- Email: [email protected]
Improving the diagnostic yield of single-operator cholangioscopy-guided biopsy of indeterminate biliary strictures: ROSE to the rescue? (with video)
ORIGINAL ARTICLE
Improving the diagnostic yield of single-operator cholangioscopy-guided biopsy of indeterminate biliary strictures: ROSE to the rescue? (with video) Shyam Varadarajulu, MD,1,* Ji Young Bang, MD, MPH,2,* Muhammad K. Hasan, MD,1 Udayakumar Navaneethan, MD,1 Robert Hawes, MD,1 Shantel Hebert-Magee, MD1 Orlando, Florida, USA
Background and Aims: Tissue diagnosis, regardless of technique or endoscope used, can be challenging in patients with indeterminate biliary strictures (IDBSs). This exploratory study evaluated the utility and role of rapid onsite evaluation of touch imprint cytology (ROSE-TIC) when single-operator cholangioscopy (SOC)-guided biopsies of IDBSs are performed. Methods: Patients with IDBSs were evaluated by intraprocedural ROSE-TIC during SOC-guided biopsy procedures. Final diagnosis was established by long-term patient follow-up in conjunction with off-site findings or surgical histology. The main outcome measure was to evaluate the utility of ROSE-TIC by determination of its operating characteristics and comparison with off-site histologic assessment. Results: Of 31 patients with IDBSs, tissue diagnosis was indeterminate at prior ERCP-guided brush and/or biopsy in 14, prior EUS-guided FNA (EUS-FNA) in 6, and a mass could not be identified at EUS in 11. The mean number of biopsies performed was 3.3 (range 1-8), and diagnostic interpretation by ROSE-TIC was diagnostic and/or suspicious for carcinoma in 15, benign in 13, atypical-reactive in 2, and bile duct intraductal papillary mucinous neoplasm in 1. Final diagnosis by surgical histology (n Z 4), death by disease (n Z 10), and patient follow-up (n Z 17) showed that the overall sensitivity of ROSE-TIC for diagnosing malignancy was 100%, specificity 88.9%, positive predictive value 86.7%, negative predictive value 100%, and diagnostic accuracy 93.5%. Conclusions: Preliminary data suggest that the diagnostic outcomes of SOC-guided biopsies in IDBSs can be significantly improved by using ROSE-TIC. This technique also may benefit centers that rely mainly on fluoroscopy-guided intraductal biopsies. (Gastrointest Endosc 2016;-:1-7.)
Establishing an accurate diagnosis in a biliary stricture can be challenging.1,2 Although ERCP is the initial modality commonly used for evaluating biliary strictures, the diagnostic sensitivities of both brush cytology and intraductal biopsies are poor. In a recent meta-analysis, the pooled sensitivity and specificity of brush cytology for
the diagnosis of malignant biliary strictures were 45% and 99%, respectively.3 Likewise, the pooled sensitivity and specificity of intraductal biopsies were only 48.1% and 99.2%, respectively. Given the suboptimal results with ERCP-based techniques, EUS-FNA is increasingly being used as an alternative technique for tissue acquisition
Abbreviations: CBD, common bile duct; EUS-FNA, EUS-guided FNA; H&E, hematoxylin and eosin; IDBS, indeterminate biliary stricture; IPMN, intraductal papillary mucinous neoplasm; PSC, primary sclerosing cholangitis; ROSE, rapid onsite evaluation; SOC, single-operator cholangioscopy; TIC, touch imprint cytology.
0016-5107/$36.00 http://dx.doi.org/10.1016/j.gie.2016.03.1497
DISCLOSURE: S. Varadarajulu and R. Hawes are consultants for Olympus Medical Systems and Boston Scientific. J. Bang is a consultant for Olympus Medical Systems. S. Hebert-Magee is a consultant for Boston Scientific. All other authors disclosed no financial relationships relevant to this publication. *Drs Varadarajulu and Bang contributed equally to the article.
Received December 22, 2015. Accepted March 26, 2016. Current affiliations: Center for Interventional Endoscopy, Florida Hospital, Orlando, Florida (1), Divison of Gastroenterology-Hepatology, Indiana University, Indianapolis, Indiana (2). Reprint requests: Shyam Varadarajulu, MD, Medical Director, Center for Interventional Endoscopy, Florida Hospital, 601 East Rollins Street, Orlando, FL 32803. If you would like to chat with an author of this article, you may contact Dr Varadarajulu at [email protected].
Copyright ª 2016 by the American Society for Gastrointestinal Endoscopy
www.giejournal.org
Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 1
Diagnostic yield of single-operator cholangioscopy-guided biopsy
in patients with biliary strictures.4 In a recent study that compared EUS-FNA with ERCP-guided brush cytology and intraductal biopsies,5 the sensitivity of EUS-FNA was superior to ERCP tissue sampling for pancreatic masses (100% vs 38%) and comparable for biliary lesions (79% for both) and indeterminate strictures (80% vs 67%). However, hilar and proximal biliary tumors are sometimes difficult to visualize and sample using EUS-FNA.6 These technical limitations have led to the evolution of cholangioscopy-based techniques for evaluating the biliary ductal system.7 Single-operator cholangioscopy (SOC) not only enables the visualization of the biliary ductal system but also facilitates tissue sampling under direct vision.8 In a retrospective study of 30 patients with extrahepatic biliary strictures in whom both EUS-FNA and ERCPguided brush cytology were inconclusive, SOC-guided biopsies (Boston Scientific Corporation, Natick, Mass) yielded a diagnostic accuracy of 77% for cholangiocarcinoma.6 A recent meta-analysis of 10 studies evaluating 456 patients demonstrated pooled sensitivity and specificity of 60.1% and 98%, respectively, for SOC-guided biopsies of malignant biliary strictures.8 Reasons for the poor diagnostic sensitivity of cholangioscopy-guided biopsies are manifold. First, SOC is a technically difficult adjunct procedure to ERCP and is graded as level III by the American Society for Gastrointestinal Endoscopy.9 Visualization of biliary strictures and procurement of specimens by using intraductal biopsy forceps requires technical dexterity and considerable procedural expertise. Second, the “mini” biopsy forceps used in conjunction with SOC have small jaws that frequently yield inadequate specimens. Also, the optimal number of SOCguided biopsy specimens required to establish a definitive histologic diagnosis is undetermined. Third, most patients referred to tertiary-care centers for evaluation of an indeterminate biliary stricture (IDBS) have an indwelling endoprosthesis placed at outside facilities. These stents induce an inflammatory response that negatively impacts the microscopic evaluation of cellular architecture by inducing reactive atypia. Finally, the presence and extent of chromosomal abnormalities such as polysomy in the setting of preneoplastic diseases such as primary sclerosing cholangitis (PSC) can be multifocal, with varying levels of predisposition to cholangiocarcinoma.10 Consequently, the diagnostic yield of blind biopsies in PSC-induced biliary strictures can be low because of sampling errors. Given these limitations, patients with IDBSs often require multiple endoscopic or radiologic interventions before a diagnosis can be established reliably. The objective of this exploratory study was to evaluate the concept and diagnostic utility of rapid onsite evaluation of touch imprint cytology (ROSE-TIC) when SOC-guided biopsies of IDBSs are performed. Our hypothesis was that if a pathologist was available onsite for rendering preliminary interpretation, then the limitations of inadequate 2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
Varadarajulu et al
specimen procurement and sampling error could be minimized, leading to improved diagnostic performance.
MATERIALS AND METHODS Study design and patients This was a retrospective study of all patients who underwent SOC-guided biopsies of IDBSs over a 30-month period from April 2013 to October 2015. Patients were identified by querying our electronic ERCP database. Included in the study were patients aged 18 years referred for evaluation of IDBS by SOC. IDBS in this study was defined as patients with bile duct strictures who, on prior ERCP, had nondiagnostic intraductal biliary brushings and/or biopsy and/or nondiagnostic EUS-FNA cytology. Excluded were patients with acute cholangitis, acute pancreatitis, or coagulation disorders. For each case, patient demographics and clinical history such as prior attempts at tissue diagnosis, prior endoscopic instrumentation, procedural findings including stricture location, number of biopsies, postprocedural adverse events, onsite and final (off-site) pathology interpretation of bile duct biopsy specimens, postoperative surgical pathology reports, inpatient medical records, and outpatient clinical and radiologic follow-up information were reviewed. Per endoscopy unit policy, all ERCP patients were routinely contacted 5 to 7 days after the procedure by an endoscopy nursing coordinator to assess for adverse events. The study protocol was approved by the Florida Hospital institutional review board (IRB no. 828895).
SOC system and procedural techniques Direct Visualization System (SpyGlass, Boston Scientific) was used for performing cholangioscopy-guided biopsies from April 2013 to January 2015, and the more recent digital version (SpyGlass DS) was used from February to October 2015. All SOC procedures were performed with the patients under general anesthesia in the prone position, by 1 of 4 therapeutic endoscopists (S.V., R.H., M.H., U.N.). Antibiotic prophylaxis was administered at the discretion of the endoscopist. Endoscopic sphincterotomy was performed, or had been performed previously, before each examination. The cholangioscope was advanced through the accessory channel of the duodenoscope. Although a guidewire was used occasionally for the passage of the older SOC system into the bile duct, biliary cannulation was achieved in all cases by using the free-hand technique with the newer system. Biliary stricture dilation was performed as needed to facilitate the passage of the cholangioscope. In general, the instrument was passed proximally, suction was used to clear bile and contrast material, sterile saline solution was infused to optimize imaging, and the cholangioscope was slowly withdrawn to perform a systematic inspection of the ductal mucosa. Suspicion of malignant stricture on www.giejournal.org
Varadarajulu et al
Diagnostic yield of single-operator cholangioscopy-guided biopsy
cholangioscopy was based on the presence of “tumor vessels” characterized by dilated tortuous vasculature coursing through the epithelium accompanied by variable degrees of an exophytic mass protruding into the lumen of the bile duct. Tissue acquisition was undertaken by using the micro-biopsy forceps (SpyBite, Boston Scientific) (Fig. 1) (Video 1, available online at www.giejournal.org).
Specimen processing The biopsy specimens were processed for ROSE by using the touch imprint technique. The biopsy specimen was carefully pressed onto the slide, allowing the superficial cells to adhere, and then gently lifted with forceps, thereby creating a touch imprint of the specimen on the slide (Video 1). The slide was air dried and stained with Diff-Quik (Baxter Diagnostics Inc, McGaw Park, Ill) for ROSE. A single pathologist onsite analyzed the TIC specimens and provided real-time diagnosis that included benign, atypical-favor reactive, nondiagnostic-insufficient material, suspicious for malignancy, malignancy, and other diagnoses. Specimens were categorized as benign when the cytology sample did not reveal malignancy or any cellular atypia, atypical-favor reactive when the specimen displayed morphologic features beyond recognizable normal tissue components or only reactive changes, nondiagnostic-insufficient when technical or sampling limitations precluded the pathologist from providing any useful information from the specimen relative to the lesion sampled, suspicious when typical features of a specific malignant neoplasm were present but were qualitatively and/or quantitatively insufficient for a conclusive diagnosis, and malignant when the specimen unequivocally displayed malignant cytologic characteristics that included adenocarcinoma or its variants. Biopsy specimens were obtained until TIC was deemed diagnostic. After TIC, the bile duct biopsy specimens were placed in 10% buffered formalin solution for fixation and were stained by using hematoxylin and eosin (H&E). All specimens were examined by a single pathologist (S.H.M.) with particular expertise in cytopathology. The final diagnosis was certified by review of both ROSE-TIC and histology specimens by a second independent pathologist, who was blinded to the findings at ROSE.
Patient follow-up Final diagnosis of malignancy was based on tissue diagnosis that consisted of review of off-site histologic material procured from SOC-guided biopsy, surgical biopsy, or Whipple resection. In patients without surgical pathology confirmation or those with benign disease, clinical records and radiologic studies were reviewed to confirm disease stability, progression, or resolution. Patients suspected to have benign disease were followed for a minimum duration of 6 months.
Outcome measures The main outcome measure was to evaluate the utility of ROSE-TIC in SOC-guided biopsy of IDBS by www.giejournal.org
Figure 1. Single-operator cholangioscopy-guided biopsy of a bile duct stricture.
determination of its operating characteristics and comparison with off-site histologic assessment.
Statistical analysis Categoric variables were summarized as frequencies and percentages, and continuous variables as means with standard deviation (SD) or medians with range. Operating characteristics for ROSE-TIC and off-site histology for diagnosis of malignancy were calculated and then compared by using the chi-square test. Receiver operating characteristic analysis was performed to compare the areas under the curve for ROSE-TIC and off-site histology by using the nonparametric method.11 Suspicious for malignancy on ROSE-TIC was considered as being synonymous with malignancy for determination of operating characteristics. Statistical significance was set as P value < .05. The dataset was compiled by using Microsoft Excel (Microsoft Corporation, Redmond, Wash), and all statistical analyses were performed by using Stata 13 (StataCorp LP, College Station, Tex).
RESULTS Patient characteristics and SOC findings Of 59 patients who underwent SOC-guided biopsies of IDBSs over the 30-month period, 31 patients were evaluated with ROSE-TIC and constituted the study cohort. The remaining 28 specimens were evaluated by off-site histology because an onsite cytopathologist was not available and hence the data were excluded from analysis. The mean age of the 31 patients was 65.2 ( SD) 17.7 years, range 22-88 years. Twenty (64.5%) were male, and all patients presented with obstructive jaundice. Prior investigations included a nondiagnostic EUS in 17 patients (54.8%; no definitive mass on EUS in 11, non-diagnostic EUS-FNA in 6 patients) and ERCP with nondiagnostic tissue acquisition (brush cytology with and/or without intraductal biopsy) in 14 (45.2%). Eighteen of 31 (58.1%) strictures were located Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 3
Diagnostic yield of single-operator cholangioscopy-guided biopsy
in the common hepatic or proximal common bile duct (CBD), 5 (16.1%) in the mid-CBD, and 8 (25.8%) in the distal CBD. Twenty-six of 31 patients (83.9%) had an indwelling plastic biliary endoprosthesis in situ. No procedural adverse events were encountered (Table 1).
ROSE-TIC findings A mean of 3.3 biopsies (range 1-8) was performed in 31 patients. Of the 31 ROSE-TIC evaluations performed, 15 (48.4%) were categorized as malignant (11 carcinoma and 4 suspicious for carcinoma) and 16 (51.6%) as nonmalignant (13 benign, 2 atypical-reactive, and 1 bile duct intraductal papillary mucinous neoplasm [IPMN]). Five patients had PSC-induced biliary strictures, and 2 of the 7 patients with cholangiocarcinoma had underlying PSC. In 12 of the 15 cases (80.0%) that were categorized as malignant by ROSE-TIC, there was complete concordance between ROSE-TIC, off-site histology, and the final diagnosis (Figs. 2A and B). In 2 of the 3 discordant cases, cytology rendered suspicious by ROSE was later proven to be bile duct IPMN by both off-site and surgical histology in 1 patient (case no. 17) and autoimmune pancreatitis in another patient (case no. 31). In the third patient (case no. 25), categorized to have malignant stricture by ROSE-TIC, off-site histology revealed only ductal atypia but was ultimately proven to be a cholangiocarcinoma by surgical histology. Of the 16 patients deemed to have nonmalignant strictures by ROSE, there was complete concordance for all 16 cases between ROSE-TIC and the final diagnosis. Two cases of reactive atypia (defined as nonmalignant) at ROSE-TIC were proven later by both off-site histology and long-term follow-up to have extrahepatic PSC and postcholecystectomy ischemic stricture. One patient with biliary IPMN at ROSE-TIC also was confirmed by both offsite and surgical histology. A summary of the correlation between ROSE-TIC, off-site histology, and final diagnosis is shown in Table 2. Although the image quality was significantly better with the second generation SOC system, there was no significant difference in the mean number of biopsies required to establish a diagnosis between the first and second generation systems (3.6 vs 3.0; P Z .43). However, there was a significant difference in the mean number of biopsies required to establish a diagnosis between patients with cholangiocarcinoma versus pancreatic adenocarcinoma (2.6 vs 4.8; P Z .02).
Final diagnosis The final diagnosis was malignancy in 13 (adenocarcinoma in 5, cholangiocarcinoma in 7, poorly differentiated carcinoma in 1) and nonmalignant lesions in 18 (benign in 11, PSC in 5, bile duct IPMN in 2); the diagnosis was established by surgical histology in 4, death by disease progression in 10, and clinical and radiologic follow-up in 17 patients, at a median follow-up of 385 days (range 82-879 days) (Table 2). 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
Varadarajulu et al
TABLE 1. Patient details and lesion characteristics Case Age, y
Sex
Stricture location
Prior workup
Survival
1
69
Male
Proximal-CHD
No mass at EUS
Dead
2
87
Male
Proximal-CHD
No mass at EUS
Dead
3
80
Female
Proximal-CHD
No mass at EUS
Dead
4
81
Female
Middle
EUS-FNA
Alive
5
58
Female
Proximal-CHD
No mass at EUS
Dead
6
68
Male
Proximal-CHD
No mass at EUS
Alive
7
86
Male
Proximal-CHD
No mass at EUS
Alive
8
56
Male
Proximal-CHD
EUS-FNA
Dead
9
45
Female
Distal
ERCP-TA
Alive
10
75
Male
Proximal-CHD
EUS-FNA
Dead
11
43
Female
Proximal-CHD
ERCP-TA
Alive
12
59
Male
Proximal-CHD
ERCP-TA
Alive
13
78
Female
Proximal-CHD
ERCP-TA
Dead
14
22
Male
Distal
No mass at EUS
Alive
15
50
Male
Middle
ERCP-TA
Alive
16
57
Male
Proximal-CHD
ERCP-TA
Dead
17
79
Male
Proximal-CHD
ERCP-TA
Alive
18
78
Female
Proximal-CHD
EUS-FNA
Dead
19
82
Male
Distal
ERCP-TA
Alive
20
84
Male
Distal
ERCP-TA
Alive
21
49
Male
Distal
No mass at EUS
Alive
22
37
Male
Middle
ERCP-TA
Alive
23
43
Female
Proximal-CHD
No mass at EUS
Alive
24
59
Male
Proximal-CHD
ERCP-TA
Alive
25
72
Male
Middle
No mass at EUS
Dead
26
88
Female
Proximal-CHD
ERCP-TA
Alive
27
34
Male
Distal
ERCP-TA
Alive
28
74
Female
Proximal-CHD
No mass at EUS
Alive
29
79
Male
Distal
EUS-FNA
Alive
30
74
Female
Middle
ERCP-TA
Alive
31
74
Male
Distal
EUS-FNA
Alive
CHD, common hepatic duct; EUS-FNA, EUS-guided FNA; TA, tissue acquisition.
Of the 13 patients with malignancy, at last follow-up, 11 were dead and 2 were undergoing chemotherapy for metastatic cancer. Of the 11 patients with benign strictures, 3 were secondary to chronic pancreatitis, 3 were postsurgical, 3 were inflammatory because of bile duct stones, 1 was secondary to autoimmune pancreatitis, and 1 was radiation-induced. All 16 patients with benign disease and/or PSC were alive at last follow-up. Both patients with bile duct IPMN underwent surgical resection with biliary bypass and were also doing well at long-term follow-up.
Operating characteristics In comparison with the final diagnosis, ROSE-TIC was truly positive for malignancy in 13 cases, truly negative in 16, falsely positive in 2, and falsely negative in none. www.giejournal.org
Varadarajulu et al
Diagnostic yield of single-operator cholangioscopy-guided biopsy
Figure 2. A, Touch imprint cytology revealing cholangiocarcinoma in a patient with proximal biliary stricture (Diff-Quik, orig. mag. 200). B, Corresponding off-site assessment of histological biopsy from the same site proving the diagnosis made by touch imprint cytology (H&E, orig. mag. 100).
Both false positive cases were interpreted as suspicious for malignancy but at final diagnosis were proven to be bile duct IPMN and autoimmune pancreatitis. Consequently, the sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy of ROSE-TIC were 100%, 88.9%, 86.7%, 100%, and 93.5%, respectively and were not significantly different from the operating characteristics of off-site histology (sensitivity 92.3%, specificity 100%, positive predictive value 100%, negative predictive value 94.7%, diagnostic accuracy 96.8%) (Table 3). There also was no significant difference between ROSE-TIC and off-site histology on receiver operating characteristic analysis (Fig. 3), with the area under the curve of 0.944 and 0.962, respectively (P Z .75). There was 100% concordance for both ROSE-TIC and off-site histologic findings between the 2 pathologists.
DISCUSSION This retrospective study is the first to demonstrate that by using ROSE in conjunction with TIC, the diagnostic sensitivity of SOC-guided biopsy of IDBSs can be significantly improved from the current 60% to 70% to >95%. The impact of ROSE on the practice of EUS-FNA has been significant.12 Several studies have shown that the presence of an onsite cytopathologist improves the diagnostic yield, decreases the number of inadequate or unsatisfactory samples, and limits the number of passes required to establish a diagnosis.13-15 ROSE has a significant impact even in challenging scenarios such as sampling pancreatic masses in the setting of chronic pancreatitis. In a prior study from our group of 300 patients with pancreatic masses of whom 75 had concomitant chronic pancreatitis, the sensitivity of EUS-FNA for detecting malignancy was 73.9% and 91.3%, respectively, for patients with and without chronic pancreatitis.16 In a similarly designed study from an expert center in Europe, the diagnostic sensitivity of EUS-FNA for malignancy was only 54% and 89%, respectively, for patients with and without concomitant chronic pancreatitis.17 The likely reason for the difference in diagnostic sensitivity between the 2 centers was the ability to limit the number of false-negative www.giejournal.org
diagnoses in patients with chronic pancreatitis by using ROSE in the former center. False-negative diagnoses are due to sampling errors, which include obtaining tissue from the wrong site or obtaining scant or uninterpretable tissue samples. These limitations can be minimized by immediate interpretation of the tissue sample and determining whether more sampling is needed until a preliminary diagnosis can be established. These observations with EUS-FNA gave us clues as to the potential impact that onsite assessment may have on SOC-guided biopsies of biliary strictures and served as the impetus for conducting the present study. The diagnostic sensitivity of SOC-guided biopsies is currently limited by specimen inadequacy and sampling error. If ROSE is made available, as with EUS-FNA, it has the potential, if confirmed in larger prospective studies, to reduce the rate of specimen inadequacy and also to minimize the likelihood of sampling error. However, the specimens obtained by SOC-guided microbiopsy forceps are small core tissue fragments and not cytologic aspirates that can be smeared easily onto slides. We therefore decided to use the TIC technique, which is a very simple procedure to perform onsite. The utility of TIC in EUS-guided tissue acquisition has been proven in a retrospective study of 109 patients in which the diagnostic accuracy of histology and TIC were 92.7% and 82.6% (not statistically significant), respectively.18 In the present study, we were able to procure adequate specimens in all 31 patients and did not encounter any sampling error. However, although we did not encounter any false-negative results, 2 patients suspected to have malignancy on ROSE-TIC were found to have bile duct IPMN and autoimmune pancreatitis at surgery. This particular pitfall is a well-recognized limitation that has been attributed to sampling error, cytopathologist misinterpretation, and translocated cell contamination.19,20 What lessons can we therefore learn from the study and how can these findings advance our management of patients with IDBS? First, performing ROSE-TIC is likely to significantly improve the diagnostic outcomes of SOCguided biopsy procedures. Even for centers that do not have access to cholangioscopy and rely only on performing fluoroscopy-guided intraductal biopsies, using the concept of ROSE-TIC is likely to yield better outcomes. Second, Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 5
Diagnostic yield of single-operator cholangioscopy-guided biopsy
Final diagnosis
1
Suspicious
Suspicious
Adenocarcinoma
2
Malignant
Adenocarcinoma
Adenocarcinoma
3
Malignant
Cholangiocarcinoma
Cholangiocarcinoma
4
Bile duct IPMN
Bile duct IPMN
Bile duct IPMN
5
Malignant
Adenocarcinoma
Cholangiocarcinoma
6
Atypical-reactive
Benign
PSC
7
Benign
Benign
Benign
8
Suspicious
Suspicious
Adenocarcinoma
9
Benign
Benign
Benign
10
Malignant
Adenocarcinoma
Cholangiocarcinoma
11
Benign
Benign
PSC
12
Benign
Benign
PSC
13
Malignant
Cholangiocarcinoma
Cholangiocarcinoma
14
Benign
Benign
PSC
15
Benign
Benign
Benign
16
Malignant
Cholangiocarcinoma
Cholangiocarcinoma
17
Suspicious*
Bile duct IPMN
Bile duct IPMN
18
Malignant
Poorly differentiated carcinoma
Poorly differentiated carcinoma
19
Malignant
Cholangiocarcinoma
Cholangiocarcinoma
20
Benign
Benign
Benign
21
Benign
Benign
PSC
22
Benign
Benign
Benign
23
Benign
Benign
Benign
24
Benign
Benign
Benign
25
Malignant
Ductal atypiay
Cholangiocarcinoma
26
Malignant
Adenocarcinoma
Adenocarcinoma
27
Benign
Benign
Benign
28
Malignant
Adenocarcinoma
Adenocarcinoma
29
Atypical-reactive
Atypical-reactive
Benign
30
Benign
Benign
Benign
31
Suspicious*
Benign
Benign
ROSE, Rapid on-site evaluation; TIC, touch implant cytology; IPMN, intraductal papillary mucinous neoplasm; PSC, primary sclerosing cholangitis. *False-positive result. yFalse-negative result.
most patients with nondiagnostic biliary brushings and/or intraductal biopsies are deemed to have indeterminate strictures and are referred to expert centers for additional workup. By facilitating a reliable, onsite diagnosis, ROSETIC is likely to decrease the number of reinterventions, thereby expediting patient care and resulting in substantial cost savings. Third, if a reliable diagnosis can be established by using ROSE-TIC in patients with biliary strictures, the technique is likely to obviate the need for EUS-FNA in a significant number of patients presenting with obstructive jaundice. This would be particularly beneficial for centers 6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
Operating characteristic (%, 95% CI) ROSE-TIC
Off-site histology
P value
Sensitivity
100 (75.3-100)
92.3 (64.0-99.8)
.308
Specificity
88.9 (65.3-98.6)
100 (81.5-100)
.146
Positive predictive value
86.7 (59.5-98.3)
100 (73.5-100)
.189
Negative predictive value
100 (79.4-100)
94.7 (74.0-99.9)
.352
Diagnostic accuracy
93.5 (78.6-99.2)
96.8 (83.8-99.9)
.554
CI, Confidence interval; ROSE-TIC, rapid on-site evaluation of touch imprint cytology.
1.00
Off-site histology
Sensitivity 0.50 0.75
ROSE-TIC
0.25
Case
TABLE 3. Comparison of operating characteristics between ROSE-TIC and off-site histology
0.00
TABLE 2. Correlation between ROSE-TIC, off-site histology, and final diagnosis
Varadarajulu et al
0.00
0.25
0.50 1-Specificity
ROSE-TIC, AUC: 0.944 Reference
0.75
1.00
Off-site histology, AUC: 0.962
Figure 3. Receiver operating characteristics analysis for rapid on-site evaluation of tissue imprint cytology and off-site histology. AUC, area under the curve; ROSE-TIC, rapid onsite evaluation-touch imprint cytology.
that do not have access to EUS-FNA because both diagnostic (tissue diagnosis) and therapeutic (biliary stenting) interventions can be undertaken in a single setting. Fourth, although our study was not designed to estimate the number of biopsies that may be required to maximize the diagnostic sensitivity, we needed to perform a mean of 3.3 biopsies per patient to achieve a high diagnostic sensitivity. We also observed that patients with jaundice due to extrinsic masses (pancreatic adenocarcinoma) required significantly more biopsies than patients with intrinsic lesions (cholangiocarcinoma) for establishing an onsite diagnosis of malignancy. Based on our experience, we recommend a minimum of 4 biopsies when SOCguided tissue acquisition procedures are performed. Our study has several limitations. First, this is a retrospective study with its inherent drawbacks. A randomized trial is currently in progress at our institution comparing onsite versus offsite techniques for evaluating biliary strictures by using SOC-guided biopsies (clinical trials registration no. NCT01815619). Second, although ROSETIC improves the diagnostic yield of SOC-guided biopsies, www.giejournal.org
Varadarajulu et al
the cost-effectiveness of this approach is unknown. It is unclear whether the technique must be adopted only for patients with prior negative workups or whether it should be routinely attempted in all patients at their index interventions. Third, given the small sample size, it was not possible to compare the diagnostic yield of ROSE-TIC by stricture location or to assess its operating characteristics by adjusting for other factors such as patient demographics, clinical presentation, or the presence of indwelling endoprostheses. Finally, because this study did not evaluate the utility of fluorescent in situ hybridization, optical coherence tomography, or probe-based confocal laser endomicroscopy in any of our patients, we could not compare the diagnostic performance of ROSETIC with these techniques. In conclusion, the present study demonstrates that the diagnostic sensitivity of SOC-guided biopsies of IDBSs can be significantly improved from the current 60% to >95% by using intraprocedural ROSE-TIC. Although it is not proven in this study, it is possible that this technique may benefit even centers that do not have access to cholangioscopy but rely mainly on fluoroscopy-guided intraductal biopsies. Randomized trials are in progress to confirm these promising observations.
Diagnostic yield of single-operator cholangioscopy-guided biopsy
7.
8.
9.
10.
11.
12.
13.
14.
15.
REFERENCES 1. Kalaitzakis E, Webster GJ. Endoscopic diagnosis of biliary tract disease. Curr Opin Gastroenterol 2012;28:273-9. 2. Rizvi S, Gores GJ. Pathogenesis, diagnosis, and management of cholangiocarcinoma. Gastroenterology 2013;145:1215-29. 3. Navaneethan U, Njei B, Lourdusamy V, et al. Comparative effectiveness of biliary brush cytology and intraductal biopsy for detection of malignant biliary strictures: a systematic review and meta-analysis. Gastrointest Endosc 2015;81:168-76. 4. Varadarajulu S, Eloubeidi MA. The role of endoscopic ultrasonography in the evaluation of pancreatico-biliary cancer. Surg Clin North Am 2010;90:251-63. 5. Weilert F, Bhat YM, Binmoeller KF, et al. EUS-FNA is superior to ERCPbased tissue sampling in suspected malignant biliary obstruction: results of a prospective, single-blind, comparative study. Gastrointest Endosc 2014;80:97-104. 6. Siddiqui AA, Mehendiratta V, Jackson W, et al. Identification of cholangiocarcinoma by using the Spyglass Spyscope system for peroral
www.giejournal.org
16.
17.
18.
19.
20.
cholangioscopy and biopsy collection. Clin Gastroenterol Hepatol 2012;10:466-71. Chen YK, Pleskow DK. SpyGlass single-operator peroral cholangiopancreatoscopy system for the diagnosis and therapy of bile-duct disorders: a clinical feasibility study (with video). Gastrointest Endosc 2007; 65:832-41. Navaneethan U, Hasan MK, Lourdusamy V, et al. Single-operator cholangioscopy and targeted biopsies in the diagnosis of indeterminate biliary strictures: a systematic review. Gastrointest Endosc 2015;82: 608-14. Cotton PB, Eisen G, Romagnuolo J, et al. Grading the complexity of endoscopic procedures: results of an ASGE working party. Gastrointest Endosc 2011;73:868-74. Eaton JE, Barr Fritcher EG, Gores GJ, et al. Biliary multifocal chromosomal polysomy and cholangiocarcinoma in primary sclerosing cholangitis. Am J Gastroenterol 2015;110:299-309. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under 2 or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837-45. Hebert-Magee S, Bae S, Varadarajulu S, et al. The presence of a cytopathologist increases the diagnostic accuracy of endoscopic ultrasoundguided fine needle aspiration cytology for pancreatic adenocarcinoma: a meta-analysis. Cytopathology 2013;24:159-71. Klapman JB, Logroño R, Dye CE, et al. Clinical impact of on-site cytopathology interpretation on endoscopic ultrasound-guided fine needle aspiration. Am J Gastroenterol 2003;98:1289-94. Iglesias-Garcia J, Dominguez-Munoz JE, Abdulkader I, et al. Influence of on-site cytopathology evaluation on the diagnostic accuracy of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) of solid pancreatic masses. Am J Gastroenterol 2011;106:1705-10. Alsohaibani F, Girgis S, Sandha GS. Does onsite cytotechnology evaluation improve the accuracy of endoscopic ultrasound-guided fine-needle aspiration biopsy? Can J Gastroenterol 2009;23:26-30. Varadarajulu S, Tamhane A, Eloubeidi MA. Yield of EUS-guided FNA of pancreatic masses in the presence or the absence of chronic pancreatitis. Gastrointest Endosc 2005;62:728-36. Fritscher-Ravens A, Brand L, Knöfel WT, et al. Comparison of endoscopic ultrasound-guided fine needle aspiration for focal pancreatic lesions in patients with normal parenchyma and chronic pancreatitis. Am J Gastroenterol 2002;97:2768-75. Vanderheyden AD, Proctor KA, Rizk MK, et al. The value of touch imprint cytology in EUS-guided Trucut biopsy. Gastrointest Endosc 2008;68:44-50. Mitsuhashi T, Ghafari S, Chang CY, et al. Endoscopic ultrasound-guided fine needle aspiration of the pancreas: cytomorphological evaluation with emphasis on adequacy assessment, diagnostic criteria and contamination from the gastrointestinal tract. Cytopathology 2006; 17:34-41. Fujii LL, Levy MJ. Pitfalls in EUS FNA. Gastrointest Endosc Clin N Am 2014;24:125-42.
Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 7
Improving the diagnostic yield of single-operator cholangioscopy-guided biopsy of indeterminate biliary strictures: ROSE to the rescue? (with video) Shyam Varadarajulu, MD,1,* Ji Young Bang, MD, MPH,2,* Muhammad K. Hasan, MD,1 Udayakumar Navaneethan, MD,1 Robert Hawes, MD,1 Shantel Hebert-Magee, MD1 Orlando, Florida, USA
Background and Aims: Tissue diagnosis, regardless of technique or endoscope used, can be challenging in patients with indeterminate biliary strictures (IDBSs). This exploratory study evaluated the utility and role of rapid onsite evaluation of touch imprint cytology (ROSE-TIC) when single-operator cholangioscopy (SOC)-guided biopsies of IDBSs are performed. Methods: Patients with IDBSs were evaluated by intraprocedural ROSE-TIC during SOC-guided biopsy procedures. Final diagnosis was established by long-term patient follow-up in conjunction with off-site findings or surgical histology. The main outcome measure was to evaluate the utility of ROSE-TIC by determination of its operating characteristics and comparison with off-site histologic assessment. Results: Of 31 patients with IDBSs, tissue diagnosis was indeterminate at prior ERCP-guided brush and/or biopsy in 14, prior EUS-guided FNA (EUS-FNA) in 6, and a mass could not be identified at EUS in 11. The mean number of biopsies performed was 3.3 (range 1-8), and diagnostic interpretation by ROSE-TIC was diagnostic and/or suspicious for carcinoma in 15, benign in 13, atypical-reactive in 2, and bile duct intraductal papillary mucinous neoplasm in 1. Final diagnosis by surgical histology (n Z 4), death by disease (n Z 10), and patient follow-up (n Z 17) showed that the overall sensitivity of ROSE-TIC for diagnosing malignancy was 100%, specificity 88.9%, positive predictive value 86.7%, negative predictive value 100%, and diagnostic accuracy 93.5%. Conclusions: Preliminary data suggest that the diagnostic outcomes of SOC-guided biopsies in IDBSs can be significantly improved by using ROSE-TIC. This technique also may benefit centers that rely mainly on fluoroscopy-guided intraductal biopsies. (Gastrointest Endosc 2016;-:1-7.)
Establishing an accurate diagnosis in a biliary stricture can be challenging.1,2 Although ERCP is the initial modality commonly used for evaluating biliary strictures, the diagnostic sensitivities of both brush cytology and intraductal biopsies are poor. In a recent meta-analysis, the pooled sensitivity and specificity of brush cytology for
the diagnosis of malignant biliary strictures were 45% and 99%, respectively.3 Likewise, the pooled sensitivity and specificity of intraductal biopsies were only 48.1% and 99.2%, respectively. Given the suboptimal results with ERCP-based techniques, EUS-FNA is increasingly being used as an alternative technique for tissue acquisition
Abbreviations: CBD, common bile duct; EUS-FNA, EUS-guided FNA; H&E, hematoxylin and eosin; IDBS, indeterminate biliary stricture; IPMN, intraductal papillary mucinous neoplasm; PSC, primary sclerosing cholangitis; ROSE, rapid onsite evaluation; SOC, single-operator cholangioscopy; TIC, touch imprint cytology.
0016-5107/$36.00 http://dx.doi.org/10.1016/j.gie.2016.03.1497
DISCLOSURE: S. Varadarajulu and R. Hawes are consultants for Olympus Medical Systems and Boston Scientific. J. Bang is a consultant for Olympus Medical Systems. S. Hebert-Magee is a consultant for Boston Scientific. All other authors disclosed no financial relationships relevant to this publication. *Drs Varadarajulu and Bang contributed equally to the article.
Received December 22, 2015. Accepted March 26, 2016. Current affiliations: Center for Interventional Endoscopy, Florida Hospital, Orlando, Florida (1), Divison of Gastroenterology-Hepatology, Indiana University, Indianapolis, Indiana (2). Reprint requests: Shyam Varadarajulu, MD, Medical Director, Center for Interventional Endoscopy, Florida Hospital, 601 East Rollins Street, Orlando, FL 32803. If you would like to chat with an author of this article, you may contact Dr Varadarajulu at [email protected].
Copyright ª 2016 by the American Society for Gastrointestinal Endoscopy
www.giejournal.org
Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 1
Diagnostic yield of single-operator cholangioscopy-guided biopsy
in patients with biliary strictures.4 In a recent study that compared EUS-FNA with ERCP-guided brush cytology and intraductal biopsies,5 the sensitivity of EUS-FNA was superior to ERCP tissue sampling for pancreatic masses (100% vs 38%) and comparable for biliary lesions (79% for both) and indeterminate strictures (80% vs 67%). However, hilar and proximal biliary tumors are sometimes difficult to visualize and sample using EUS-FNA.6 These technical limitations have led to the evolution of cholangioscopy-based techniques for evaluating the biliary ductal system.7 Single-operator cholangioscopy (SOC) not only enables the visualization of the biliary ductal system but also facilitates tissue sampling under direct vision.8 In a retrospective study of 30 patients with extrahepatic biliary strictures in whom both EUS-FNA and ERCPguided brush cytology were inconclusive, SOC-guided biopsies (Boston Scientific Corporation, Natick, Mass) yielded a diagnostic accuracy of 77% for cholangiocarcinoma.6 A recent meta-analysis of 10 studies evaluating 456 patients demonstrated pooled sensitivity and specificity of 60.1% and 98%, respectively, for SOC-guided biopsies of malignant biliary strictures.8 Reasons for the poor diagnostic sensitivity of cholangioscopy-guided biopsies are manifold. First, SOC is a technically difficult adjunct procedure to ERCP and is graded as level III by the American Society for Gastrointestinal Endoscopy.9 Visualization of biliary strictures and procurement of specimens by using intraductal biopsy forceps requires technical dexterity and considerable procedural expertise. Second, the “mini” biopsy forceps used in conjunction with SOC have small jaws that frequently yield inadequate specimens. Also, the optimal number of SOCguided biopsy specimens required to establish a definitive histologic diagnosis is undetermined. Third, most patients referred to tertiary-care centers for evaluation of an indeterminate biliary stricture (IDBS) have an indwelling endoprosthesis placed at outside facilities. These stents induce an inflammatory response that negatively impacts the microscopic evaluation of cellular architecture by inducing reactive atypia. Finally, the presence and extent of chromosomal abnormalities such as polysomy in the setting of preneoplastic diseases such as primary sclerosing cholangitis (PSC) can be multifocal, with varying levels of predisposition to cholangiocarcinoma.10 Consequently, the diagnostic yield of blind biopsies in PSC-induced biliary strictures can be low because of sampling errors. Given these limitations, patients with IDBSs often require multiple endoscopic or radiologic interventions before a diagnosis can be established reliably. The objective of this exploratory study was to evaluate the concept and diagnostic utility of rapid onsite evaluation of touch imprint cytology (ROSE-TIC) when SOC-guided biopsies of IDBSs are performed. Our hypothesis was that if a pathologist was available onsite for rendering preliminary interpretation, then the limitations of inadequate 2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
Varadarajulu et al
specimen procurement and sampling error could be minimized, leading to improved diagnostic performance.
MATERIALS AND METHODS Study design and patients This was a retrospective study of all patients who underwent SOC-guided biopsies of IDBSs over a 30-month period from April 2013 to October 2015. Patients were identified by querying our electronic ERCP database. Included in the study were patients aged 18 years referred for evaluation of IDBS by SOC. IDBS in this study was defined as patients with bile duct strictures who, on prior ERCP, had nondiagnostic intraductal biliary brushings and/or biopsy and/or nondiagnostic EUS-FNA cytology. Excluded were patients with acute cholangitis, acute pancreatitis, or coagulation disorders. For each case, patient demographics and clinical history such as prior attempts at tissue diagnosis, prior endoscopic instrumentation, procedural findings including stricture location, number of biopsies, postprocedural adverse events, onsite and final (off-site) pathology interpretation of bile duct biopsy specimens, postoperative surgical pathology reports, inpatient medical records, and outpatient clinical and radiologic follow-up information were reviewed. Per endoscopy unit policy, all ERCP patients were routinely contacted 5 to 7 days after the procedure by an endoscopy nursing coordinator to assess for adverse events. The study protocol was approved by the Florida Hospital institutional review board (IRB no. 828895).
SOC system and procedural techniques Direct Visualization System (SpyGlass, Boston Scientific) was used for performing cholangioscopy-guided biopsies from April 2013 to January 2015, and the more recent digital version (SpyGlass DS) was used from February to October 2015. All SOC procedures were performed with the patients under general anesthesia in the prone position, by 1 of 4 therapeutic endoscopists (S.V., R.H., M.H., U.N.). Antibiotic prophylaxis was administered at the discretion of the endoscopist. Endoscopic sphincterotomy was performed, or had been performed previously, before each examination. The cholangioscope was advanced through the accessory channel of the duodenoscope. Although a guidewire was used occasionally for the passage of the older SOC system into the bile duct, biliary cannulation was achieved in all cases by using the free-hand technique with the newer system. Biliary stricture dilation was performed as needed to facilitate the passage of the cholangioscope. In general, the instrument was passed proximally, suction was used to clear bile and contrast material, sterile saline solution was infused to optimize imaging, and the cholangioscope was slowly withdrawn to perform a systematic inspection of the ductal mucosa. Suspicion of malignant stricture on www.giejournal.org
Varadarajulu et al
Diagnostic yield of single-operator cholangioscopy-guided biopsy
cholangioscopy was based on the presence of “tumor vessels” characterized by dilated tortuous vasculature coursing through the epithelium accompanied by variable degrees of an exophytic mass protruding into the lumen of the bile duct. Tissue acquisition was undertaken by using the micro-biopsy forceps (SpyBite, Boston Scientific) (Fig. 1) (Video 1, available online at www.giejournal.org).
Specimen processing The biopsy specimens were processed for ROSE by using the touch imprint technique. The biopsy specimen was carefully pressed onto the slide, allowing the superficial cells to adhere, and then gently lifted with forceps, thereby creating a touch imprint of the specimen on the slide (Video 1). The slide was air dried and stained with Diff-Quik (Baxter Diagnostics Inc, McGaw Park, Ill) for ROSE. A single pathologist onsite analyzed the TIC specimens and provided real-time diagnosis that included benign, atypical-favor reactive, nondiagnostic-insufficient material, suspicious for malignancy, malignancy, and other diagnoses. Specimens were categorized as benign when the cytology sample did not reveal malignancy or any cellular atypia, atypical-favor reactive when the specimen displayed morphologic features beyond recognizable normal tissue components or only reactive changes, nondiagnostic-insufficient when technical or sampling limitations precluded the pathologist from providing any useful information from the specimen relative to the lesion sampled, suspicious when typical features of a specific malignant neoplasm were present but were qualitatively and/or quantitatively insufficient for a conclusive diagnosis, and malignant when the specimen unequivocally displayed malignant cytologic characteristics that included adenocarcinoma or its variants. Biopsy specimens were obtained until TIC was deemed diagnostic. After TIC, the bile duct biopsy specimens were placed in 10% buffered formalin solution for fixation and were stained by using hematoxylin and eosin (H&E). All specimens were examined by a single pathologist (S.H.M.) with particular expertise in cytopathology. The final diagnosis was certified by review of both ROSE-TIC and histology specimens by a second independent pathologist, who was blinded to the findings at ROSE.
Patient follow-up Final diagnosis of malignancy was based on tissue diagnosis that consisted of review of off-site histologic material procured from SOC-guided biopsy, surgical biopsy, or Whipple resection. In patients without surgical pathology confirmation or those with benign disease, clinical records and radiologic studies were reviewed to confirm disease stability, progression, or resolution. Patients suspected to have benign disease were followed for a minimum duration of 6 months.
Outcome measures The main outcome measure was to evaluate the utility of ROSE-TIC in SOC-guided biopsy of IDBS by www.giejournal.org
Figure 1. Single-operator cholangioscopy-guided biopsy of a bile duct stricture.
determination of its operating characteristics and comparison with off-site histologic assessment.
Statistical analysis Categoric variables were summarized as frequencies and percentages, and continuous variables as means with standard deviation (SD) or medians with range. Operating characteristics for ROSE-TIC and off-site histology for diagnosis of malignancy were calculated and then compared by using the chi-square test. Receiver operating characteristic analysis was performed to compare the areas under the curve for ROSE-TIC and off-site histology by using the nonparametric method.11 Suspicious for malignancy on ROSE-TIC was considered as being synonymous with malignancy for determination of operating characteristics. Statistical significance was set as P value < .05. The dataset was compiled by using Microsoft Excel (Microsoft Corporation, Redmond, Wash), and all statistical analyses were performed by using Stata 13 (StataCorp LP, College Station, Tex).
RESULTS Patient characteristics and SOC findings Of 59 patients who underwent SOC-guided biopsies of IDBSs over the 30-month period, 31 patients were evaluated with ROSE-TIC and constituted the study cohort. The remaining 28 specimens were evaluated by off-site histology because an onsite cytopathologist was not available and hence the data were excluded from analysis. The mean age of the 31 patients was 65.2 ( SD) 17.7 years, range 22-88 years. Twenty (64.5%) were male, and all patients presented with obstructive jaundice. Prior investigations included a nondiagnostic EUS in 17 patients (54.8%; no definitive mass on EUS in 11, non-diagnostic EUS-FNA in 6 patients) and ERCP with nondiagnostic tissue acquisition (brush cytology with and/or without intraductal biopsy) in 14 (45.2%). Eighteen of 31 (58.1%) strictures were located Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 3
Diagnostic yield of single-operator cholangioscopy-guided biopsy
in the common hepatic or proximal common bile duct (CBD), 5 (16.1%) in the mid-CBD, and 8 (25.8%) in the distal CBD. Twenty-six of 31 patients (83.9%) had an indwelling plastic biliary endoprosthesis in situ. No procedural adverse events were encountered (Table 1).
ROSE-TIC findings A mean of 3.3 biopsies (range 1-8) was performed in 31 patients. Of the 31 ROSE-TIC evaluations performed, 15 (48.4%) were categorized as malignant (11 carcinoma and 4 suspicious for carcinoma) and 16 (51.6%) as nonmalignant (13 benign, 2 atypical-reactive, and 1 bile duct intraductal papillary mucinous neoplasm [IPMN]). Five patients had PSC-induced biliary strictures, and 2 of the 7 patients with cholangiocarcinoma had underlying PSC. In 12 of the 15 cases (80.0%) that were categorized as malignant by ROSE-TIC, there was complete concordance between ROSE-TIC, off-site histology, and the final diagnosis (Figs. 2A and B). In 2 of the 3 discordant cases, cytology rendered suspicious by ROSE was later proven to be bile duct IPMN by both off-site and surgical histology in 1 patient (case no. 17) and autoimmune pancreatitis in another patient (case no. 31). In the third patient (case no. 25), categorized to have malignant stricture by ROSE-TIC, off-site histology revealed only ductal atypia but was ultimately proven to be a cholangiocarcinoma by surgical histology. Of the 16 patients deemed to have nonmalignant strictures by ROSE, there was complete concordance for all 16 cases between ROSE-TIC and the final diagnosis. Two cases of reactive atypia (defined as nonmalignant) at ROSE-TIC were proven later by both off-site histology and long-term follow-up to have extrahepatic PSC and postcholecystectomy ischemic stricture. One patient with biliary IPMN at ROSE-TIC also was confirmed by both offsite and surgical histology. A summary of the correlation between ROSE-TIC, off-site histology, and final diagnosis is shown in Table 2. Although the image quality was significantly better with the second generation SOC system, there was no significant difference in the mean number of biopsies required to establish a diagnosis between the first and second generation systems (3.6 vs 3.0; P Z .43). However, there was a significant difference in the mean number of biopsies required to establish a diagnosis between patients with cholangiocarcinoma versus pancreatic adenocarcinoma (2.6 vs 4.8; P Z .02).
Final diagnosis The final diagnosis was malignancy in 13 (adenocarcinoma in 5, cholangiocarcinoma in 7, poorly differentiated carcinoma in 1) and nonmalignant lesions in 18 (benign in 11, PSC in 5, bile duct IPMN in 2); the diagnosis was established by surgical histology in 4, death by disease progression in 10, and clinical and radiologic follow-up in 17 patients, at a median follow-up of 385 days (range 82-879 days) (Table 2). 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
Varadarajulu et al
TABLE 1. Patient details and lesion characteristics Case Age, y
Sex
Stricture location
Prior workup
Survival
1
69
Male
Proximal-CHD
No mass at EUS
Dead
2
87
Male
Proximal-CHD
No mass at EUS
Dead
3
80
Female
Proximal-CHD
No mass at EUS
Dead
4
81
Female
Middle
EUS-FNA
Alive
5
58
Female
Proximal-CHD
No mass at EUS
Dead
6
68
Male
Proximal-CHD
No mass at EUS
Alive
7
86
Male
Proximal-CHD
No mass at EUS
Alive
8
56
Male
Proximal-CHD
EUS-FNA
Dead
9
45
Female
Distal
ERCP-TA
Alive
10
75
Male
Proximal-CHD
EUS-FNA
Dead
11
43
Female
Proximal-CHD
ERCP-TA
Alive
12
59
Male
Proximal-CHD
ERCP-TA
Alive
13
78
Female
Proximal-CHD
ERCP-TA
Dead
14
22
Male
Distal
No mass at EUS
Alive
15
50
Male
Middle
ERCP-TA
Alive
16
57
Male
Proximal-CHD
ERCP-TA
Dead
17
79
Male
Proximal-CHD
ERCP-TA
Alive
18
78
Female
Proximal-CHD
EUS-FNA
Dead
19
82
Male
Distal
ERCP-TA
Alive
20
84
Male
Distal
ERCP-TA
Alive
21
49
Male
Distal
No mass at EUS
Alive
22
37
Male
Middle
ERCP-TA
Alive
23
43
Female
Proximal-CHD
No mass at EUS
Alive
24
59
Male
Proximal-CHD
ERCP-TA
Alive
25
72
Male
Middle
No mass at EUS
Dead
26
88
Female
Proximal-CHD
ERCP-TA
Alive
27
34
Male
Distal
ERCP-TA
Alive
28
74
Female
Proximal-CHD
No mass at EUS
Alive
29
79
Male
Distal
EUS-FNA
Alive
30
74
Female
Middle
ERCP-TA
Alive
31
74
Male
Distal
EUS-FNA
Alive
CHD, common hepatic duct; EUS-FNA, EUS-guided FNA; TA, tissue acquisition.
Of the 13 patients with malignancy, at last follow-up, 11 were dead and 2 were undergoing chemotherapy for metastatic cancer. Of the 11 patients with benign strictures, 3 were secondary to chronic pancreatitis, 3 were postsurgical, 3 were inflammatory because of bile duct stones, 1 was secondary to autoimmune pancreatitis, and 1 was radiation-induced. All 16 patients with benign disease and/or PSC were alive at last follow-up. Both patients with bile duct IPMN underwent surgical resection with biliary bypass and were also doing well at long-term follow-up.
Operating characteristics In comparison with the final diagnosis, ROSE-TIC was truly positive for malignancy in 13 cases, truly negative in 16, falsely positive in 2, and falsely negative in none. www.giejournal.org
Varadarajulu et al
Diagnostic yield of single-operator cholangioscopy-guided biopsy
Figure 2. A, Touch imprint cytology revealing cholangiocarcinoma in a patient with proximal biliary stricture (Diff-Quik, orig. mag. 200). B, Corresponding off-site assessment of histological biopsy from the same site proving the diagnosis made by touch imprint cytology (H&E, orig. mag. 100).
Both false positive cases were interpreted as suspicious for malignancy but at final diagnosis were proven to be bile duct IPMN and autoimmune pancreatitis. Consequently, the sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy of ROSE-TIC were 100%, 88.9%, 86.7%, 100%, and 93.5%, respectively and were not significantly different from the operating characteristics of off-site histology (sensitivity 92.3%, specificity 100%, positive predictive value 100%, negative predictive value 94.7%, diagnostic accuracy 96.8%) (Table 3). There also was no significant difference between ROSE-TIC and off-site histology on receiver operating characteristic analysis (Fig. 3), with the area under the curve of 0.944 and 0.962, respectively (P Z .75). There was 100% concordance for both ROSE-TIC and off-site histologic findings between the 2 pathologists.
DISCUSSION This retrospective study is the first to demonstrate that by using ROSE in conjunction with TIC, the diagnostic sensitivity of SOC-guided biopsy of IDBSs can be significantly improved from the current 60% to 70% to >95%. The impact of ROSE on the practice of EUS-FNA has been significant.12 Several studies have shown that the presence of an onsite cytopathologist improves the diagnostic yield, decreases the number of inadequate or unsatisfactory samples, and limits the number of passes required to establish a diagnosis.13-15 ROSE has a significant impact even in challenging scenarios such as sampling pancreatic masses in the setting of chronic pancreatitis. In a prior study from our group of 300 patients with pancreatic masses of whom 75 had concomitant chronic pancreatitis, the sensitivity of EUS-FNA for detecting malignancy was 73.9% and 91.3%, respectively, for patients with and without chronic pancreatitis.16 In a similarly designed study from an expert center in Europe, the diagnostic sensitivity of EUS-FNA for malignancy was only 54% and 89%, respectively, for patients with and without concomitant chronic pancreatitis.17 The likely reason for the difference in diagnostic sensitivity between the 2 centers was the ability to limit the number of false-negative www.giejournal.org
diagnoses in patients with chronic pancreatitis by using ROSE in the former center. False-negative diagnoses are due to sampling errors, which include obtaining tissue from the wrong site or obtaining scant or uninterpretable tissue samples. These limitations can be minimized by immediate interpretation of the tissue sample and determining whether more sampling is needed until a preliminary diagnosis can be established. These observations with EUS-FNA gave us clues as to the potential impact that onsite assessment may have on SOC-guided biopsies of biliary strictures and served as the impetus for conducting the present study. The diagnostic sensitivity of SOC-guided biopsies is currently limited by specimen inadequacy and sampling error. If ROSE is made available, as with EUS-FNA, it has the potential, if confirmed in larger prospective studies, to reduce the rate of specimen inadequacy and also to minimize the likelihood of sampling error. However, the specimens obtained by SOC-guided microbiopsy forceps are small core tissue fragments and not cytologic aspirates that can be smeared easily onto slides. We therefore decided to use the TIC technique, which is a very simple procedure to perform onsite. The utility of TIC in EUS-guided tissue acquisition has been proven in a retrospective study of 109 patients in which the diagnostic accuracy of histology and TIC were 92.7% and 82.6% (not statistically significant), respectively.18 In the present study, we were able to procure adequate specimens in all 31 patients and did not encounter any sampling error. However, although we did not encounter any false-negative results, 2 patients suspected to have malignancy on ROSE-TIC were found to have bile duct IPMN and autoimmune pancreatitis at surgery. This particular pitfall is a well-recognized limitation that has been attributed to sampling error, cytopathologist misinterpretation, and translocated cell contamination.19,20 What lessons can we therefore learn from the study and how can these findings advance our management of patients with IDBS? First, performing ROSE-TIC is likely to significantly improve the diagnostic outcomes of SOCguided biopsy procedures. Even for centers that do not have access to cholangioscopy and rely only on performing fluoroscopy-guided intraductal biopsies, using the concept of ROSE-TIC is likely to yield better outcomes. Second, Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 5
Diagnostic yield of single-operator cholangioscopy-guided biopsy
Final diagnosis
1
Suspicious
Suspicious
Adenocarcinoma
2
Malignant
Adenocarcinoma
Adenocarcinoma
3
Malignant
Cholangiocarcinoma
Cholangiocarcinoma
4
Bile duct IPMN
Bile duct IPMN
Bile duct IPMN
5
Malignant
Adenocarcinoma
Cholangiocarcinoma
6
Atypical-reactive
Benign
PSC
7
Benign
Benign
Benign
8
Suspicious
Suspicious
Adenocarcinoma
9
Benign
Benign
Benign
10
Malignant
Adenocarcinoma
Cholangiocarcinoma
11
Benign
Benign
PSC
12
Benign
Benign
PSC
13
Malignant
Cholangiocarcinoma
Cholangiocarcinoma
14
Benign
Benign
PSC
15
Benign
Benign
Benign
16
Malignant
Cholangiocarcinoma
Cholangiocarcinoma
17
Suspicious*
Bile duct IPMN
Bile duct IPMN
18
Malignant
Poorly differentiated carcinoma
Poorly differentiated carcinoma
19
Malignant
Cholangiocarcinoma
Cholangiocarcinoma
20
Benign
Benign
Benign
21
Benign
Benign
PSC
22
Benign
Benign
Benign
23
Benign
Benign
Benign
24
Benign
Benign
Benign
25
Malignant
Ductal atypiay
Cholangiocarcinoma
26
Malignant
Adenocarcinoma
Adenocarcinoma
27
Benign
Benign
Benign
28
Malignant
Adenocarcinoma
Adenocarcinoma
29
Atypical-reactive
Atypical-reactive
Benign
30
Benign
Benign
Benign
31
Suspicious*
Benign
Benign
ROSE, Rapid on-site evaluation; TIC, touch implant cytology; IPMN, intraductal papillary mucinous neoplasm; PSC, primary sclerosing cholangitis. *False-positive result. yFalse-negative result.
most patients with nondiagnostic biliary brushings and/or intraductal biopsies are deemed to have indeterminate strictures and are referred to expert centers for additional workup. By facilitating a reliable, onsite diagnosis, ROSETIC is likely to decrease the number of reinterventions, thereby expediting patient care and resulting in substantial cost savings. Third, if a reliable diagnosis can be established by using ROSE-TIC in patients with biliary strictures, the technique is likely to obviate the need for EUS-FNA in a significant number of patients presenting with obstructive jaundice. This would be particularly beneficial for centers 6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2016
Operating characteristic (%, 95% CI) ROSE-TIC
Off-site histology
P value
Sensitivity
100 (75.3-100)
92.3 (64.0-99.8)
.308
Specificity
88.9 (65.3-98.6)
100 (81.5-100)
.146
Positive predictive value
86.7 (59.5-98.3)
100 (73.5-100)
.189
Negative predictive value
100 (79.4-100)
94.7 (74.0-99.9)
.352
Diagnostic accuracy
93.5 (78.6-99.2)
96.8 (83.8-99.9)
.554
CI, Confidence interval; ROSE-TIC, rapid on-site evaluation of touch imprint cytology.
1.00
Off-site histology
Sensitivity 0.50 0.75
ROSE-TIC
0.25
Case
TABLE 3. Comparison of operating characteristics between ROSE-TIC and off-site histology
0.00
TABLE 2. Correlation between ROSE-TIC, off-site histology, and final diagnosis
Varadarajulu et al
0.00
0.25
0.50 1-Specificity
ROSE-TIC, AUC: 0.944 Reference
0.75
1.00
Off-site histology, AUC: 0.962
Figure 3. Receiver operating characteristics analysis for rapid on-site evaluation of tissue imprint cytology and off-site histology. AUC, area under the curve; ROSE-TIC, rapid onsite evaluation-touch imprint cytology.
that do not have access to EUS-FNA because both diagnostic (tissue diagnosis) and therapeutic (biliary stenting) interventions can be undertaken in a single setting. Fourth, although our study was not designed to estimate the number of biopsies that may be required to maximize the diagnostic sensitivity, we needed to perform a mean of 3.3 biopsies per patient to achieve a high diagnostic sensitivity. We also observed that patients with jaundice due to extrinsic masses (pancreatic adenocarcinoma) required significantly more biopsies than patients with intrinsic lesions (cholangiocarcinoma) for establishing an onsite diagnosis of malignancy. Based on our experience, we recommend a minimum of 4 biopsies when SOCguided tissue acquisition procedures are performed. Our study has several limitations. First, this is a retrospective study with its inherent drawbacks. A randomized trial is currently in progress at our institution comparing onsite versus offsite techniques for evaluating biliary strictures by using SOC-guided biopsies (clinical trials registration no. NCT01815619). Second, although ROSETIC improves the diagnostic yield of SOC-guided biopsies, www.giejournal.org
Varadarajulu et al
the cost-effectiveness of this approach is unknown. It is unclear whether the technique must be adopted only for patients with prior negative workups or whether it should be routinely attempted in all patients at their index interventions. Third, given the small sample size, it was not possible to compare the diagnostic yield of ROSE-TIC by stricture location or to assess its operating characteristics by adjusting for other factors such as patient demographics, clinical presentation, or the presence of indwelling endoprostheses. Finally, because this study did not evaluate the utility of fluorescent in situ hybridization, optical coherence tomography, or probe-based confocal laser endomicroscopy in any of our patients, we could not compare the diagnostic performance of ROSETIC with these techniques. In conclusion, the present study demonstrates that the diagnostic sensitivity of SOC-guided biopsies of IDBSs can be significantly improved from the current 60% to >95% by using intraprocedural ROSE-TIC. Although it is not proven in this study, it is possible that this technique may benefit even centers that do not have access to cholangioscopy but rely mainly on fluoroscopy-guided intraductal biopsies. Randomized trials are in progress to confirm these promising observations.
Diagnostic yield of single-operator cholangioscopy-guided biopsy
7.
8.
9.
10.
11.
12.
13.
14.
15.
REFERENCES 1. Kalaitzakis E, Webster GJ. Endoscopic diagnosis of biliary tract disease. Curr Opin Gastroenterol 2012;28:273-9. 2. Rizvi S, Gores GJ. Pathogenesis, diagnosis, and management of cholangiocarcinoma. Gastroenterology 2013;145:1215-29. 3. Navaneethan U, Njei B, Lourdusamy V, et al. Comparative effectiveness of biliary brush cytology and intraductal biopsy for detection of malignant biliary strictures: a systematic review and meta-analysis. Gastrointest Endosc 2015;81:168-76. 4. Varadarajulu S, Eloubeidi MA. The role of endoscopic ultrasonography in the evaluation of pancreatico-biliary cancer. Surg Clin North Am 2010;90:251-63. 5. Weilert F, Bhat YM, Binmoeller KF, et al. EUS-FNA is superior to ERCPbased tissue sampling in suspected malignant biliary obstruction: results of a prospective, single-blind, comparative study. Gastrointest Endosc 2014;80:97-104. 6. Siddiqui AA, Mehendiratta V, Jackson W, et al. Identification of cholangiocarcinoma by using the Spyglass Spyscope system for peroral
www.giejournal.org
16.
17.
18.
19.
20.
cholangioscopy and biopsy collection. Clin Gastroenterol Hepatol 2012;10:466-71. Chen YK, Pleskow DK. SpyGlass single-operator peroral cholangiopancreatoscopy system for the diagnosis and therapy of bile-duct disorders: a clinical feasibility study (with video). Gastrointest Endosc 2007; 65:832-41. Navaneethan U, Hasan MK, Lourdusamy V, et al. Single-operator cholangioscopy and targeted biopsies in the diagnosis of indeterminate biliary strictures: a systematic review. Gastrointest Endosc 2015;82: 608-14. Cotton PB, Eisen G, Romagnuolo J, et al. Grading the complexity of endoscopic procedures: results of an ASGE working party. Gastrointest Endosc 2011;73:868-74. Eaton JE, Barr Fritcher EG, Gores GJ, et al. Biliary multifocal chromosomal polysomy and cholangiocarcinoma in primary sclerosing cholangitis. Am J Gastroenterol 2015;110:299-309. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under 2 or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837-45. Hebert-Magee S, Bae S, Varadarajulu S, et al. The presence of a cytopathologist increases the diagnostic accuracy of endoscopic ultrasoundguided fine needle aspiration cytology for pancreatic adenocarcinoma: a meta-analysis. Cytopathology 2013;24:159-71. Klapman JB, Logroño R, Dye CE, et al. Clinical impact of on-site cytopathology interpretation on endoscopic ultrasound-guided fine needle aspiration. Am J Gastroenterol 2003;98:1289-94. Iglesias-Garcia J, Dominguez-Munoz JE, Abdulkader I, et al. Influence of on-site cytopathology evaluation on the diagnostic accuracy of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) of solid pancreatic masses. Am J Gastroenterol 2011;106:1705-10. Alsohaibani F, Girgis S, Sandha GS. Does onsite cytotechnology evaluation improve the accuracy of endoscopic ultrasound-guided fine-needle aspiration biopsy? Can J Gastroenterol 2009;23:26-30. Varadarajulu S, Tamhane A, Eloubeidi MA. Yield of EUS-guided FNA of pancreatic masses in the presence or the absence of chronic pancreatitis. Gastrointest Endosc 2005;62:728-36. Fritscher-Ravens A, Brand L, Knöfel WT, et al. Comparison of endoscopic ultrasound-guided fine needle aspiration for focal pancreatic lesions in patients with normal parenchyma and chronic pancreatitis. Am J Gastroenterol 2002;97:2768-75. Vanderheyden AD, Proctor KA, Rizk MK, et al. The value of touch imprint cytology in EUS-guided Trucut biopsy. Gastrointest Endosc 2008;68:44-50. Mitsuhashi T, Ghafari S, Chang CY, et al. Endoscopic ultrasound-guided fine needle aspiration of the pancreas: cytomorphological evaluation with emphasis on adequacy assessment, diagnostic criteria and contamination from the gastrointestinal tract. Cytopathology 2006; 17:34-41. Fujii LL, Levy MJ. Pitfalls in EUS FNA. Gastrointest Endosc Clin N Am 2014;24:125-42.
Volume
-,
No.
-
: 2016 GASTROINTESTINAL ENDOSCOPY 7