Single-operator peroral pancreatoscopy in the preoperative diagnostics of suspected main duct intraductal papillary mucinous neoplasms: efficacy and novel insights on complications

Background Distinguishing intraductal papillary mucinous neoplasms (IPMNs) from other pancreatic cystic lesions is essential since IPMNs carry the risk of becoming malignant. Differentiating the main pancreatic duct involving IPMNs (MD-IPMNs) through conventional imaging is deficient. Single-operator peroral pancreatoscopy (SOPP) represents a promising method offering additional information on suspected lesions in the pancreatic main duct (MD). We aimed to determine the role of SOPP in the preoperative diagnostics of suspected MD-IPMNs and identify factors contributing to SOPP-related complications. Materials and Methods In this primarily retrospective study, SOPPs were performed at three high-volume centers on suspected MD-IPMNs. Primary outcome was the clinical impact of SOPP to subsequent patient care. Additionally, we documented post-SOPP complications and analyzed several assumed patient- and procedure-related risk factors. Results One hundred and one (101) SOPPs were performed. Subsequent clinical management was affected due to the findings in 86 (85%) cases. Surgery was planned for 29 (29%) patients. A condition other than IPMN explaining MD dilatation was found in 28 (28%) cases. In 35 (35%) cases, follow-up with MRI was continued. Post-SOPP pancreatitis occurred in 20 (20%) patients and one of them was fatal. A decrease in odds of post-SOPP pancreatitis was seen as the MD diameter increases (OR 0.714 for 1.0 mm increase in MD diameter, CI 95% 0.514–0.993, p = 0.045). Furthermore, a correlation between lower MD diameter values and higher severity post-SOPP pancreatitis was seen (TJT = 599, SE = 116.6, z = − 2.31; p = 0.020). History of pancreatitis after endoscopic retrograde cholangiopancreatography was a confirmed risk factor for post-SOPP pancreatitis. Conclusions between complications and other risk factors could not be drawn. Conclusion SOPP aids clinical decision-making in suspected MD-IPMNs. Risk for post-SOPP pancreatitis is not negligible compared to non-invasive imaging methods. The risk for pancreatitis decreases as the diameter of the MD increases.

Although "fish-eye" papillae and oozing of mucus are visualized during duodenoscopy point towards IPMN, these signs are not seen consistently in all patients with MD-IPMN [22]. Endoscopic retrograde cholangiopancreatography (ERCP) is not recommended due to the risk of adverse events, alongside a lower sensitivity and specificity in identifying pancreatic cystic lesions compared to conventional imaging and EUS [10]. Single-operator peroral pancreatoscopy (SOPP), however, has shown promising results by adding visual inspection of the MD and the possibility of gathering visually guided biopsies from suspected areas [22][23][24][25][26][27][28].
Here, we aimed to establish the role of SOPP in the preoperative diagnostics of suspected MD-IPMNs. Secondarily, we aimed to identify factors contributing to SOPP-related adverse events.

Material and methods
In this multicenter cohort study, patients were collected from three high-volume endoscopic centers in Scandinavia. Although a few patients were prospectively included, this study is in essence retrospective. The study population consisted of all patients who underwent SOPP due to MD dilatation and suspected MD-IPMN or mixed-type IPMN (MT-IPMN) at Helsinki University Hospital (HUH), between 2012 and 2019, at Karolinska University Hospital (KUH), between 2015 and 2019, and at Skåne University Hospital (SUS), between 2017 and 2019. Between 2017 and 2019, seven additional patients were recruited prospectively at HUH, all of whom granted their written consent before the procedure. All patient data from KUH and SUS were collected retrospectively. This study received ethical approval from University of Helsinki (document number HUS/428/2017) and Etikprövningsmyndigheten (document number 2021:03989). This study is registered at ClinicalTrials.gov (Identifier: NCT03062124).
All patients were discussed at an MDT conference prior to SOPP. The patients had undergone MRI, CT, or both prior to the MDT conference. In all cases, imaging findings demonstrated cystic appearance of the pancreas and MD dilatation (≥ 5 mm) with the possibility of MD/MT-IPMN. However, consensus on further clinical management could not be reached during MDT conferences based on imaging findings, and further investigations were needed.
The following indications for SOPP were identified: (1) equal possibility for diagnosis other than MD/MT-IPMN that explains the cystic appearance of the pancreas and MD dilatation. Alternative diagnoses such as SB-IPMN, chronic pancreatitis, stricture of the MD, or pancreatic cystadenomas were suspected. A MDT conference recommended SOPP for excluding pathologies other than MD/MT-IPMN before planning surgery. (2) MD/MT-IPMN was highly likely, but the complication risk for pancreatic surgery was increased due to comorbidities such as cardiovascular diseases, prior liver transplant, and abnormal blood coagulation. Some patients were indecisive about surgery. A MDT conference recommended SOPP to gain visual confirmation of MD/ MT-IPMN and biopsies on target lesions. Surgery would be reconsidered, if the specimens retrieved provided evidence on high-grade dysplasia (HGD) or malignancy. (3) Pancreatic surgery was indicated based on imaging findings. Due to comorbidities the patient was not eligible for total pancreatectomy. However, a pancreaticoduodenectomy or a distal pancreatic resection would be possible. A MDT conference recommended SOPP to evaluate extension of IPMN lesions in MD in order to make a decision on surgery.
The equipment used for SOPPs primarily relied on Spy-Glass DS Direct Visualization System (SpyGlass DS; Boston Scientific, USA). However, 15 (15%) procedures in the early years of this study were performed using older SpyGlass (Boston Scientific, USA) equipment. Indefinite tumorous alterations were examined and biopsied (SpyBite™) as an indicative of MD-IPMN, and the extension of these lesions in the MD (tail, body, neck or head) was noted. Irrigation fluid and brush cytology samples from the MD were collected whenever suitable. All procedures were performed by experienced specialists.
As a key outcome, we determined how often the visual appearance of MD and/or MD irrigation fluid samples and biopsies taken had any impact on further clinical management. Additionally, we collected data from medical charts regarding adverse events (Cotton consensus criteria) [29], diameter of the MD (MRI scan), history of pancreatitis or prior post-ERCP pancreatitis, prophylactic usage of nonsteroidal anti-inflammatory drugs (NSAIDs), and pancreatic stents. Procedure-related data, such as time consumption, timing of the pancreatic sphincterotomy (PS), papillary, or MD orifice balloon dilatation, and occurrence of parallel therapeutic interventions performed [e.g., radiofrequency ablation (RFA) and electrohydraulic lithotripsy (EHL)] were collected. Data on general patient characteristics such as gender, age, the American Society of Anesthesiologists' physical status classification (ASA), serum carbohydrate antigen 19-9 levels (s-Ca19-9), and body mass index (BMI) were registered.

Statistical analysis
Data are reported as the number of cases and as percentages for categorical variables. For continuous variables with a normal distribution, mean and standard deviation (SD) are presented. For continuous variables with non-normal distribution, median and range are presented. Fisher's exact test was used to evaluate the statistical significance of nominal or dichotomous results. The Shapiro-Wilks test was used to test the normality of the distribution of continuous variables. The Mann-Whitney U-test was used to calculate the differences in non-normally distributed continuous variables. We used logistic regression to analyze the relationships between dependent dichotomous variables and independent variables. Jonckheere-Terpstra test was used to determine the trends between ordinal independent variables and continuous dependent variables. Two-sided p values < 0.05 were considered significant. All statistical analyses were performed using IBM's SPSS Statistics (Armonk, NY, USA).

Results
In total, 101 SOPPs were performed on 84 patients (14 patients had two or more SOPPs). 30 (30%) were performed at HUH, 69 (68%) at KUH, and 2 (2.0%) at SUS. The median time from the index SOPP until data analysis was 3 years (range 1-8 years). Patient and procedural characteristics are presented in Tables 1 and 2.

Impact on further patient care
All patients were discussed at a MDT conference following the SOPP. Of the procedures, 86 (85%) provided new information benefiting the patient and impacting further care. Indications and findings on SOPPs, and their impact on further clinical management are presented in Table 4.
For 35 (35%) patients, follow-up with MRI was recommended based on findings in SOPP. In 29 (29%) SOPPs, the MD was seen disease-free and a SB-IPMN diagnosis was set. These patients had only benign or LGD findings in irrigation fluid samples. In four patients with SB-IPMN, SOPP discovered a coexisting condition (chronic pancreatitis n = 2, squamous cell metaplasia n = 1, and papillary neuroendocrine tumor (NET) n = 1). Based on SOPP findings, MDT conference recommended close follow-up with MRI to additional 6 (5.9%) patients with MD-IPMN polyps seen in MD. These patients had a high risk for surgical complications and only benign or LGD findings in biopsies and irrigation fluid samples retrieved during SOPP.
Other causes for MD dilatation were discovered in 28 (28%) patients. These included pancreatic MD stones (n = 10), chronic strictures caused by chronic pancreatitis or prior stenting (n = 16), one papillary neuroendocrine tumor (NET), and one squamous cell metaplasia. Of these patients, four were diagnosed with concurrent SB-IPMN, and for them, surveillance with MRI continued. Based on findings on SOPP, recurrent surveillance with MRI was terminated in 24 patient cases. However, of those patients, 2 (2.0%) were diagnosed with pancreatic cancer within the following 4 months after SOPP. For these two patients, SOPP was considered non-beneficial.
For 7 (6.9%) patients, the SOPP was considered as a failure and therefore not contributable to the final decision [failure to cannulate into the pancreatic MD (n = 5)    and inconclusive findings (n = 2)]. Furthermore, 6 patients (5.9%) with MD-IPMN findings in SOPP refused or were unfit for surgery. A pancreaticoduodenectomy was the most common procedure performed (n = 19, 66%) on patients who underwent surgery. The procedures performed are shown in Table 5. Among the 29 operated patients, visual inspection during SOPP correctly identified the intraductal pathology in 26 (90%) cases. In eight operated patients, SOPP suggested HGD or malignancy in biopsies or cytologic samples. Of these patients, 6 (75%) had HGD or malignancy in the final histopathology report. In one case, biopsies taken during SOPP failed to present HGD later discovered in the surgical specimen. The final pathological anatomic diagnosis (PAD) following surgery and their corresponding findings in SOPP appear in Table 6.

Visual appearance of MD
Optical assessment of intraductal pathological findings could be obtained in 47 (47%) patients (n = 37 MD-IPMN, n = 10 pancreatic MD stones), and in another 37 (37%) patients, the MD was free from any abnormalities. Moreover, visual appearance is also a prerequisite for taking directed biopsies from pathological findings, including non-tumorous lesions such as chronic pancreatitis and iatrogenic strictures caused by previous stents. Table 7 summarizes the details for samples taken during SOPP.

Brush cytology samples
Brush cytology samples were collected during 51 (50%) SOPPs. Typically, multiple samples were taken from each patient. Brush cytology samples confirmed 5 (9.8%) cases of malignant cells and 17 (33%) cases of atypia. In the remaining samples, the finding was benign or insufficient (see Table 7 for details). Fluorescence in situ hybridization (FISH) was performed on the brush cytology samples of three (3.0%) patients. Of these three cases, one demonstrated no aneuploidy. Another one revealed aneuploidy with diploidy of the chromosomes 3, 7, 17, and the chromosome locus 9p21. FISH failed in the third case due to too few cells.

Irrigation fluid samples
A total of 57 irrigation fluid samples were collected. Among these samples, 4 (7.0%) showed signs of a malignant or HGD disease and 11 (19%) samples indicated atypia or LGD. FISH was performed on one patient's irrigation fluid sample. However, it failed due to insufficient number of cells. From the irrigation fluid samples, 18 (32%) samples contained mucus (see Table 7 for details).
Among the 10 patients with parallel therapeutic interventions, five adverse events occurred to 3 (30%) patients [mild pancreatitis after RFA (n = 1), mild pancreatitis and bleeding after EHL + DASE (n = 1), and severe pancreatitis and pancreatic duct bleeding after RFA (n = 1)]. Of the patients who were not subjected to any additional interventions, 17 (18%) experienced a procedural associated complication. No statistically significant difference (p = 0.39) between the groups was found.

Diameter of the main pancreatic duct
A Mann-Whitney U-test indicated that the median MD diameter for patients without post-SOPP pancreatitis (median = 8.0 mm) was significantly greater than for patients with post-SOPP pancreatitis (median = 6.5 mm, p = 0.021) ( Table 8).
Analyzed as a continuous variable in logistic regression, a statistically significant decrease in odds ratio (OR) of post-SOPP pancreatitis was observed as the MD diameter increases (OR 0.714 for 1.0 mm increase in MD diameter, CI 95% 0.514-0.993; p = 0.045). Using the median value for MD diameter (median = 7.0 mm) as a cut-off value, logistic regression test showed a lower probability for post-SOPP pancreatitis if the MD diameter is ≥ 7.0 mm (OR 0.334, CI 95% 0.120-0.928; p = 0.035). A Jonckheere-Terpstra test for ordered alternatives showed a statistically significant trend suggesting that lower median MD diameter value correlates with higher severity (scale from "mild", "moderate" to "severe") post-SOPP pancreatitis (T JT = 599, SE = 116.6, z = − 2.31; p = 0.020). Table 9 summarizes median MD diameters and severity of pancreatitis.

History of acute or chronic pancreatitis
In 39 (39%) patient cases, there was a history of acute pancreatitis (n = 14, 14%), chronic pancreatitis (n = 22, 22%), or both (n = 3, 3.0%). Of the SOPPs performed on patients with prior acute pancreatitis, 2 (12%) led to mild pancreatitis, while 18 (21%) SOPPs performed on patients without prior acute pancreatitis resulted in a complication (p = 0.513). Of the patients with history of chronic pancreatitis, 6 (24%) suffered from post-SOPP pancreatitis. Of the complications recorded, four were graded mild and the other two were graded moderate or severe. Meanwhile, 14 (18%) of the SOPPs performed on patients without a history of chronic pancreatitis led to post-SOPP pancreatitis (p = 0.569) ( Table 10).

History of post-ERCP pancreatitis
One or more prior ERCPs had been performed on 55 (54%) patients. Of those 55 patients, 12 (22%) had a history of post-ERCP pancreatitis. Of these patients with prior post-ERCP pancreatitis, 5 (42%) suffered from post-SOPP pancreatitis. Meanwhile 4 (9.3%) patients with prior ERCP, but without prior complications, had post-SOPP pancreatitis (p = 0.017). A moderate or severe post-SOPP pancreatitis occurred 3 (25%) times to patients with history of post-ERCP pancreatitis. No cases of moderate or severe post-SOPP pancreatitis were observed in patients with prior ERCP without complications (p = 0.008) ( Table 10).

Papillary or MD orifice balloon dilatation
To enable entry of the SOPP device into the MD, balloon dilatation of the papillary region was performed on 11 (11%) patients. Balloons with a diameter from 4 to 6 mm were used. Of the patients with balloon dilatation, 3 (27%) experienced post-SOPP pancreatitis. Of the SOPPs performed without balloon dilatation, 17 (19%) resulted in post-SOPP pancreatitis (p = 0.452) ( Table 10).

Procedure time
The duration of the procedure was recorded from the beginning of the duodenoscopy through the removal of SOPP equipment from the pancreatic duct. Please note that the time includes both conventional ERCP and SOPP. The Mann-Whitney U-test revealed no statistically significant difference between procedures followed by post-SOPP complications and those with no respective complications based on the duration of the procedure (see Table 11 for details). Please note that the procedure time for six patients was unknown. Therefore, this analysis included 95 SOPPs only.

Discussion
In 86 (85%) presumed IPMN cases, SOPP yielded additional diagnostic information benefiting the patient and impacting clinical decision-making. Previous studies have reported 64% to 77% of cholangiopancreatoscopies to carry clinical value [22,30,31]. The higher impact rate in this study is explained by patient selection, which differed compared to previous studies. The cohort in this study consisted not only of patients with uncertain IPMN findings, but also of patients with probable MD-IPMN, but whose comorbidities limited options for surgical treatment. We managed to demonstrate that SOPP aids clinical decision-making also in the latter patient group.
Previous studies have shown that collecting pancreatic juice with SOPP is feasible, and pancreatic juice cytology carries a diagnostic value in identifying IPMNs [28]. Studies also indicate a high specificity for detecting malignancy with biopsy specimens and irrigation fluid samples taken during SOPP [25,27]. In this study, we identified three cases where the MD was clear from IPMN during SOPP, although irrigation fluid samples suggested IPMN disease with HGD or malignant features. In addition, visually guided biopsies taken from lesions identified in an MD and/or brush cytology samples retrieved from an MD revealed seven cases of malignancy or HGD. In this study, 75% of surgically treated patients displaying HGD or malignancy in biopsies or cytologic samples retrieved during SOPP had HGD or malignant findings in the final pathologist's report. However, in one case, samples taken during SOPP failed to present HGD discovered later in the surgical specimen. In addition, there were two cases where SOPP failed to detect a malignancy diagnosed within 4 months of the procedure. Our study suggests that visually guided biopsies and irrigation fluid samples contribute to the diagnostic value of SOPP although risk for false negatives should be kept in mind.
The complication rate in this study was 20%. Previous studies found total complication rates of 7 to 13% for peroral  cholangiopancreatoscopies [32][33][34]. In our study, patients underwent pancreatoscopy, and no cholangioscopies were performed. A study addressing only pancreatoscopies showed a post-SOPP pancreatitis rate of 17% [22]. Another study suggested that pancreatoscopy was associated with a higher adverse events rate of 19.8% as compared to 9.6% for cholangioscopy [31]. In this study, parallel therapeutic interventions performed during SOPPs contribute to the total complication rate (complication rate 30% versus 18% without therapeutic interventions, p = 0.390). Although the finding lacks statistical significance, it should be noted that all bleeding complications (n = 2) and 33% (n = 3) of all severe and moderate types of pancreatitis we observed occurred following SOPPs with parallel therapeutic interventions.
Overall, the total complication rate in this study did not differ from previous studies on peroral pancreatoscopy. A normal MD measures 1.5 to 3.5 mm in diameter [35]. In order to enter the MD, it needs to be enlarged, since the diameter of the SOPP device is approximately 3.3 mm (10 Fr) [36]. In this study, no SOPPs were performed on patients whose MD diameter fell below 5 mm. We demonstrated a decrease in odds of post-SOPP pancreatitis as the MD diameter increases (OR 0.714 for 1.0 mm increase in MD diameter, CI 95% 0.514-0.993; p = 0.045). A lower probability for post-SOPP pancreatitis was demonstrated, when the MD diameter is ≥ 7.0 mm (OR 0.334, CI 95% 0.120-0.928; p = 0.035). Furthermore, we were able to establish a trend suggesting that lower median MD diameter value correlates with higher severity post-SOPP pancreatitis (T JT = 599, SE = 116.6, z = − 2.31; p = 0.020). Although suggested by Reuterwall et al. [31], no studies exist which address the connection between the diameter of the MD and post-SOPP pancreatitis. When planning an SOPP for a patient with MD diameter under 7.0 mm, the risk for post-SOPP pancreatitis should be carefully considered.
Multiple meta-analyses suggest that using prophylactic NSAIDs prior to conventional ERCP results in a lower post-ERCP pancreatitis rate [37][38][39]. Studies concerning prophylactic NSAID use specifically prior to SOPP remain scarce. In this study, patients administered with prophylactic NSAIDs experienced a lower post-SOPP pancreatitis rate compared to patients without prophylactic NSAIDs (n = 2 (8.0%) vs n = 18 (24%); p = 0.146). Due to a low number of patients, definitive conclusions cannot be made. However, the results agree with the recent studies concerning post-ERCP complications [37][38][39] and current guidelines on ERCP-related adverse events [40].
Current guidelines recommend prophylactic pancreatic stenting only in select patients at a higher risk for post-ERCP pancreatitis [40]. Furthermore, it has been suggested that prophylactic pancreatic stenting among patients with IPMN might even increase the risk for post-ERCP pancreatitis [41]. In this study, the difference in post-SOPP complication rates between patients with a prophylactic pancreatic stent versus patients without pancreatic stents was not significant (n = 7 (16%) vs n = 13 (23%); p = 0.456). The role of prophylactic pancreatic stenting remains controversial.
Due to the size of the SOPP device, a generous PS is typically needed in order to enter the pancreatic duct safely [42]. To our knowledge, no comprehensive studies have addressed the issue of whether there is a difference in adverse events rates, if the PS is performed simultaneously as SOPP versus previously during a different procedure. Scattered studies report individual cases of post-SOPP pancreatitis among both patients with prior PS and with PS performed simultaneously with SOPP [43,44]. Interestingly, in this study, patients with prior PS exhibited a lower rate of moderate or severe post-SOPP pancreatitis compared with patients who underwent PS simultaneously with SOPP (n = 3 (5.6%) vs n = 7 (18%); p = 0.087). However, this finding was not statistically significant.
Prior pancreatitis and post-ERCP pancreatitis are known risk factors for post-ERCP pancreatitis [45]. The role of papillary balloon dilation is controversial [46]. In this study, a statistically significant difference was seen in post-SOPP pancreatitis rates between patients with prior post-ERCP pancreatitis (n = 5, 42%) and patients with prior ERCP without pancreatitis (n = 4, 9.3%, p = 0.017). Our results on the subject are similar to prior findings [45]. However, a statistically significant connection between post-SOPP pancreatitis and history of acute or chronic pancreatitis, or balloon dilatation of the papilla or MD orifice could not be reached. Essentially, this first Scandinavian observational study demonstrates unambiguously the benefits of SOPP in preoperative diagnostics of suspected MD-IPMNs. However, an even larger study material should be collected to identify and confirm additional factors contributing to post-SOPP complications. In addition, we note several limitations to our study. First, the study material was collected from three different centers consisting of patients treated at different time periods. During these time periods, two different types of SOPP instrumentation were used. Second, guidelines concerning the management of MD dilatation, the use of prophylactic NSAIDs, and prophylactic stents have changed during these time periods. Furthermore, SOPPs with parallel therapeutic interventions that might contribute to the complication rate were performed in one study center only.
To conclude, patients with a dilated MD and suspected MD-IPMN lesions, but with no clear decision on subsequent care, benefit from SOPP. SOPP yields additional diagnostic information and has the potential to reduce needless surgery or unnecessary repeated surveillance with MRI within selected patients. However, the risk for post-procedural complications, especially pancreatitis is not negligible. To reduce the risk of post-SOPP complications, IPMN diagnostics with SOPP should be performed in large volume endoscopic centers with high expertise and comprehensive experience on SOPPs. Patients referred to SOPP due to indefinite IPMN findings should be discussed beforehand in a MDT conference to define the indication and expected benefit of the procedure. Patients fit for surgery with definite imaging findings supporting MD-IPMN diagnosis should be referred to surgery without performing SOPP. History of post-ERCP pancreatitis is a risk factor for post-SOPP pancreatitis. The odds for developing post-procedural pancreatitis decreases as the diameter of the MD increases. Lower values of MD diameter correlate with higher severity of post-SOPP pancreatitis. Although statistical significance was not reached here, parallel therapeutic interventions performed during SOPP might increase the risk for complications. Furthermore, the rate for moderate and severe complications might decrease when PS is performed separately from the actual SOPP, and when prophylactic NSAIDs are administered. However, these findings are not statistically significant. Finally, the use of pancreatic stents remains controversial.
Funding Open access funding provided by Umea University.

Declarations
Disclosures Roberto Valente received a grant from Boston Scientific for lecturing at the Finnish Pancreatic Club 2021. Urban Arnelo is consultant to Ambu and Boston Scientific and has received grants for investigator initiated and industry-sponsored studies from Boston Scientific. Sini Vehviläinen, Niklas Fagerström, Hanna Seppänen, Marianne Udd, Outi Lindström, Harri Mustonen, Fredrik Swahn, and Leena Kylänpää declare no conflict of interest or financial ties to disclose.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.