et al
Value of microforceps biopsy for pancreatic cysts
Figure 2. EUS image of microbiopsy forceps wide open inside a pancreatic cyst.
that were either diagnostic or nondiagnostic divided by the total number of patients. The nondiagnostic cases were all the cases that were nondiagnostic by cytology and histology plus the acquisition failure cases. Level I analysis referred to the ability to distinguish a mucinous from a nonmucinous cyst. Level II analysis referred to the ability to detect a cyst at high risk for malignancy (high-grade dysplasia or adenocarcinoma). Level III analysis referred to the ability to make a specific diagnosis of the pancreatic cyst (such as an IPMN, serous cystadenoma, cPNET, or MCN). On cytology evaluation, a cyst was diagnostic of a mucinous cyst if there was extracellular mucin or mucinous epithelium considered lesional; otherwise, the cyst was classified as nonmucinous.12 Transgastric FNAs with nonspecific gastric foveolar type epithelium without thick extracellular mucin were defined as nondiagnostic. Pancreatic cyst fluid CEA >192 ng/mL was considered diagnostic for a mucinous cyst. Cytology was diagnostic of a high-risk cyst if there were cells diagnostic of a malignant cyst or a cyst at high-risk for malignancy such as a mucinous cyst with high-grade epithelial dysplasia12 or a cystic neuroendocrine tumor.13 The histology evaluation of the MFB tissue followed standard histology definitions for epithelial type: mucinous epithelium with cytoplasmic mucin visible on routine hematoxylin and eosin stain (Figs. 4 and 5), serous epithelium (nonmucinous cuboidal epithelium with or without confirmation with special stains to support cytoplasmic glycogen) (Fig. 6), acinar epithelium (granular nonmucinous cytoplasm with confirmation of exocrine granules with ancillary testing) (Fig. 7) and endocrine epithelium (with or without confirmation with ancillary tests) (Fig. 8). Histology was diagnostic of a high-risk cyst with the diagnosis of a mucinous cyst with high-grade dysplasia or invasive adenocarcinoma or neuroendocrine tumor. The presence of subepithelial ovarian type stroma defined an MCN, and the absence of such stroma defined an IPMN. Fibrous tissue without epithelium www.giejournal.org
Figure 3. Microbiopsy forceps after obtaining a pinch biopsy specimen from a pancreatic cyst.
was a tissue acquisition success, but nondiagnostic for cyst type. In patients who underwent pancreatic surgery in a 1-year follow-up period, surgical histology results were used as a criterion standard to determine the specificity and accuracy of MFB.
Data analysis The Statistical Package for Social Sciences for Windows version 16.0 (SPSS Inc, Chicago, Ill, USA) was used to analyze the data. A chi-square test was used for comparison of categorical variables. The value of P < .05 was considered statistically significant.
RESULTS Data regarding 42 patients who had EUS-FNA and MFB (23 female and 19 male) were entered over a 1-year period into the study from multiple centers for data analysis. The mean age of the patients was 69 years (range 27-91). Cysts were located in the head of the pancreas in 16 patients (38.1%), in the body in 17 (40.5%), and in the tail in 9 (21.4%). The mean size of the cysts was 28.2 mm (range 12-60 mm), with 25 of 42 (60%) cysts septated and 17 of 42 (40%) unilocular. A mural nodule was seen in 4 patients (9.6%), and cysts and an adjacent mass were observed in 1 (2.4%).
Tissue acquisition yield Cyst fluid was obtained from 37 patients; the EUS-FNA fluid acquisition yield was 88.1% (37/42). The microforceps tissue acquisition yield was 90.4% (38/42). The pinch biopsy specimen was not sufficient for histology Volume
-,
No.
-
: 2018 GASTROINTESTINAL ENDOSCOPY 3
Value of microforceps biopsy for pancreatic cysts
Basar et al
Figure 4. Microbiopsy forceps biopsy specimen of an intraductal papillary mucinous neoplasm with low-grade dysplasia. Note the absence of subepithelial ovarian-type stroma (H&E, orig. mag. 200).
Figure 6. Microbiopsy forceps biopsy specimen of a serous cystadenoma. Note that a nonmucinous cuboidal epithelium lines the fibrous septum (H&E, orig. mag. 400).
Diagnostic tissue yield
Figure 5. Papillary epithelium with high-grade dysplasia consistent with an intraductal papillary mucinous neoplasm with high-grade dysplasia. The resected tumor demonstrated a microinvasive adenocarcinoma (H&E, orig. mag. 200).
processing in 4 of 42 patients (9.6%) (2 from a cyst wall and 2 from a septum). Overall, 30 of 42 successful MFB specimens (71.4%) were obtained from the cyst wall, 6 of 42 (14.4%) from septations, 1 of 42 (2.4%) from an adjacent mass, and 1 of 42 (2.4%) from a mural nodule. The microforceps tissue acquisition yield was not significantly different between the groups according to the presence or absence of septations or size and location of cysts (P Z .122; P Z .761, and P Z .176; respectively) (Table 1). During the procedure and before discharge, no serious adverse events were observed. One patient had mild abdominal pain, which lasted for an hour after the procedure. Another patient had self-limited, intracystic bleeding, which did not progress and did not result in any symptoms. 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2018
Level I analysis (diagnostic ability to distinguish a mucinous from nonmucinous cyst). The level I diagnostic tissue yield by using cyst fluid cytology was 47.6% (20/42) and was not statistically significantly different from the MFB histology diagnostic yield of 61.9% (26/42) (P Z .188) (Table 2). Excluding acquisition failures, 45.9% of cases (17/37) were nondiagnostic by cyst fluid cytology, and 31.5% of cases (12/38) were nondiagnostic by MFB (Table 3). In addition, 12 of 37 cysts (32.4%) had a CEA level >192 ng/mL, when acquisition failures were excluded. The average CEA among mucinous cysts was 893.1 ng/mL, as compared with 47.4 ng/mL among nonmucinous cysts (P Z .003). Aspirated cyst fluid CEA levels >192 ng/mL were found in 3 patients in the EUSFNA group (with only cyst cytology results) and in 4 patients in the MFB group (with only cyst histology results) whose results were nondiagnostic according to Level I analysis. Level II analysis (risk of malignancy). The level II diagnostic tissue yield by using cyst fluid cytology of 54.7.0% (23/42) was not statistically significantly different when compared with MFB of 71.5% (30/42) (P Z .113) (Table 2). Cytology was diagnostic for low-grade dysplastic epithelium in 50.0% of patients (21/42), and MFB was diagnostic for low-grade dysplastic epithelium in 64.2% of patients (27/42). Cytology diagnosed high-grade dysplastic epithelium in 2 patients (4.8%), and MFB diagnosed high-grade dysplastic epithelium in 3 patients (7.1%). When we excluded the acquisition failures, the cyst fluid did not contain a valuable epithelium for grading dysplasia in 37.8% of patients (14/37), and the MFB specimen did not contain epithelium for grading in 21.0% of patients (8/38) (Table 3). www.giejournal.org
Basar et al
Value of microforceps biopsy for pancreatic cysts
Figure 7. Microbiopsy forceps biopsy of an acinar cell cystadenoma A, with immunohistochemical stain for trypsin B, supporting the diagnosis (H&E, orig. mag. 400).
TABLE 1. Tissue (diagnostic and nondiagnostic) acquisition yield relative to cyst features by microforceps biopsy Tissue acquisition yield
P value
Unilocular (17)
17/17 (100%)
.122
Septated (25)
21/25 (84%)
Morphology of cyst
Size of cyst <30 mm (24)
22/24 (92%)
30 mm (18)
16/18 (89%)
.761
Location of cyst Head (16)
16/16 (100%)
Body (17)
15/17 (88%)
Tail (9) Figure 8. Small, uniform polygonal cells with coarse chromatin typical of solid well-differentiated neuroendocrine tumors is seen in this forceps biopsy of a cystic well-differentiated neuroendocrine tumor (H&E, orig. mag. 200).
Level III analysis (specific cyst diagnosis). There was a significant difference in the level III diagnostic tissue yield between cyst fluid cytology (2/42) and MFB histology (15/42) (4.8% and 35.7%; P Z .001; respectively), (Table 2). The 15 patients who had a specific histology cyst diagnosis by MFB compared with cytology are shown in Table 4.
Validation of MFB and cytology In a 1-year follow-up period, 17% of patients (7/42) underwent pancreatic surgery. Specific MFB results were concordant with surgical histology in 6 of 7 patients (85%). There was concordance between the specific cytology diagnosis and surgical histology in 1 of 7 (15%) patients (Table 5).
DISCUSSION We show that tissue obtained by MFB adds valuable, diagnostic tissue in the EUS-FNA evaluation of pancreatic www.giejournal.org
.176
7/9 (78%)
cysts. In our study, the MFB tissue acquisition yield was 90%. MFB was found to be superior to cytology for providing a specific cyst diagnosis and was comparable to cytology in detecting a mucinous cyst and a high-risk cyst. The use of the disposable microforceps was not associated with any significant adverse events other than intracystic bleeding in 1 patient. The goal of the management of patients with pancreatic cysts is to accurately define and resect cysts before progression of malignancy, while avoiding unnecessary follow-up procedures and surgery. Because it is known from other types of tissues that dysplasia does not always progress to cancer,14,15 and the natural history of pancreatic cysts is not completely clear, the accurate diagnosis and follow-up criteria for pancreatic cysts is still a challenging issue. The premalignant cysts of the pancreas such as mucinous cysts (including IPMNs and MCNs) can be cured before developing an invasive adenocarcinoma. However, not all patients with pancreatic cystic neoplasms require surgery and not all the patients are good candidates for surgery. Not all the premalignant mucinous cysts are at high risk of malignancy, and surgery is not the best Volume
-,
No.
-
: 2018 GASTROINTESTINAL ENDOSCOPY 5
Value of microforceps biopsy for pancreatic cysts
Basar et al
TABLE 2. Diagnostic tissue yield of EUS-guided FNA cyst fluid cytology and of microforceps biopsy histology EUS-guided FNA (cytology) 42 patients
Microforceps biopsy (histology) 42 patients
Diagnostic
20 (47.6%)
26 (61.9%)
Nondiagnostic*
22 (52.4%)
16 (38.1%)
Diagnostic
23 (54.7%)
30 (71.5%)
Nondiagnostic*
19 (45.3%)
12 (28.5%)
Diagnosis
P value
Level I
.188
Level II
.113
Level III
.001
Diagnostic Nondiagnostic*
2 (4.8%)
15 (35.7%)
40 (95.2%)
27 (64.3%)
Level I–diagnostic ability to differentiate a mucinous from nonmucinous cyst. Level II–diagnostic of degree of dysplasia. Level III–diagnostic of specific type of cyst. *Nondiagnostic: the combinations of the cases that were nondiagnostic by cytology and histology plus the acquisition failure cases.
TABLE 3. Number of patients according to diagnostic tissue yields
Diagnosis Mucinous vs nonmucinous
TABLE 4. The results of cytology (by EUS-guided FNA) in 15 patients who had a specific histology cyst diagnosis (by MFB)
EUS-guided FNA (cytology) 42 patients
Microforceps biopsy (histology) 42 patients
20 (47.6%)
26 (61.9%)
6 IPMNs
6/6 Nondiagnostic 3/3 Nondiagnostic
Specific histology diagnosis by MFB
Specific cytology diagnosis by EUS-guided FNA
Nondiagnostic
17 (40.5%)
12 (28.5%)
3 SCNs
Acquisition failures*
5 (11.9%)
4 (9.6%)
2 MCNs
2/2 Nondiagnostic
2 PDACs
1/2 PDACs
Low-grade (risk) epithelium
21 (50.0%)
27 (64.2%)
High-grade (risk) epithelium
2 (4.8%)
3 (7.1%)
Nondiagnostic
14 (33.3%)
8 (19.1%)
Acquisition failures*
5 (11.9%)
4 (9.6%)
Specific diagnosis present
2 (4.8%)
15 (35.7%)
No specific diagnosis
35 (83.3%)
23 (54.7%)
Acquisition failures*
5 (11.9%)
4 (9.6%)
*The number of patients from whom cyst fluid was not obtained and the microforceps biopsy did not procure sufficient tissue for histology.
choice in elderly patients with comorbid conditions in which conservative follow-up would be a better option. For example, most patients with branch duct–IPMNs.16 In addition, distinguishing between a branch duct–IPMN and MCN is important because all MCNs are resected regardless of grade because of the ease of the usual distal pancreatectomy compared with the expense and anxiety of life-long surveillance for progression to malignancy in the typical middle-aged woman.17 Therefore, determining cyst type as well as the malignant potential of a cyst dictates the best treatment strategy while avoiding unnecessary surgery. A multidisciplinary approach with clinical, radiologic, and cytologic information including biochemical and genetic 6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2018
1/2 Nondiagnostic (false positive low-risk mucinous cyst) 1 cPNET
1/1 Nondiagnostic (false positive high-risk mucinous cyst)
1 Acinar cystadenoma
1/1 Nondiagnostic (false positive low-risk mucinous cyst)
MFB, Microforceps biopsy; IPMN, intraductal papillary neoplasm; SCN, serous cystic neoplasm; MCN, mucinous cystic neoplasm; PDAC, pancreatic ductal adenocarcinoma; cPNET, cystic pancreatic neuroendocrine tumor.
analysis is generally recommended to diagnose a pancreatic cyst.17 Cyst fluid can be aspirated during EUS-FNA and analyzed for biochemical testing including amylase and CEA levels, cytology analysis and genetic mutational analysis.18 The accuracy of EUS-FNA cytology is very good with a very low false-positive rate but a relatively high falsenegative rate.19 In clinical practice, a high cyst fluid CEA level with the presence of mucin on cytology is the most common approach to differentiating a mucinous cyst from a nonmucinous cyst. Although cytology can detect a high-risk cyst by diagnosing a cPNET, high-grade mucinous dysplasia or invasive carcinoma with over 90% specificity, most cysts evaluated are not high risk, and the specificity of diagnosing a mucinous cyst in general is only 50%.7-10 Furthermore, cytology alone is insufficient for diagnosis in up to 60% of patients because of the absence of epithelial www.giejournal.org
Basar et al
Value of microforceps biopsy for pancreatic cysts
TABLE 5. Comparison of the diagnostic results of 7 patients who underwent pancreatic surgery Specific cytology diagnosis by EUS-guided FNA Nondiagnostic* Nondiagnostic* (mucinous cyst with LGD) PDACz
Specific histology diagnosis by MFB
Surgical histology diagnosis
Serous cystadenomay
Serous cystadenoma
PDAC arising from mucinous cyst with HGDy
PDAC arising from IPMN
PDAC arising from IPMN with HGDy
PDAC arising from IPMN with HGD
Nondiagnostic*
IPMN with LGDy
IPMN with LGD
Nondiagnostic*
Nondiagnostic*
MCN with LGD
Nondiagnostic* (mucinous cyst with LGD)
Acinar cystadenomay
Acinar cystadenoma
Nondiagnostic* (mucinous cyst with HGD)
cPNETy
cPNET
MFB, Microforceps biopsy; LGD, low-grade dysplasia; PDAC, pancreatic ductal adenocarcinoma; HGD, high-grade dysplasia; IPMN, intraductal papillary neoplasm; MCN, mucinous cystic neoplasm; cPNET, cystic pancreatic neuroendocrine tumor. *Nondiagnostic: the nondiagnostic cases by cytology and histology or the acquisition failure cases. ySpecific MFB histology was concordant with surgical histology. zSpecific cytology was concordant with surgical histology.
cells in cyst fluid, the presence of GI contamination, the heterogeneity of mucinous cyst dysplasia, and the lack of experience and expertise in cytology interpretation.7-10 In addition, the fragile cells of serous cystadenoma rarely survive processing for diagnosis,20 and the subepithelial ovarian-type stroma required for the diagnosis of MCN is not present in aspirated cyst fluid.21 Although EUS-FNA cyst fluid cytology has value in the differentiation between mucinous/non-mucinous cysts and accurately diagnosing high-risk cysts, it cannot usually specifically define the type of cyst. MFB provides tissue from the cyst wall, septations, and mural nodules for histology thus providing tissue for ancillary studies such as immunohistochemistry. The subepithelial cyst wall tissue can be biopsied by micro-forceps to show the ovarian type stroma, which would further differentiate a MCN from an IPMN.22 Furthermore, MFB can provide tissue for ancillary studies to confirm the diagnosis of serous cystadenoma, cPNET, and acinar cell cystadenoma.23 Although statistically not significant, the MFB histology increased the number of mucinous/nonmucinous diagnostic cyst results in a moderate proportion of patients (20-26 of 42) in our study. On the other hand, cytology alone was diagnostic of degree of dysplasia in 23 of 42 patients (54.7%) (21 were low-grade epithelium, 2 were high-grade epithelium). Furthermore, the pinch biopsy specimens were diagnostic of the degree of dysplasia in 30 of 42 patients (71.5%), with only MFB (27 low-grade epithelium, 3 high-grade epithelium). The MFB histology again increased the detection of low-grade epithelium in a moderate proportion of patients (from 21 patients to 27 patients out of 42 patients). Cyst fluid CEA (CEA >192 ng/mL) has been shown to be one of the most accurate methods for identifying a mucinous cyst, whereas CEA alone is neither a marker to differentiate a malignant pancreatic cyst from benign nor to differentiate the www.giejournal.org
subgroups of mucinous cysts.7,24 In our study, 32.4% of the aspirated cysts had CEA levels >192 ng/mL, supporting a mucinous cyst. Interestingly, CEA levels >192 ng/mL were detected in 3 patients whose EUS-FNA could not differentiate between a mucinous and a nonmucinous cyst, and in 4 patients whose MFB histology could not differentiate between a mucinous and a nonmucinous cyst. Although CEA lacks the ability to identify the degree of dysplasia and the specific cyst type, adding cyst fluid CEA analysis to MFB may help differentiate mucinous from nonmucinous cysts. Most importantly, we demonstrated that the diagnostic yield of the MFB for a specific type of cyst was significantly higher than the diagnostic cytology yield. Moreover, when specific diagnostic results with MFB histology were compared with cytology results, none of the 6 patients with IPMN, none of the 3 patients with serous cystadenoma, and none of the 2 patients with MCN could be diagnosed. Additionally, a patient with a cPNET was misdiagnosed as having a high-risk mucinous cyst, and a patient with pancreatic ductal adenocarcinoma and another patient with acinar cystadenoma were misdiagnosed as having low-risk mucinous cysts. However, it has to be kept in mind that the overall ability to provide a specific diagnosis with only MFB was merely 36%. Moreover, although most of the patients enrolled into the study did not require pancreatic surgery over a 1-year follow-up period, specific MFB diagnoses were far more concordant, compared with a specific cytology diagnosis with surgical histology diagnoses. In conclusion, our study showed that the MFB was not affected by the size, location, or septations of the cyst and provided a specific diagnosis of the subtype of pancreatic cyst significantly better than cytology. In addition, MFB was slightly better than cytology in differentiating between mucinous and nonmucinous cysts and detecting a high-risk cyst. The main limitation of our study was that our outcomes were based on a relatively small sample size in a retrospective Volume
-,
No.
-
: 2018 GASTROINTESTINAL ENDOSCOPY 7
Value of microforceps biopsy for pancreatic cysts
Basar et al
collection of data and were without a criterion standard diagnosis in the majority of patients. This study likely represents the early experience with this novel tissue collection device. Nevertheless, our findings demonstrated that an EUS-guided MFB provided diagnostic tissue from a pancreatic cyst at a rate superior to cyst fluid cytology.
REFERENCES 1. Yoon WJ, Brugge WR. Pancreatic cystic neoplasms: diagnosis and management. Gastroenterol Clin North Am 2012;41:103-18. 2. Brugge WR. Diagnosis and management of cystic lesions of the pancreas. J Gastrointest Oncol 2015;6:375-88. 3. Mortelé KJ. Cystic pancreatic neoplasms: imaging features and management strategy. Semin Roentgenol 2013;48:253-63. 4. Brugge WR. The use of EUS to diagnose cystic neoplasms of the pancreas. Gastrointest Endosc 2009;69:S203-9. 5. Yoon WJ, Brugge WR. Endoscopic ultrasound and pancreatic cystic lesions-diagnostic and therapeutic applications. Endosc Ultrasound 2012;1:75-9. 6. Brugge WR, Lewandrowski K, Lee-Lewandrowski E, et al. Diagnosis of pancreatic cystic neoplasms: a report of the cooperative pancreatic cyst study. Gastroenterology 2004;126:1330-6. 7. Cizginer S, Turner BG, Bilge AR, et al. Cyst fluid carcinoembryonic antigen is an accurate diagnostic marker of pancreatic mucinous cysts [published correction appears in Pancreas. 2013;42:728]. Pancreas 2011;40:1024-8. 8. Zhan XB, Wang B, Liu F, et al. Cyst fluid carcino embryonic antigen concentration and cytology by endosonography-guided fine needle aspiration in predicting malignant pancreatic mucinous cystic neoplasms. J Dig Dis 2013;14:191-5. 9. Pitman MB, Genevay M, Yaeger K, et al. High-grade atypical epithelial cells in pancreatic mucinous cysts are a more accurate predictor of malignancy than “positive” cytology. Cancer Cytopathol 2010;118: 434-40. 10. Springer S, Wang Y, Dal Molin M, et al. A combination of molecular markers and clinical features improve the classification of pancreatic cysts. Gastroenterology 2015;149:1501-10.
8 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2018
11. Pitman MB, Lewandrowski K, Shen J, et al. Pancreatic cysts: preoperative diagnosis and clinical management. Cancer Cytopathol 2010;118: 1-13. 12. Pitman MB, Centeno BA, Daglilar ES, et al. Cytological criteria of highgrade epithelial atypia in the cyst fluid of pancreatic intraductal papillary mucinous neoplasms. Cancer Cytopathol 2014;122:40-7. 13. Morales-Oyarvide V, Yoon WJ, Ingkakul T, et al. Cystic pancreatic neuroendocrine tumors: the value of cytology in pre-operative diagnosis. Cancer Cytopathology 2014;122:435-44. 14. Wani S, Puli SR, Shaheen NJ, et al. Esophageal adenocarcinoma in Barrett’s esophagus after endoscopic ablative therapy: a meta-analysis and systematic review. Am J Gastroenterol 2009;104:502-13. 15. Ostör AG. Natural history of cervical intraepithelial neoplasia: a critical review. Int J Gynecol Pathol 1993;12:186-92. 16. Schmidt CM. Pancreatic cyst cytology: optimization of cancer risk profiling. Cancer Cytopathol 2013;121:2-3. 17. Pitman MB, Brugge WR, Warshaw AL. The value of cyst fluid analysis in the pre-operative evaluation of pancreatic cysts. J Gastrointest Oncol 2011;2:195-8. 18. Brugge WR, De Witt J, Klapman JB, et al. Techniques for cytologic sampling of pancreatic and bile duct lesions: The Papanicolaou Society of Cytopathology Guidelines. Cytojournal 2014;11:2. 19. Woolf KM, Liang H, Sletten ZJ, et al. False-negative rate of endoscopic ultrasound-guided fine-needle aspiration for pancreatic solid and cystic lesions with matched surgical resections as the gold standard: one institution’s experience. Cancer Cytopathol 2013;121:449-58. 20. Belsley NA, Pitman MB, Lauwers GY, et al. Serous cystadenomas of the pancreas: diagnostic performance of imaging and fine needle aspiration biopsy. Cancer Cytopathol 2008;114:102-10. 21. Centeno BA, Stelow E, Pitman MB. Pancreatic cytohistology. Cytohistology of small tissue samples. Published in association with the Papanicolaou Society of Cytopathology. KR Geisinger, MB Pitman, series editors. Cambridge, UK: Cambridge University Press; 2015. 22. Adsay NV. Cystic neoplasia of the pancreas: pathology and biology. J Gastrointest Surg 2008;12:401-4. 23. Chen AL, Misdraji J, Brugge WR, et al. Acinar cell cystadenoma: a challenging cytology diagnosis, facilitated by moray() micro-forceps biopsy. Diagn Cytopathol 2017;45:557-60. 24. Scourtas A, Dudley JC, Brugge WR, et al. Preoperative characteristics and cytological features of 136 histologically confirmed pancreatic mucinous cystic neoplasms. Cancer 2017;125:169-77.
www.giejournal.org
Feasibility and safety of microforceps biopsy in the diagnosis of pancreatic cysts Omer Basar, MD,1 Osman Yuksel, MD,2 Dennis J. Yang, MD,3 Jason Samarasena, MD,4 David Forcione, MD,1 Christopher J. DiMaio, MD,5 Mihir S. Wagh, MD,6 Kenneth Chang, MD,4 Brenna Casey, MD,1 Carlos Fernandez-del Castillo, MD,7 Martha B. Pitman, MD,8,* William R. Brugge, MD1,* Boston, Massachusetts, USA; Ankara, Turkey; Gainesville, Florida; Aurora, California; New York, New York; Denver, Colorado, USA
Background and Aims: The tissue acquisition and diagnostic yield of cyst fluid cytology is low-to-moderate and rarely provides a specific diagnosis. The aim of this study was to compare the tissue acquisition and diagnostic tissue yield of microforceps biopsy (MFB) with cyst fluid cytology. Methods: In this multicenter study, data of 42 patients who had cysts both aspirated by EUS-guided FNA (EUS-FNA) and biopsy specimens were then obtained with an MFB device, were collected. Cytology analysis of cyst fluid and histologic analysis of biopsy specimens were done. Acquisition yield was defined as percentage of patients with tissue present in the aspirate or biopsy. Diagnostic tissue yield was evaluated at 3 levels: the ability of differentiation between mucinous and/or nonmucinous cysts, detection of high risk for malignancy, and specific cyst type diagnosis. Results: The mean patient age was 69 years. Sixteen pancreatic cysts (38.1%) were located in the head, 17 (40.5%) in the body, and 9 (21.4%) in the tail. The mean cyst size was 28.2 mm (12-60 mm); 25 of 42 (60%) were septated. The EUS-FNA tissue (fluid) acquisition yield was 88.1% (37/42). The MFB tissue acquisition yield was 90.4% (38/42). The diagnostic cytology yield to differentiate between mucinous and/or nonmucinous cysts was 47.6% (20/42), and the MFB histologic yield to differentiate between mucinous and/or nonmucinous cysts was 61.9% (26/42) (P Z .188). The percentage of cysts at high risk for malignancy by cytology was 54.7% (23/42), and MFB was 71.5% (30/42) (P Z .113). However, the ability of MFB to provide a specific cyst type diagnosis was 35.7% (15/42), and that for cytology was 4.8% (2/42) (P Z .001). Surgical histology was concordant with that of MFB in 6 of 7 patients (85%), and with that of cytology in 1 of 7 patients (15%). Conclusion: The cyst tissue acquisition yield for MFBs was 90%. Although cytology of cyst fluid and MFB were comparable in distinguishing mucinous and nonmucinous cysts and detecting cysts at high risk for malignancy, MFB was far superior to cytology for providing a specific cyst diagnosis. (Gastrointest Endosc 2018;-:1-8.)
Abbreviations: CEA, carcinoembryonic antigen; cPNET, cystic pancreatic neuroendocrine tumor; EUS-FNA, EUS-guided FNA; IPMN, intraductal papillary neoplasm; MCN, mucinous cystic neoplasm; MFB, microforceps biopsy. DISCLOSURE: O. Basar has received a grant from U.S. Endoscopy Group, Inc. D. Yang is a consultant for US Endoscopy. J. Samarasena is a consultant for US Endoscopy. C. DiMaio is a consultant and speaker for Boston Scientific and Medtronic. M. Wagh is a consultant for Boston Scientific and Medtronic. K. Chang is a consultant for Olympus and Cook Medical and a speaker for Medtronic. W. Brugge does research for Boston Scientific. All other authors disclosed no financial relationships relevant to this publication. *Authors Pitman and Brugge contributed equally to the article. Copyright ª 2018 by the American Society for Gastrointestinal Endoscopy
www.giejournal.org
0016-5107/$36.00 https://doi.org/10.1016/j.gie.2018.02.039 Received September 22, 2017. Accepted February 20, 2018. Current affiliations: Pancreas Biliary Center, Gastrointestinal Unit, Massachusetts General Hospital, Boston, Massachusetts (1), Department of Gastroenterology, Eryaman Private Hospital, Ankara, Turkey (2), Division of Gastroenterology, University of Florida Health, Gainesville, Florida (3), Division of Gastroenterology and Hepatology, University of California, Aurora, California (4), Division of Gastroenterology, Mount Sinai Medical Center, New York, New York (5), Division of Gastroenterology and Hepatology, University of Colorado, Denver, Colorado (6), Department of General and Gastrointestinal Surgery; (7), Department of Pathology, Massachusetts General Hospital, Boston, Massachusetts (8). Reprint requests: Omer Basar, MD, 3-H GI Associates, Zero Emerson Place, Blossom St, Massachusetts General Hospital, Boston, MA 02114.
Volume
-,
No.
-
: 2018 GASTROINTESTINAL ENDOSCOPY 1
Value of microforceps biopsy for pancreatic cysts
Basar et al
Primary pancreatic cysts are broadly divided into nonneoplastic and neoplastic, the latter termed pancreatic cystic neoplasms. Pancreatic cystic neoplasms are further classified into nonmucinous neoplasms (serous cystadenomas), mucinous neoplasms (intraductal papillary neoplasms [IPMN]), and mucinous cystic neoplasms [MCN]). In addition to primary pancreatic cysts, solid tumors can become cystic secondarily from degeneration, including pancreatic ductal adenocarcinoma and cystic pancreatic neuroendocrine tumors (cPNETs). The risk of malignancy in pancreatic cysts usually dictates the need for surgical resection.1,2 The most commonly used tools for both the diagnosis and differentiation of pancreatic cysts are cross-sectional imaging and EUS.3,4 Because imaging alone has some limitations, cyst fluid analysis obtained by EUS-guided FNA (EUS-FNA) aids in further evaluation.4-6 Cytology is limited because of scant cellularity and difficulty in detecting thin, watery, extracellular mucin, but cytology has been shown to be highly accurate for the diagnosis of a cyst at high risk for malignancy.7-10 However, cytology is dependent on cells being shed into the cyst fluid for analysis. Interpretation challenges arise from GI contamination, degenerative changes of the cells, heterogeneity of the cyst lining epithelium, and lack of experience and interpretive expertise.11 Recently, a U.S. Food and Drug Administration– approved single-use Moray microforceps biopsy device (U.S. Endoscopy, Mentor, Ohio) (Fig. 1) has been designed for use in EUS procedures to enable sampling from cysts that can be accessed with a 19-gauge EUS-FNA needle. The aim of this study was to compare the tissue acquisition and diagnostic yield of the microforceps biopsy (MFB) with cyst fluid cytology.
PATIENTS AND METHODS This study is a retrospective, controlled, open label, and multicenter investigation including Massachusetts General Hospital, University of Florida Health, University of California at Irvine, Mount Sinai Hospital, and University of Colorado. The patients gave their consent for EUS-FNA, cyst fluid analysis, and for use of the MFB device in the evaluation of pancreatic cysts. The institutional review boards of each center approved the study. The data of all patients (initial and subsequent) undergoing MFBs between 2015 and 2016 were collected, without a predefined study protocol. Exclusion criteria included patients with a bleeding diathesis, women with known pregnancy, patients with a history of pancreatic cancer, patients with acute pancreatitis or a high clinical suspicion of a pseudocyst or abscess, patients with a solid pancreatic mass or a clinically suspected pancreatic adenocarcinoma with cystic degeneration, and patients with cysts of extra-pancreatic origin. 2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2018
Figure 1. Moray microforceps. Permission granted by US Endoscopy.
EUS-FNA and pancreatic cyst biopsy In this multicenter study, data of 42 patients who had cysts aspirated by EUS-FNA were collected, and biopsy specimens were then obtained with an MFB device. Before the start of the procedure, all patients were given broad-spectrum antibiotics. The number, location, size of the cyst, presence or absence of septations, mural nodule and an adjacent mass on EUS were recorded. After evaluating the cyst with EUS, a 19-gauge Flex EUS needle (Boston Scientific, Marlborough, Mass) was placed into the cyst cavity, and the cyst fluid was aspirated first. Without removing the needle, an MFB device was introduced through the needle (Fig. 2), and pinch biopsy specimens were obtained from the cyst wall, septations, nodules, or adjacent masses if present (Fig. 3). The standardized order of obtaining biopsy specimens was the adjacent mass, mural nodule, cyst wall, and septations, respectively. If the biopsy specimen seemed insufficient in size for histology, additional passes were made. Aspirated cyst fluid was sent for CEA evaluation and for cytology analysis by using a cytospin preparation stained with routine Papanicolaou stain. The MFB specimens were sent in formalin and processed as a routine histology specimen. Adverse events including intra-cystic bleeding, pain, or pancreatitis related to the procedure were assessed during the procedure and for 2 hours after the procedure before discharge.
Cytology and histology evaluation Failure to obtain fluid by EUS-FNA or tissue by MFB were considered acquisition failures and were classified as nondiagnostic. The term tissue acquisition yield referred to the ability to collect fluid or tissue for analysis and was calculated as the number of patients with fluid or tissue obtained by aspiration or MFB, which was either diagnostic or nondiagnostic, divided by the total number of patients. The diagnostic tissue yield was evaluated at 3 levels (defined below) and calculated as the number of cases www.giejournal.org
Basar et al
Value of microforceps biopsy for pancreatic cysts
Figure 2. EUS image of microbiopsy forceps wide open inside a pancreatic cyst.
that were either diagnostic or nondiagnostic divided by the total number of patients. The nondiagnostic cases were all the cases that were nondiagnostic by cytology and histology plus the acquisition failure cases. Level I analysis referred to the ability to distinguish a mucinous from a nonmucinous cyst. Level II analysis referred to the ability to detect a cyst at high risk for malignancy (high-grade dysplasia or adenocarcinoma). Level III analysis referred to the ability to make a specific diagnosis of the pancreatic cyst (such as an IPMN, serous cystadenoma, cPNET, or MCN). On cytology evaluation, a cyst was diagnostic of a mucinous cyst if there was extracellular mucin or mucinous epithelium considered lesional; otherwise, the cyst was classified as nonmucinous.12 Transgastric FNAs with nonspecific gastric foveolar type epithelium without thick extracellular mucin were defined as nondiagnostic. Pancreatic cyst fluid CEA >192 ng/mL was considered diagnostic for a mucinous cyst. Cytology was diagnostic of a high-risk cyst if there were cells diagnostic of a malignant cyst or a cyst at high-risk for malignancy such as a mucinous cyst with high-grade epithelial dysplasia12 or a cystic neuroendocrine tumor.13 The histology evaluation of the MFB tissue followed standard histology definitions for epithelial type: mucinous epithelium with cytoplasmic mucin visible on routine hematoxylin and eosin stain (Figs. 4 and 5), serous epithelium (nonmucinous cuboidal epithelium with or without confirmation with special stains to support cytoplasmic glycogen) (Fig. 6), acinar epithelium (granular nonmucinous cytoplasm with confirmation of exocrine granules with ancillary testing) (Fig. 7) and endocrine epithelium (with or without confirmation with ancillary tests) (Fig. 8). Histology was diagnostic of a high-risk cyst with the diagnosis of a mucinous cyst with high-grade dysplasia or invasive adenocarcinoma or neuroendocrine tumor. The presence of subepithelial ovarian type stroma defined an MCN, and the absence of such stroma defined an IPMN. Fibrous tissue without epithelium www.giejournal.org
Figure 3. Microbiopsy forceps after obtaining a pinch biopsy specimen from a pancreatic cyst.
was a tissue acquisition success, but nondiagnostic for cyst type. In patients who underwent pancreatic surgery in a 1-year follow-up period, surgical histology results were used as a criterion standard to determine the specificity and accuracy of MFB.
Data analysis The Statistical Package for Social Sciences for Windows version 16.0 (SPSS Inc, Chicago, Ill, USA) was used to analyze the data. A chi-square test was used for comparison of categorical variables. The value of P < .05 was considered statistically significant.
RESULTS Data regarding 42 patients who had EUS-FNA and MFB (23 female and 19 male) were entered over a 1-year period into the study from multiple centers for data analysis. The mean age of the patients was 69 years (range 27-91). Cysts were located in the head of the pancreas in 16 patients (38.1%), in the body in 17 (40.5%), and in the tail in 9 (21.4%). The mean size of the cysts was 28.2 mm (range 12-60 mm), with 25 of 42 (60%) cysts septated and 17 of 42 (40%) unilocular. A mural nodule was seen in 4 patients (9.6%), and cysts and an adjacent mass were observed in 1 (2.4%).
Tissue acquisition yield Cyst fluid was obtained from 37 patients; the EUS-FNA fluid acquisition yield was 88.1% (37/42). The microforceps tissue acquisition yield was 90.4% (38/42). The pinch biopsy specimen was not sufficient for histology Volume
-,
No.
-
: 2018 GASTROINTESTINAL ENDOSCOPY 3
Value of microforceps biopsy for pancreatic cysts
Basar et al
Figure 4. Microbiopsy forceps biopsy specimen of an intraductal papillary mucinous neoplasm with low-grade dysplasia. Note the absence of subepithelial ovarian-type stroma (H&E, orig. mag. 200).
Figure 6. Microbiopsy forceps biopsy specimen of a serous cystadenoma. Note that a nonmucinous cuboidal epithelium lines the fibrous septum (H&E, orig. mag. 400).
Diagnostic tissue yield
Figure 5. Papillary epithelium with high-grade dysplasia consistent with an intraductal papillary mucinous neoplasm with high-grade dysplasia. The resected tumor demonstrated a microinvasive adenocarcinoma (H&E, orig. mag. 200).
processing in 4 of 42 patients (9.6%) (2 from a cyst wall and 2 from a septum). Overall, 30 of 42 successful MFB specimens (71.4%) were obtained from the cyst wall, 6 of 42 (14.4%) from septations, 1 of 42 (2.4%) from an adjacent mass, and 1 of 42 (2.4%) from a mural nodule. The microforceps tissue acquisition yield was not significantly different between the groups according to the presence or absence of septations or size and location of cysts (P Z .122; P Z .761, and P Z .176; respectively) (Table 1). During the procedure and before discharge, no serious adverse events were observed. One patient had mild abdominal pain, which lasted for an hour after the procedure. Another patient had self-limited, intracystic bleeding, which did not progress and did not result in any symptoms. 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2018
Level I analysis (diagnostic ability to distinguish a mucinous from nonmucinous cyst). The level I diagnostic tissue yield by using cyst fluid cytology was 47.6% (20/42) and was not statistically significantly different from the MFB histology diagnostic yield of 61.9% (26/42) (P Z .188) (Table 2). Excluding acquisition failures, 45.9% of cases (17/37) were nondiagnostic by cyst fluid cytology, and 31.5% of cases (12/38) were nondiagnostic by MFB (Table 3). In addition, 12 of 37 cysts (32.4%) had a CEA level >192 ng/mL, when acquisition failures were excluded. The average CEA among mucinous cysts was 893.1 ng/mL, as compared with 47.4 ng/mL among nonmucinous cysts (P Z .003). Aspirated cyst fluid CEA levels >192 ng/mL were found in 3 patients in the EUSFNA group (with only cyst cytology results) and in 4 patients in the MFB group (with only cyst histology results) whose results were nondiagnostic according to Level I analysis. Level II analysis (risk of malignancy). The level II diagnostic tissue yield by using cyst fluid cytology of 54.7.0% (23/42) was not statistically significantly different when compared with MFB of 71.5% (30/42) (P Z .113) (Table 2). Cytology was diagnostic for low-grade dysplastic epithelium in 50.0% of patients (21/42), and MFB was diagnostic for low-grade dysplastic epithelium in 64.2% of patients (27/42). Cytology diagnosed high-grade dysplastic epithelium in 2 patients (4.8%), and MFB diagnosed high-grade dysplastic epithelium in 3 patients (7.1%). When we excluded the acquisition failures, the cyst fluid did not contain a valuable epithelium for grading dysplasia in 37.8% of patients (14/37), and the MFB specimen did not contain epithelium for grading in 21.0% of patients (8/38) (Table 3). www.giejournal.org
Basar et al
Value of microforceps biopsy for pancreatic cysts
Figure 7. Microbiopsy forceps biopsy of an acinar cell cystadenoma A, with immunohistochemical stain for trypsin B, supporting the diagnosis (H&E, orig. mag. 400).
TABLE 1. Tissue (diagnostic and nondiagnostic) acquisition yield relative to cyst features by microforceps biopsy Tissue acquisition yield
P value
Unilocular (17)
17/17 (100%)
.122
Septated (25)
21/25 (84%)
Morphology of cyst
Size of cyst <30 mm (24)
22/24 (92%)
30 mm (18)
16/18 (89%)
.761
Location of cyst Head (16)
16/16 (100%)
Body (17)
15/17 (88%)
Tail (9) Figure 8. Small, uniform polygonal cells with coarse chromatin typical of solid well-differentiated neuroendocrine tumors is seen in this forceps biopsy of a cystic well-differentiated neuroendocrine tumor (H&E, orig. mag. 200).
Level III analysis (specific cyst diagnosis). There was a significant difference in the level III diagnostic tissue yield between cyst fluid cytology (2/42) and MFB histology (15/42) (4.8% and 35.7%; P Z .001; respectively), (Table 2). The 15 patients who had a specific histology cyst diagnosis by MFB compared with cytology are shown in Table 4.
Validation of MFB and cytology In a 1-year follow-up period, 17% of patients (7/42) underwent pancreatic surgery. Specific MFB results were concordant with surgical histology in 6 of 7 patients (85%). There was concordance between the specific cytology diagnosis and surgical histology in 1 of 7 (15%) patients (Table 5).
DISCUSSION We show that tissue obtained by MFB adds valuable, diagnostic tissue in the EUS-FNA evaluation of pancreatic www.giejournal.org
.176
7/9 (78%)
cysts. In our study, the MFB tissue acquisition yield was 90%. MFB was found to be superior to cytology for providing a specific cyst diagnosis and was comparable to cytology in detecting a mucinous cyst and a high-risk cyst. The use of the disposable microforceps was not associated with any significant adverse events other than intracystic bleeding in 1 patient. The goal of the management of patients with pancreatic cysts is to accurately define and resect cysts before progression of malignancy, while avoiding unnecessary follow-up procedures and surgery. Because it is known from other types of tissues that dysplasia does not always progress to cancer,14,15 and the natural history of pancreatic cysts is not completely clear, the accurate diagnosis and follow-up criteria for pancreatic cysts is still a challenging issue. The premalignant cysts of the pancreas such as mucinous cysts (including IPMNs and MCNs) can be cured before developing an invasive adenocarcinoma. However, not all patients with pancreatic cystic neoplasms require surgery and not all the patients are good candidates for surgery. Not all the premalignant mucinous cysts are at high risk of malignancy, and surgery is not the best Volume
-,
No.
-
: 2018 GASTROINTESTINAL ENDOSCOPY 5
Value of microforceps biopsy for pancreatic cysts
Basar et al
TABLE 2. Diagnostic tissue yield of EUS-guided FNA cyst fluid cytology and of microforceps biopsy histology EUS-guided FNA (cytology) 42 patients
Microforceps biopsy (histology) 42 patients
Diagnostic
20 (47.6%)
26 (61.9%)
Nondiagnostic*
22 (52.4%)
16 (38.1%)
Diagnostic
23 (54.7%)
30 (71.5%)
Nondiagnostic*
19 (45.3%)
12 (28.5%)
Diagnosis
P value
Level I
.188
Level II
.113
Level III
.001
Diagnostic Nondiagnostic*
2 (4.8%)
15 (35.7%)
40 (95.2%)
27 (64.3%)
Level I–diagnostic ability to differentiate a mucinous from nonmucinous cyst. Level II–diagnostic of degree of dysplasia. Level III–diagnostic of specific type of cyst. *Nondiagnostic: the combinations of the cases that were nondiagnostic by cytology and histology plus the acquisition failure cases.
TABLE 3. Number of patients according to diagnostic tissue yields
Diagnosis Mucinous vs nonmucinous
TABLE 4. The results of cytology (by EUS-guided FNA) in 15 patients who had a specific histology cyst diagnosis (by MFB)
EUS-guided FNA (cytology) 42 patients
Microforceps biopsy (histology) 42 patients
20 (47.6%)
26 (61.9%)
6 IPMNs
6/6 Nondiagnostic 3/3 Nondiagnostic
Specific histology diagnosis by MFB
Specific cytology diagnosis by EUS-guided FNA
Nondiagnostic
17 (40.5%)
12 (28.5%)
3 SCNs
Acquisition failures*
5 (11.9%)
4 (9.6%)
2 MCNs
2/2 Nondiagnostic
2 PDACs
1/2 PDACs
Low-grade (risk) epithelium
21 (50.0%)
27 (64.2%)
High-grade (risk) epithelium
2 (4.8%)
3 (7.1%)
Nondiagnostic
14 (33.3%)
8 (19.1%)
Acquisition failures*
5 (11.9%)
4 (9.6%)
Specific diagnosis present
2 (4.8%)
15 (35.7%)
No specific diagnosis
35 (83.3%)
23 (54.7%)
Acquisition failures*
5 (11.9%)
4 (9.6%)
*The number of patients from whom cyst fluid was not obtained and the microforceps biopsy did not procure sufficient tissue for histology.
choice in elderly patients with comorbid conditions in which conservative follow-up would be a better option. For example, most patients with branch duct–IPMNs.16 In addition, distinguishing between a branch duct–IPMN and MCN is important because all MCNs are resected regardless of grade because of the ease of the usual distal pancreatectomy compared with the expense and anxiety of life-long surveillance for progression to malignancy in the typical middle-aged woman.17 Therefore, determining cyst type as well as the malignant potential of a cyst dictates the best treatment strategy while avoiding unnecessary surgery. A multidisciplinary approach with clinical, radiologic, and cytologic information including biochemical and genetic 6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2018
1/2 Nondiagnostic (false positive low-risk mucinous cyst) 1 cPNET
1/1 Nondiagnostic (false positive high-risk mucinous cyst)
1 Acinar cystadenoma
1/1 Nondiagnostic (false positive low-risk mucinous cyst)
MFB, Microforceps biopsy; IPMN, intraductal papillary neoplasm; SCN, serous cystic neoplasm; MCN, mucinous cystic neoplasm; PDAC, pancreatic ductal adenocarcinoma; cPNET, cystic pancreatic neuroendocrine tumor.
analysis is generally recommended to diagnose a pancreatic cyst.17 Cyst fluid can be aspirated during EUS-FNA and analyzed for biochemical testing including amylase and CEA levels, cytology analysis and genetic mutational analysis.18 The accuracy of EUS-FNA cytology is very good with a very low false-positive rate but a relatively high falsenegative rate.19 In clinical practice, a high cyst fluid CEA level with the presence of mucin on cytology is the most common approach to differentiating a mucinous cyst from a nonmucinous cyst. Although cytology can detect a high-risk cyst by diagnosing a cPNET, high-grade mucinous dysplasia or invasive carcinoma with over 90% specificity, most cysts evaluated are not high risk, and the specificity of diagnosing a mucinous cyst in general is only 50%.7-10 Furthermore, cytology alone is insufficient for diagnosis in up to 60% of patients because of the absence of epithelial www.giejournal.org
Basar et al
Value of microforceps biopsy for pancreatic cysts
TABLE 5. Comparison of the diagnostic results of 7 patients who underwent pancreatic surgery Specific cytology diagnosis by EUS-guided FNA Nondiagnostic* Nondiagnostic* (mucinous cyst with LGD) PDACz
Specific histology diagnosis by MFB
Surgical histology diagnosis
Serous cystadenomay
Serous cystadenoma
PDAC arising from mucinous cyst with HGDy
PDAC arising from IPMN
PDAC arising from IPMN with HGDy
PDAC arising from IPMN with HGD
Nondiagnostic*
IPMN with LGDy
IPMN with LGD
Nondiagnostic*
Nondiagnostic*
MCN with LGD
Nondiagnostic* (mucinous cyst with LGD)
Acinar cystadenomay
Acinar cystadenoma
Nondiagnostic* (mucinous cyst with HGD)
cPNETy
cPNET
MFB, Microforceps biopsy; LGD, low-grade dysplasia; PDAC, pancreatic ductal adenocarcinoma; HGD, high-grade dysplasia; IPMN, intraductal papillary neoplasm; MCN, mucinous cystic neoplasm; cPNET, cystic pancreatic neuroendocrine tumor. *Nondiagnostic: the nondiagnostic cases by cytology and histology or the acquisition failure cases. ySpecific MFB histology was concordant with surgical histology. zSpecific cytology was concordant with surgical histology.
cells in cyst fluid, the presence of GI contamination, the heterogeneity of mucinous cyst dysplasia, and the lack of experience and expertise in cytology interpretation.7-10 In addition, the fragile cells of serous cystadenoma rarely survive processing for diagnosis,20 and the subepithelial ovarian-type stroma required for the diagnosis of MCN is not present in aspirated cyst fluid.21 Although EUS-FNA cyst fluid cytology has value in the differentiation between mucinous/non-mucinous cysts and accurately diagnosing high-risk cysts, it cannot usually specifically define the type of cyst. MFB provides tissue from the cyst wall, septations, and mural nodules for histology thus providing tissue for ancillary studies such as immunohistochemistry. The subepithelial cyst wall tissue can be biopsied by micro-forceps to show the ovarian type stroma, which would further differentiate a MCN from an IPMN.22 Furthermore, MFB can provide tissue for ancillary studies to confirm the diagnosis of serous cystadenoma, cPNET, and acinar cell cystadenoma.23 Although statistically not significant, the MFB histology increased the number of mucinous/nonmucinous diagnostic cyst results in a moderate proportion of patients (20-26 of 42) in our study. On the other hand, cytology alone was diagnostic of degree of dysplasia in 23 of 42 patients (54.7%) (21 were low-grade epithelium, 2 were high-grade epithelium). Furthermore, the pinch biopsy specimens were diagnostic of the degree of dysplasia in 30 of 42 patients (71.5%), with only MFB (27 low-grade epithelium, 3 high-grade epithelium). The MFB histology again increased the detection of low-grade epithelium in a moderate proportion of patients (from 21 patients to 27 patients out of 42 patients). Cyst fluid CEA (CEA >192 ng/mL) has been shown to be one of the most accurate methods for identifying a mucinous cyst, whereas CEA alone is neither a marker to differentiate a malignant pancreatic cyst from benign nor to differentiate the www.giejournal.org
subgroups of mucinous cysts.7,24 In our study, 32.4% of the aspirated cysts had CEA levels >192 ng/mL, supporting a mucinous cyst. Interestingly, CEA levels >192 ng/mL were detected in 3 patients whose EUS-FNA could not differentiate between a mucinous and a nonmucinous cyst, and in 4 patients whose MFB histology could not differentiate between a mucinous and a nonmucinous cyst. Although CEA lacks the ability to identify the degree of dysplasia and the specific cyst type, adding cyst fluid CEA analysis to MFB may help differentiate mucinous from nonmucinous cysts. Most importantly, we demonstrated that the diagnostic yield of the MFB for a specific type of cyst was significantly higher than the diagnostic cytology yield. Moreover, when specific diagnostic results with MFB histology were compared with cytology results, none of the 6 patients with IPMN, none of the 3 patients with serous cystadenoma, and none of the 2 patients with MCN could be diagnosed. Additionally, a patient with a cPNET was misdiagnosed as having a high-risk mucinous cyst, and a patient with pancreatic ductal adenocarcinoma and another patient with acinar cystadenoma were misdiagnosed as having low-risk mucinous cysts. However, it has to be kept in mind that the overall ability to provide a specific diagnosis with only MFB was merely 36%. Moreover, although most of the patients enrolled into the study did not require pancreatic surgery over a 1-year follow-up period, specific MFB diagnoses were far more concordant, compared with a specific cytology diagnosis with surgical histology diagnoses. In conclusion, our study showed that the MFB was not affected by the size, location, or septations of the cyst and provided a specific diagnosis of the subtype of pancreatic cyst significantly better than cytology. In addition, MFB was slightly better than cytology in differentiating between mucinous and nonmucinous cysts and detecting a high-risk cyst. The main limitation of our study was that our outcomes were based on a relatively small sample size in a retrospective Volume
-,
No.
-
: 2018 GASTROINTESTINAL ENDOSCOPY 7
Value of microforceps biopsy for pancreatic cysts
Basar et al
collection of data and were without a criterion standard diagnosis in the majority of patients. This study likely represents the early experience with this novel tissue collection device. Nevertheless, our findings demonstrated that an EUS-guided MFB provided diagnostic tissue from a pancreatic cyst at a rate superior to cyst fluid cytology.
REFERENCES 1. Yoon WJ, Brugge WR. Pancreatic cystic neoplasms: diagnosis and management. Gastroenterol Clin North Am 2012;41:103-18. 2. Brugge WR. Diagnosis and management of cystic lesions of the pancreas. J Gastrointest Oncol 2015;6:375-88. 3. Mortelé KJ. Cystic pancreatic neoplasms: imaging features and management strategy. Semin Roentgenol 2013;48:253-63. 4. Brugge WR. The use of EUS to diagnose cystic neoplasms of the pancreas. Gastrointest Endosc 2009;69:S203-9. 5. Yoon WJ, Brugge WR. Endoscopic ultrasound and pancreatic cystic lesions-diagnostic and therapeutic applications. Endosc Ultrasound 2012;1:75-9. 6. Brugge WR, Lewandrowski K, Lee-Lewandrowski E, et al. Diagnosis of pancreatic cystic neoplasms: a report of the cooperative pancreatic cyst study. Gastroenterology 2004;126:1330-6. 7. Cizginer S, Turner BG, Bilge AR, et al. Cyst fluid carcinoembryonic antigen is an accurate diagnostic marker of pancreatic mucinous cysts [published correction appears in Pancreas. 2013;42:728]. Pancreas 2011;40:1024-8. 8. Zhan XB, Wang B, Liu F, et al. Cyst fluid carcino embryonic antigen concentration and cytology by endosonography-guided fine needle aspiration in predicting malignant pancreatic mucinous cystic neoplasms. J Dig Dis 2013;14:191-5. 9. Pitman MB, Genevay M, Yaeger K, et al. High-grade atypical epithelial cells in pancreatic mucinous cysts are a more accurate predictor of malignancy than “positive” cytology. Cancer Cytopathol 2010;118: 434-40. 10. Springer S, Wang Y, Dal Molin M, et al. A combination of molecular markers and clinical features improve the classification of pancreatic cysts. Gastroenterology 2015;149:1501-10.
8 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2018
11. Pitman MB, Lewandrowski K, Shen J, et al. Pancreatic cysts: preoperative diagnosis and clinical management. Cancer Cytopathol 2010;118: 1-13. 12. Pitman MB, Centeno BA, Daglilar ES, et al. Cytological criteria of highgrade epithelial atypia in the cyst fluid of pancreatic intraductal papillary mucinous neoplasms. Cancer Cytopathol 2014;122:40-7. 13. Morales-Oyarvide V, Yoon WJ, Ingkakul T, et al. Cystic pancreatic neuroendocrine tumors: the value of cytology in pre-operative diagnosis. Cancer Cytopathology 2014;122:435-44. 14. Wani S, Puli SR, Shaheen NJ, et al. Esophageal adenocarcinoma in Barrett’s esophagus after endoscopic ablative therapy: a meta-analysis and systematic review. Am J Gastroenterol 2009;104:502-13. 15. Ostör AG. Natural history of cervical intraepithelial neoplasia: a critical review. Int J Gynecol Pathol 1993;12:186-92. 16. Schmidt CM. Pancreatic cyst cytology: optimization of cancer risk profiling. Cancer Cytopathol 2013;121:2-3. 17. Pitman MB, Brugge WR, Warshaw AL. The value of cyst fluid analysis in the pre-operative evaluation of pancreatic cysts. J Gastrointest Oncol 2011;2:195-8. 18. Brugge WR, De Witt J, Klapman JB, et al. Techniques for cytologic sampling of pancreatic and bile duct lesions: The Papanicolaou Society of Cytopathology Guidelines. Cytojournal 2014;11:2. 19. Woolf KM, Liang H, Sletten ZJ, et al. False-negative rate of endoscopic ultrasound-guided fine-needle aspiration for pancreatic solid and cystic lesions with matched surgical resections as the gold standard: one institution’s experience. Cancer Cytopathol 2013;121:449-58. 20. Belsley NA, Pitman MB, Lauwers GY, et al. Serous cystadenomas of the pancreas: diagnostic performance of imaging and fine needle aspiration biopsy. Cancer Cytopathol 2008;114:102-10. 21. Centeno BA, Stelow E, Pitman MB. Pancreatic cytohistology. Cytohistology of small tissue samples. Published in association with the Papanicolaou Society of Cytopathology. KR Geisinger, MB Pitman, series editors. Cambridge, UK: Cambridge University Press; 2015. 22. Adsay NV. Cystic neoplasia of the pancreas: pathology and biology. J Gastrointest Surg 2008;12:401-4. 23. Chen AL, Misdraji J, Brugge WR, et al. Acinar cell cystadenoma: a challenging cytology diagnosis, facilitated by moray() micro-forceps biopsy. Diagn Cytopathol 2017;45:557-60. 24. Scourtas A, Dudley JC, Brugge WR, et al. Preoperative characteristics and cytological features of 136 histologically confirmed pancreatic mucinous cystic neoplasms. Cancer 2017;125:169-77.
www.giejournal.org