Technical success, clinical efficacy, and safety of endoscopic minor papilla interventions for symptomatic pancreatic diseases

Abstract

BACKGROUND

Endoscopic minor papilla intervention (EMPI) is an option for diagnosing or treating symptomatic pancreatic diseases in cases with failed pancreatic duct deep cannulation via the major papilla, pancreas divisum with obstruction of the minor papilla, or an abnormal patulous orifice of the minor papilla during endoscopic retrograde cholangiopancreatography (ERCP). However, the relatively low patency and small opening of the minor papillae pose technical challenges.

AIM

To evaluate the technical success, clinical success, stone clearance, and safety profile of EMPI for diagnosis and treatment of symptomatic pancreatic diseases.

METHODS

Patients diagnosed with symptomatic pancreatic diseases and EMPI between February 1996 and February 2023 were included. The primary outcomes were the initial technical success, defined as successful deep cannulation via the minor papilla (DCMP; access of the guidewire to the upstream pancreatic duct via the minor papilla) alone, overall technical success, defined as successful DCMP alone and successful DCMP with additional needle-knife precut minor papillotomy (NKPMP), and immediate clinical success, defined as > 50% improvement in abdominal pain after therapeutic EMPI. Secondary outcomes included long-term clinical success at 1, 3, and 7 years, pancreatic stone clearance, and procedure-related early and late adverse events (AEs).

RESULTS

Overall, 43 patients, 32 with obstructive chronic pancreatitis, 8 with pancreatic divisum, and 3 with intraductal papillary mucinous neoplasm were included. The initial and overall technical success rates were 74.4% (32/43) and 88.4% (38/43), respectively. The immediate clinical success rate was 79.1% (34/43), and the long-term clinical success rates at 1, 3, and 7 years were 74.7%, 55.3%, and 41.5%, respectively, among the 22 patients with a follow-up period of 57.5 (7-266) months. Complete and partial success of pancreatic stone clearance was achieved in 53.9% (7/13) and 15.4% (2/13), respectively. Early AEs included post-ERCP pancreatitis (PEP, n = 5) and self-limiting bleeding (n = 1); surgery therapy was required for 1 case with severe PEP and conservative management for the other 4 with mild PEP. Late AEs included minor papilla stricture (n = 1) after endoscopic minor papillotomy and pancreatic duct stricture (n = 1) after double pancreatic stent placement; no specific treatment was implemented for these events.

CONCLUSION

EMPI is feasible, effective, and safe for symptomatic pancreatic diseases, in terms of the technical and clinical success, stone clearance, and incidence and severity of AEs. NKPMP appears to enhance technical success. However, potential risks of developing PEP and late AEs should be kept in mind.

Key Words: Endoscopic minor papilla intervention; Symptomatic pancreatic diseases; Obstructive chronic pancreatitis; Pancreas division; Abdominal pain improvement

Core Tip: The study evaluated the technical success, clinical success, stone clearance, and safety of endoscopic minor papilla intervention (EMPI) for symptomatic pancreatic diseases. Initial and overall technical success rates for deep cannulation via the minor papilla were 74.4% and 88.4%, respectively. Immediate clinical success (pain improvement) was 79.1%, and the long-term clinical success rates at 1, 3, and 7 years were 74.7%, 55.3%, and 41.5%, respectively. Stone clearance was achieved in 69.2% of cases. Post-endoscopic retrograde cholangiopancreatography pancreatitis was a main early adverse event, developed in 11.6% of cases. EMPI is a feasible, effective, and safe modality for symptomatic pancreatic diseases.

INTRODUCTION

Obstructive chronic pancreatitis (CP) develops as a secondary complication caused by an obstruction in the main pancreatic duct (MPD) due to a stone, stricture, and/or localized mass or by an obstruction in the sphincter or minor papilla mainly in the presence of pancreas divisum[1]. The obstruction results in upstream pancreatic duct dilation, secondary acinar cell atrophy, and fibrosis[2] and consequently increases pressure within the pancreatic duct and pancreatic parenchyma, which further leads to pancreatic ischemia and is a major cause of recurrent or chronic abdominal pain[3,4]. At present, several national guidelines[1,2,5,6] recommend endoscopic retrograde cholangiopancreatography (ERCP), a minimally invasive therapeutic procedure, is utilized as the first-line treatment for pancreatic diseases including obstructive CP (OCP), whereas surgery intervention is often reserved as the final treatment choice if endoscopic approaches is unsuccessful[1]. Deep cannulation of the MPD through the major duodenal papilla is typically required for ERCP treatment of OCP, but it is technically challenging and often unsuccessful in symptomatic cases of OCP with complete downstream obstruction of the MPD caused by a stone or stricture. For these cases, as well as those with pancreas divisum who develop abdominal pain due to obstruction of minor papilla, and those with intraductal papillary mucinous neoplasm (IPMN), which causes a fish eye-like appearance or enlargement of the minor papilla, endoscopic minor papilla intervention (EMPI) through ERCP is a diagnostic and therapeutic option[7-9]. During an EMPI procedure, a therapeutic intervention, also called minor papilla endotherapy (MPE), such as endoscopic minor papillotomy (EMP), stricture dilation, stenting, and/or stone therapy, or a diagnostic intervention[8], such as peroral pancreatoscopy or intraductal ultrasound (IDUS) via the minor papilla is performed. However, the minor papilla has a smaller opening with a lower patency rate (ranging from 12% to 82%), compared with the major papilla[7], posing technical challenges for EMPI. Although the injection of secretin has been utilized to enhance the success rate of cannulation by aiding in the observation of papillary opening and pancreatic fluid drainage for easier cannulation, nearly one-third of patients (31.6%) experience unsuccessful cannulation[10]. Thus, it is recommended that in cases where the deep cannulation via the minor papilla (DCMP) procedure is unsuccessful, the free-hand needle-knife precut minor papillotomy (NKPMP) technique be used[11,12], followed by the pull-type sphincterotome to complete the EMP[13]. NKPMP, especially if the precutting fails, is associated with a risk of developing post-ERCP pancreatitis (PEP) in patients with complete pancreas divisum where the ventral duct is not communicated with the dorsal duct or those with "pseudodivisum" (mimicking a complete pancreas divisum[7]), which is associated with complete obstruction in the downstream MPD, as local edema of the orifice of the minor papilla caused by the NKPMP procedure could obstruct the only pancreatic drainage pathway.

EMPI is well documented for its application in cases with pancreas divisum, a congenital variation of the pancreas[13]. The number of symptomatic OCP cases treated with EMPI is relatively smaller than that of cases with pancreas divisum[11]. For symptomatic pancreas divisum, EMP followed by short-term placement of 5 Fr prophylactic pancreatic duct stents (PPS) is usually sufficient to relieve minor papilla obstruction and prevent PEP or early papilla stenosis; however, balloon dilation of the minor papilla is associated with a higher risk of adverse events (AEs) including PEP, even with PPS placement[12]. By contrast, EMPI for OCP usually requires stone fragmentation/retrieval in the upstream MPD after EMP and/or balloon dilation of the minor papilla or stenting to bypass the stone in the downstream MPD and accessory pancreatic duct (APD) stent placement to treat the strictures and pseudocyst drainage, which increases the complexity of treatment. Currently, there is no consensus on the use of NKPMP and balloon dilation in EMPI. Additionally, whether the different morphologies of long and short-type APD affect the success rate of DCMP is not yet fully understood. We have applied EMPI to treat patients with symptomatic pancreatic diseases including OCP after unsuccessful MPD deep cannulation and pancreas divisum since 1996[14].

Therefore, we conducted the present retrospective study of 27-year clinical data to assess the technical success of EMPI through ERCP, immediate and long-term clinical outcomes of MPE, and early and late AEs for treating symptomatic pancreatic diseases.

MATERIALS AND METHODS

Patients

This study was a retrospective analysis of hospitalized patients with symptomatic pancreatic diseases who consecutively underwent EMPI via ERCP at Heilongjiang Provincial Hospital (Heilongjiang, China) between July 1996 and July 2023. Patients who met the following criteria were included in the present study: (1) Patients with OCP accompanied with abdominal pain due to stones and/or strictures of the downstream MPD and sphincter or symptomatic patients with OCP with a reverse-Z type of MPD, when deep canulation of MPD via the major papilla during ERCP failed; (2) Patients with pancreas divisum accompanied with abdominal pain due to obstruction or suspected obstruction of the minor papilla; and (3) Patients with suspected IPMN accompanied with abdominal pain and abnormal orifice of the minor papilla shown during ERCP. Patients with incomplete data were excluded.

Demographic and clinical data including sex, age, clinical symptoms, duration of abdominal pain, and pain intensity were collected from the hospitalized medical records and thorough outpatient follow-ups, or telephone interviews. Data related to ERCP imaging and EMPI procedures were collected from the Digestive Endoscopy Center databases. The indications for EMPI, APD morphology classification, pancreatic stone features (location, number, size, and stone density at X-ray), serum amylase levels, and liver function tests were also collected.

This retrospective study was approved by the Ethics Committee of Heilongjiang Provincial Hospital, and informed consent from the patients was waived due to the retrospective nature.

Diagnosis and classification of indications for EMPI

OCP caused by a pancreatic stone and/or stricture in the downstream MPD or a pancreatic sphincter stricture, usually associated with upstream MPD dilation, was diagnosed through radiological imaging examination. Pancreas divisum was categorized into the following three groups according to clinical presentation: Pancreatic-type pain (PP) without pancreatitis, recurrent acute pancreatitis (RAP), and CP[15]. In addition, pancreas divisum was also classified into three anatomic variants: (1) Classic (or complete) pancreas divisum with total failure of fusion between the ventral and dorsal ducts; (2) Pancreas divisum with absent ventral duct; and (3) Incomplete pancreas divisum[16], and the incomplete variant was further classified into three types (i.e. types 1-3) based on branch fusion[17]. The abovementioned (2) should belong to a special type of pancreas divisum due to the absent ventral pancreatic duct. IPMN, a neoplasm of intraductal epithelium of mucin-secreting cells growing in the MPD and/or its branches[18] that may progress from benign to invasive lesion, was diagnosed by imaging (i.e. pancreatography), pancreatoscopy, and pathology. Specifically, IPMNs were classified by imaging, as branch duct-type, main duct-type, and mixed duct-type[18]. IPMNs were histologically classified into four types including intestinal, gastric, pancreaticobiliary, and oncocytic types based on the cytoarchitectural features and immunophenotypes[19], and categorized as low-grade and high-grade IPMN, and IPMN with invasive carcinoma[18].

In addition, magnetic resonance cholangiopancreatography (MRCP), which is the gold standard for diagnosing pancreas divisum, with a specificity of 97%-99%[10], was performed. In patients without pancreas divisum, the APD was classified into long-type and short-type based on the classification of embryology[17]. In patients with pancreas divisum, the dorsal and ventral duct systems were described, and the dilation of the dorsal duct was defined when the diameter was ≥ 2 mm as previously described[10,20].

EMPI procedures

The therapeutic EMPI (or MPE) procedures included EMP, stricture dilation, stenting, and pancreatic stone therapy after DCMP alone or with additional NKPMP, and diagnostic intervention procedures included peroral pancreatoscopy (ultra-thin pancreatoscopy or SpyGlass DS) and IDUS via the minor papilla.

For EMPI, DCMP was initially performed by using a pull-type minor papillotome (MT-25; Cook Medical, Bloomington, IN, United States), a tapered cannula, or a standard pull-type papillotome, along with a guidewire (0.018-inch, 0.025-inch, or 0.035-inch). Generally, two cannulation methods were carried out. The first one was the direct cannulation method via the minor papilla, mostly wire-guided. The second one was the rendezvous technique via the major papilla for the reverse-Z type of MPD[11,21,22]. After the guidewire accessed to the pancreatic duct via the minor papilla, pancreatic ductography was performed. In cases where DCMP was unsuccessful, freehand NKPMP was performed as previously described[11-13].

After successful DCMP alone or with additional NKPMP, EMP (incision length 3-6 mm) was performed using a pull-type papillotome, or the minor papilla was dilated using a 5-8.5 Fr plastic bougie over guidewire. After EMP or bougie dilation, an indwelling 5 Fr PPS was placed to prevent PEP, or a stent (7-8.5 Fr) was placed for ductal decompression or pseudocyst drainage, and a single 10 Fr stent or double side-by-side 7 Fr stents were placed in the pancreatic duct for stricture therapy. In patients with pancreatic duct stones, a basket or balloon was used to remove small (< 5 mm) stones. Large (≥ 5 mm[5,23]) radiopaque stones located in the head/body of the pancreas were removed with a 4-6 mm balloon dilation for minor papilla alone or in combination with extracorporeal shock wave lithotripsy (ESWL), and large radiolucent stones were mechanically fragmented within the pancreatic duct.

As a diagnostic intervention procedure, ultra-thin pancreatoscopy (0.8 mm in diameter, PF-8P; Olympus America, Inc., Center Valley, PA, United States), SpyGlass DS Direct Visualization System (Boston Scientific, Marlborough, MA, United States), which direct enables visualization of the ERCP blind area of the pancreatic system, or IDUS (UM-29R; Olympus), which provides real-time, cross-sectional imaging of pancreatobiliary ducts and surrounding structures during ERCP using a high-frequency ultrasound transducer, was performed via the minor papilla after DCMP or EMP. If a tumorous lesion was suspected in PD, biopsy specimens were taken under ERCP or SpyGlass.

Follow-up

Follow-up visits were conducted every 3 months during the first year after treatment, either by an outpatient department visit or telephone interview, to assess abdominal pain improvement. Afterward, an annual telephone follow-up was performed unless there was a recurrence of abdominal pain. Patients with abdominal pain during the follow-up period were asked to undergo laboratory tests (amylase and liver function enzymes) and imaging examinations at the outpatient department, and those with obvious abnormal findings were admitted to the hospital for an ERCP procedure depending on their general condition.

Primary and secondary outcomes

The following primary outcomes were measured in all patients undergoing EMPI: (1) Initial technical success, defined as successful DCMP (i.e. access of the guidewire to the upstream pancreatic duct via minor papilla) alone; (2) Overall technical success, defined as successful DCMP alone and successful DCMP with an additional NKPMP procedure after an unsuccessful DCMP procedure; and (3) Immediate clinical success (pain improvement), defined as > 50% improvement in abdominal pain after therapeutic EMPI. A patient self-report scale of 0-10 was used, where 0 indicated no pain, and 10 indicated severe, unbearable, and continuous pain. Scores 0-1 were considered as no pain, 2-3 as mild pain, 4-6 as moderate pain, and 7 or more as severe pain as defined previously[24].

Secondary outcomes were long-term clinical success (i.e. cumulative clinical success), pancreatic stone clearance, and EMPI-related AEs. Long-term clinical success was defined as the absence or decrease of abdominal pain without the need for any analgesics during follow-up for patients with immediate clinical success after EMPI procedures. Pancreatic stone clearance was defined as complete success (> 90% clearance rate), partial success (50%-90% clearance rate), or unsuccess (< 50%, clearance rate)[25]. EMPI-related AEs included early AEs such as PEP, bleeding, and perforation within 48 hours after the EMPI procedures, and late AEs, defined as complications over 6 months such as minor papilla stenosis and pancreatic duct stricture in patients who underwent EMPI procedures and followed up at least 6 months after MPE.

Statistical analyses

The data were analyzed using the international standard statistical programming software SAS 9.4 (SAS Institute, Cary, NC, United States). The numerical variables are expressed as the mean ± SD, or median (range), depending on the normality of the distribution. Categorical data are expressed as percentages, and statistically analyzed with Fisher's exact probability method; odds ratio (OR) and 95% confidence interval (CI) were calculated. P < 0.05 was considered statistically significant. Long-term clinical success after MPE, i.e. EMP, stenting, and stone removal, was evaluated by applying the Kaplan-Meier method.

RESULTS

Characteristics of the patients

A total of 43 patients, 32 with OCP, 8 with pancreas divisum, and 3 with IPMN, were enrolled in this study (Figure 1 and Table 1); 31 (72%) were males, and 12 (28%) were females, with a mean age of 48.4 ± 12.7 years. The median duration of abdominal pain was 24 (range 0.5-206 months). Detailed information on the severity of abdominal pain (n = 43), the indications for EMPI (n = 43), the classification of APD (n = 21), and laboratory test findings (n = 42) are summarized in Table 1. The baseline characteristics of 16 cases with pancreatic duct stones are shown in Supplementary Table 1.

Figure 1 Technical details of endoscopic minor papilla intervention procedures in representative patients. The solid line indicates therapeutic intervention and unsolid line indicates diagnostic intervention. ¹Including 4 patients with obstructive chronic pancreatitis (3 of them with impacted stones) and 1 patient with pancreas divisum and chronic pancreatitis. ²Two or more interventional procedures were performed in some individual patients. ³Two patients received both needle-knife precut minor papillotomy (NKPMP) and endoscopic minor papillotomy (EMP). ⁴A 5 Fr stent (n = 10) or 7-8.5 Fr stent (n = 20) was used and ultimately exchanged in some cases with a single 10 Fr stent (n = 2) and double 7 Fr stents (n = 4) to treat the stricture. ⁵Four patients received diagnostic intervention after therapeutic intervention procedures including EMP (n = 4) and balloon dilation of the minor papilla and stricture (n = 1) for SpyGlass passing through the stricture and balloon sweep (n = 2) for mucin removal. IDUS: Intraductal ultrasound; PD: Pancreatic duct.

Table 1 Baseline characteristics of patients who underwent endoscopic minor papilla intervention procedures for symptomatic pancreatic diseases.

Characteristics	Value
Demographics (n = 43)
Sex
Male	31 (72.1)
Female	12 (27.9)
Age (year)	48.7 ± 12.7
Abdominal pain symptom (n = 43)
Duration of abdominal pain (months)	24 (0.5-206)
Pain requiring hospitalization	30 (73.2)
Chronic recurrent pain requiring analgesic drugs	26 (61.9)
Narcotics	9 (34.6)
NSAIDs	17 (65.4)
Concurrent cholangitis/bile stasis (n = 28)
Cholangitis	10 (24.4)
Bile stasis	18 (43.9)
Indications for EMPI (n = 43)1
OCP	32 (74.4)
OCP with obstruction of MPD or pancreatic sphincter	27 (62.8)
OCP due to a downstream MPD stone or stricture	22 (51.2)2
OCP due to a pancreatic sphincter stricture	2 (4.7)
OCP due to a tumorous stricture	3 (7.0)
OCP with reverse-Z type of MPD	4 (9.3)3
OCP with dorsal duct stones	1 (2.3)4
Pancreas divisum	8 (18.6)5
IPMN	3 (7.0)
Classification of APD (n = 21)
Long-type	13 (61.9)
Short-type	8 (38.1)
Laboratory tests (n = 42)
Elevated serum amylase or lipase	6 (14.3)
Abnormal liver enzyme	12 (28.6)

¹One or more indications may exist in 1 case.

²Including 12 cases with both stones and strictures, 7 cases of impacted stones, and 8 cases with severe strictures; 3 cases were associated with a pseudocyst.

³Two cases with strictures and two cases with both stones and strictures.

⁴One case of obstructive chronic pancreatitis due to complex pancreaticobiliary maljunction with pancreas divisum accompanied by obstructing stones in the dorsal duct and common bile duct.

⁵Including 6 cases with complete pancreas divisum, 1 with pancreas divisum without the ventral duct, and 1 with incomplete pancreas divisum (type 1 with a ventral duct stricture). Clinically, 5, 1, and 2 had pancreatic-type pain, recurrent acute pancreatitis, and chronic pancreatitis, respectively.

Data are expressed as n (%), mean ± SD, or median (range), where appropriate. APD: Accessory pancreatic duct; EMPI: Endoscopic minor papilla intervention; IPMN: Intraductal papillary mucinous neoplasm; MPD: Main pancreatic duct; NSAIDs: Non-steroidal anti-inflammatory drugs; OCP: Obstructive chronic pancreatitis.

Overall, 90 EMPI procedure sessions were performed in the 43 patients, with a median of two (range 1-6) sessions per patient, and the technical details for these patients are presented in Figure 1. DCMP with direct cannulation via the minor papilla and the rendezvous technique were conducted in 41 (95.4%) and 2 (4.7%) cases, respectively, and was successful in 32 patients (Figure 2A-D). In the 4 cases with the reverse-Z type of MPD, DCMP was successful with direct cannulation in two and with the rendezvous technique in the other two (Figure 2A-D). Additional NKPMP was successful in 6 of the 11 patients with unsuccessful DCMP. Eventually, 38 patients with overall technical success underwent therapeutic EMPI procedures, including pancreatic stenting (n = 30, 5 Fr stent in 10 and 7-10 Fr stent in 20), and balloon dilation (n = 9, dilation of the minor papilla in 3, of the stricture in 1, or of the both in 5; Figure 1) for pancreatic stone removal (n = 7), stent placement (n = 1), and SpyGlass DS (n = 1). In our practice, 4-6 mm balloons were used; the 8 mm balloon was used only in 1 case because of the presence of obvious dilated APD. Among the 13 patients with pancreatic stones, 9, 1, 3, and 3 patients were treated with basket/balloon (n = 8), mechanical lithotripsy (n = 1), EMP alone (n = 1), and stenting bypass (n = 3) for accompanied pseudocyst drainage, respectively (Figure 1 and Supplementary Figure 1).

Figure 2 Illustration of the technical success of deep cannulation via the minor papilla in two patients. A-D: Patient 1 receiving the rendezvous technique: Pancreatography shows the reverse-Z type of the main pancreatic duct (MPD) with a proximal stricture (arrow) of MPD, and the guidewire antegrade into the narrowed accessory pancreatic duct (arrowhead; A); Under duodenoscopy, the guidewire retrogradely enters from the major papilla and antegrade through the minor papilla (arrow) into the duodenum (B); The papillotome is inserted into the minor papilla over the other end of the guidewire pulled back from the biopsy channel (C); Pancreatic ductography demonstrates the guidewire (arrow) adjustment entering upstream of the MPD with successful deep cannulation (D); E-H: Patient 2 receiving additional free-hand needle-knife precut minor papillotomy: Endoscopic view shows the needle-knife precut minor papillotomy in progress (E); The papillotome is inserted into the minor papilla after incisions of the minor papilla (F); Pancreatic ductography displays a short type accessory pancreatic duct (APD) and a proximal stricture (arrow) of MPD, and upstream MPD dilation (G); and The guidewire (arrow) successfully passes through the APD into the upstream MPD, achieving successful deep cannulation (H).

Primary outcomes

Initial technical success with DCMP alone was achieved in 32 (74.4%) of the 43 cases; the success rates were not significantly different between patients with long-type APD (69.2%, 9/13) and those with short-type (37.5%, 3/8) APD (OR = 3.749, 95%CI: 0.587-23.932, P = 0.203). In the 11 cases without initial technical success, technical success was achieved in 6 cases (54.5%) with an additional NKPMP procedure (Figure 2F-H). Consequently, the overall technical success rate increased to 88.4% (38/43; Table 2). Moreover, 34 (79.1%) of the 43 patients achieved immediate clinical success (pain improvement); the immediate clinical success was achieved in 26, 5, and 3 patients with OCP (n = 32), pancreas divisum (n = 8), and IPMN (n = 3), respectively (Table 2). In addition, the immediate clinical success was achieved in 33 patients with overall technical success (n = 38) and one of those without (n = 5), respectively (OR = 26.394, 95%CI: 2.432-286.467; P = 0.005). The stents successfully passed through the stricture and were appropriately placed in all 30 patients (including 25 OCP cases and 5 pancreas divisum cases) who underwent pancreatic ductal stenting (Figure 1). In the 2 OCP cases in whom the 5 Fr stents initially failed to pass through the severely obstructed area, we utilized the microcatheter technique to dilate the stricture and successfully placed the stents and achieved pancreatic ductal decompression. Noticeably, there was a significant retraction of the APD caliber with rapid pain improvement and no more use of any analgesics after the additional NKPMP procedure in 1 patient with severe abdominal pain due to alcoholic OCP and requiring at least three injections of pethidine daily although overall technical success was not achieved in this case.

Table 2 Primary and secondary outcomes of endoscopic minor papilla intervention for symptomatic pancreatic diseases (n = 43).

Outcomes	Value
Primary outcomes
Initial technical success of DCMP (n = 43)	32 (74.4)
Overall technical success (n = 43)	38 (88.4)1
Immediate clinical success (pain improvement; n = 43)	34 (79.1)2
OCP (n = 32)	26 (81.3)
Pancreas divisum (n = 8)	5 (62.5)
IPMN (n = 3)	3 (100.0)
Secondary outcomes
Long-term clinical success (n = 22)
Cumulative clinical success (%)
1-year	74.7
3-year	55.3
7-year	41.5
Pancreatic stone therapy (n = 16)
Complete success	7 (43.8)3
Partial success	2 (12.5)3
Unsuccess	7 (43.8)
Stenting bypass	3 (18.8)
EMP alone	1 (6.3)
Failed DCMP	3 (18.8)
Early adverse events (n = 43)	6 (14.0)
Post-ERCP pancreatitis	5 (11.5)4
Self-limited pancreatic duct bleeding	1 (2.3)
Late (> 6 months) adverse events (n = 22)5	2 (9.1)
Minor papilla stricture	1 (4.6)
Main pancreatic duct stricture	1 (4.5)

¹Additional free-hand needle knife precut minor papillotomy achieved technical success.

²Of the 9 patients with no pain improvement, 4 with unsuccessful deep cannulation and 5 received minor papilla endotherapy but failed to achieve > 50% improvement in abdominal pain.

³Clearance success of pancreatic stones was observed in 9 patients, 7 with complete success, and 2 with partial success.

⁴Including 1 severe and 4 mild post-endoscopic retrograde cholangiopancreatography

pancreatitis.

⁵Of the 34 patients with immediate clinical success after endoscopic minor papilla intervention procedures, 22 were followed up for a period of 57.5 (7-266) months.

Data are expressed as n (%) unless otherwise specified. DCMP: Deep cannulation via the minor papilla; ERCP: Endoscopic retrograde cholangiopancreatography; IPMN: Intraductal papillary mucinous neoplasm; OCP: Obstructive chronic pancreatitis.

Secondary outcomes

In 34 (79.1%) patients with immediate clinical success, 3 patients with tumorous stricture were not followed up, and 9 patients, including 1 with IPMN, 7 with OCP, and 1 with pancreas divisum, were lost to follow-up within 6 months after EMPI. Thus, 22 patients were evaluated for long-term clinical success with Kaplan-Meier analysis, which showed that the cumulative clinical success at 1, 3, and 7 years were 74.7%, 55.3%, and 41.5%, respectively, during a follow-up period of median 57.5 (range 7-266) months (Figure 3). In 1 patient with type 1 incomplete pancreas divisum with PP where the flow of pancreatic juice from the major papilla was affected due to a very thin ventral duct and thin communicating branch of the dorsal duct, and there was a dorsal duct obstruction due to minor papilla stricture, immediate and long-term clinical success was achieved after EMP (Supplementary Figure 2A and B). After a 20-year follow-up, there were no further occurrences of abdominal pain. In another patient with pancreas divisum, severe abdominal pain due to RAP was successfully relieved with MPE but experienced a recurrence of RAP 3 years after the EMPI procedure, which progressed into CP despite a patent minor papilla. This patient had been repeatedly hospitalized during a 22-year follow-up period.

Figure 3 Kaplan-Meier curves showing the long-term clinical success (pain free) after endoscopic minor papilla intervention procedures. Cumulative clinical success at 1 year, 3 years, and 7 years were 747%, 55.3%, and 41.5%, respectively, in 22 patients (19 with obstructive chronic pancreatitis and 3 with pancreas divisum) who underwent endotherapy) during a follow-up period of median 57.5 (7-266) months.

Pancreatic stones were present in 16 of the 43 cases. DCMP was unsuccessful in 3 of 16 patients. Complete and partial success was achieved in 7 (43.8%) and 2 (12.5%) patients, respectively. Four (25.0%) cases with stones impacted in the downstream MPD through the minor papilla route failed to achieve stone clearance (Figure 1); stenting was performed in 3 cases (18.8%) to bypass the stones into the upstream MPD and EMP alone was performed in 1 patient. Immediate clinical success was obtained in 12 (75.0%) of the 16 patients. Noticeably, in 1 patient with pancreas divisum, a very rare complex pancreaticobiliary maljunction (PBM) was observed, along with stones in both the dorsal duct and common bile duct, which were completely cleared (Figure 4). Recurrence of pancreatic stone was observed in 3 (33.3%) of the 9 cases with complete and partial success during follow-up period of 42 ± 32 (range 3-103) months; 2 with early recurrence (within 6 months) and 1 with late recurrence (19 month).

Figure 4 A case of complex pancreaticobiliary maljunction with complete pancreatic divisum and accompanied with stones in both the dorsal duct and the common bile duct. A: Magnetic resonance cholangiopancreatography (MRCP) shows notable localized dilation of the dorsal duct (arrow) and a pancreatic stone (arrow), accompanied by a markedly narrowed upstream dorsal duct (arrowhead) and dilation of the intrahepatic and extrahepatic bile ducts. The ventral duct is not visualized; B: MRCP displays the presence of a common bile duct stone (arrow) and dilated dorsal duct, along with dilation of the intrahepatic and extrahepatic bile ducts (arrowhead); C: Cholangiography via the major papilla shows localized dilation of the dorsal duct and a pancreatic stone (arrowhead), indicating an anomalous junction between the dorsal duct and common bile duct; D: Cholangiography via the major papilla shows a large stone in the dorsal duct (8 mm × 12 mm; arrowhead), along with dilation of the intrahepatic and extrahepatic bile ducts progressing towards a choledochal cyst (arrow); E: Endoscopic sphincterotomy is being performed; F: The common bile duct stone is extracted using a basket; G: Endoscopic minor papillotomy (arrow) is being carried out and the major papilla can be seen in distance; H: Pancreatography via the minor papilla after removed the pancreatic stone reveals localized dilation of the dorsal duct (arrowhead) and a narrowed upstream dorsal duct (arrow); I: Computed tomography scan indicates no agenesis of the dorsal pancreas (i.e. absence of the pancreatic body and tail (arrow).

EMPI-related early AEs occurred in 6 (14.0%) of the 43 cases, including PEP in 5 (11.6%) cases (4 mild and 1 severe; 3 due to unsuccessful additional NKPMP and 2 due to EMP) and self-limiting bleeding in 1 (2.3%) case due to balloon dilation for APD stricture. The only case who developed severe PEP was from a patient undergoing NKPMP. The incidence rates of PEP were 27.3% (3/11) and 7.1% (2/28), respectively, in patients undergoing free-hand NKPMP after unsuccessful DCMP and those with EMP after successful DCMP (OR = 4.875, 95%CI: 0.689- 34.501; P = 0.125; Supplementary Table 2). The four mild PEP cases received conservative management, and the one severe PEP case received surgical therapy, and all recovered well. The sex, age, pancreas divisum, balloon dilation, and pancreatic duct stenting were not associated with the occurrence of PEP (Supplementary Table 2).

Late AEs were observed in 2 of 22 patients with long-term clinical success. A severe post-EMP minor papilla stricture, leading to the development of recurrent CP, was reported in 1 (3.1%) with PP-type pancreas divisum 11 years after the procedure (Supplementary Figure 2C-F), and two sessions of mechanical dilation therapy for stricture of minor papilla were performed for stricture. An upstream de novo long-segment MPD stricture developed 6 months after 7Fr pancreatic double stent placement for the dilated downstream MPD stricture in another patient (3.1%) with alcoholic OCP (Supplementary Figure 3), who was followed up without further stenting.

Diagnostic EMPI techniques and outcomes

During or after the therapeutic procedures, 4 patients underwent diagnostic EMPI procedures, including ultra-thin pancreatoscopy (n = 2), IDUS (n = 1), and SpyGlass DS (n = 1) for further assessment of the patients (Figure 1). Three cases were suspected of having IPMN, and one was suspected of having a malignant tumor. In the first case, duodenoscopy showed both the major and minor papilla with a patulous orifice and fisheye-like appearance (Figure 5A and B). Pancreatography via the minor papilla showed a 2.5 cm-sized cystic lesion communicating with MPD in the pancreatic head, and the ultra-thin pancreatoscopy showed some mucus in slightly dilated MPD without any papillary tumor. Thus, this patient was diagnosed by imaging with branch duct-type IPMN, which was not able to be biopsied via the minor papilla.

Figure 5 Illustrations of the branch duct-type and main duct-type intraductal papillary mucinous neoplasms in 3 patients. A and B: Patient 1 with branch duct-type intraductal papillary mucinous neoplasm (IPMN): Duodenoscopy shows the patulous orifice of the major papilla with a fish eye-like appearance (arrow; A), and the patulous orifice of the minor papilla with a fish eye-like appearance (arrow) and mucus flowing out from the orifice before the patient underwent ultra-thin pancreatoscopy via the minor papilla (B); C-F: Patient 2 with main duct-type IPMN: Duodenoscopy shows the enlarged major papilla (arrowhead) with doing deep cannulation (but unsuccessful), and the patulous orifice of the minor papilla (arrow; C), histological examination shows villous papillary architecture composed of pseudostratified tall columnar cells with cigar-shaped enlarged nuclei and basophilic cytoplasm with a large amount of apical mucin, and low-grade and intestinal-type IPMN (hematoxylin and eosin staining, 40 × and 100 ×, respectively; D and E), and a large amount of gelatinous mucous flowing out from both the major (arrowhead) and minor papilla (arrow) 10 months after an endoscopic minor papilla intervention procedure (F); G-J: Patient 3 with main duct-type IPMN: Duodenoscopy shows the enlarged minor papilla (arrow), with mucus visible at the opening (G), the major papilla (arrow) of normal size, with a narrower opening (failed deep cannulation (H), and pancreatic ductography reveals notably dilated (13 mm) dorsal duct (arrows) without filling defects (arrowhead) in the dorsal ducts of the pancreatic body and tail, and the ventral duct (I), and endoscopic minor papillotomy is being performed by using a pull-type papillotome (J).

In the second case, duodenoscopy showed only the minor papilla with an enlarged orifice and fisheye-like appearance (Figure 5C), and IDUS and pancreatography via the minor papilla showed papillary protruding lesions and several filling defects with irregular MPD dilation in the pancreatic head, respectively. Thus, this patient was diagnosed by imaging as having main duct-type IPMN. A large amount of viscous mucus was discharged from both the major papilla and minor papilla after a 10-month follow-up (Figure 5C-F), which can rule out the diagnosis of complete pancreas divisum. In the third case, the minor papilla was enlarged with a small amount of mucus at the opening (Figure 5G) while the major papilla had a normal appearance (Figure 5H). Pancreatography via the minor papilla showed notably dilated (13 mm) dorsal duct without filling defects in the dorsal ducts of the pancreatic body and tail, and the ventral duct (Figure 5I), and then endoscopic minor papillotomy was performed (Figure 5J). Ultra-thin pancreatoscopy revealed multiple tumors inside the dorsal duct at the body and tail of the pancreas, such as frond-like and nodular or papillary protrusions, with atypical vascular images, indicating the possibility of malignancy (Video 1). Histological examination of biopsies taken under ERCP via the minor papilla showed tumors with a villous papillary architecture composed of pseudostratified tall columnar cells, low-grade and intestinal type IPMN (Figure 5D and E) in the second and third cases. Fifteen years later, the third patient underwent pancreatectomy (the body and tail of the pancreas) with splenectomy, partial portal vein resection, and artificial vascular replacement for a tumor measuring 5 cm × 3.5 cm × 1.5 cm. The pathological examination revealed IPMN with associated invasive carcinoma (colloid carcinoma without local lymph node metastasis). As of the most recent follow-up, the patient had survived for four years and five months.

The fourth patient who was suspected of having a malignant tumor developed a de novo long-segment stricture in the upstream MPD with a suspected malignant lesion following 6-month pancreatic stent placement. SpyGlass DS showed an inflammatory stricture, and pathological examination on a biopsy taken from the stricture site under ERCP showed chronic inflammation only (Supplementary Figure 3).

DISCUSSION

In the present study, patients with symptomatic pancreatic diseases received EMPI procedures for treating the primary disease and/or pancreatic duct decompression to relieve abdominal pain. The initial technical success with DCMP was achieved in 74.4% (32/43) of the patients. The success rates were 73.2% (30/41) and 100% (2/2), respectively, with the direct canulation via the minor papilla and rendezvous techniques, and 69.2% (9/13) and (37.5% (3/8), respectively, for long-type and short-type APD. The overall technical success rate increased to 88.4% (38/43) with additional NKPMP. Moreover, the immediate clinical success was achieved in 79.1% (34/43); the rate was 86.8% (33/38) in patients who underwent EMPI vs 20.0% (1/5) in those who did not. The long-term clinical success in patients with both the overall technical success and immediate clinical success and a follow-up period of at least 6 months were 74.7%, 55.3%, and 41.5%, respectively, at 1, 3, and 7 years. Of the 16 patients with pancreatic stones, complete and partial clearance success rates were achieved in 53.9% and 15.4%, respectively, with EMP. Early AEs occurred in 6 (14.0%) of the 43 cases: 5 with PEP. Two late AEs, a post-EMP minor papilla stricture and an upstream long-segment MPD stricture, were observed in 32 patients during the follow-up.

A high initial technical success rate of 74.4% for DCMP was achieved with the direct cannulation method via the minor papilla or the rendezvous technique via the major papilla. Moreover, the overall technical success rate even increased to 88.4% with the addition of NKPMP. These data indicate that free-hand NKPMP can be applied to improve the technical success of EMPI in patients who have a clear appearance and raised minor papilla when deep cannulation or pancreatography via the minor papilla was unsuccessful. Theoretically, the long-type APD converges with the MPD at the pancreatic neck, forming a straight line with the upstream MPD, whereas the short-type APD has an angled convergence with the MPD near the branch of the lower head of the MPD, and thus the short-type MPD potentially affects the success of DCMP[7,26]). However, in the present study, there was no significant difference in the initial deep cannulation success between long-type and short-type APD, with the former showing a tendency to be better than the latter (i.e. 69.2% vs 37.5%).

After successful DCMP, MPE procedures, including EMP, stricture dilation, stenting, pancreatic stone therapy, and pseudocyst drainage, achieved an immediate clinical success (pain improvement) rate in 79.1% of patients with technical success. The rate appeared to be higher in OCP patients (81.3%, 26/32) than in patients with pancreas divisum (62.5%, 5/8), which may be attributed to obstruction relief by stenting, pancreatic stone clearance, and/or pseudocyst drainage in OCP patients. Pain improvement was achieved in all 3 patients with IPMN. This is because EMPI procedures, such as EMP or mucus removal with a balloon, are simple to perform for IPMN due to a patulous orifice of the minor papilla and pain improvement can be quickly achieved; however, stenting is not suitable due to thick mucus. Noticeably, 1 of 5 patients with unsuccessful deep cannulation even after NKPMP also achieved immediate clinical success. He required at least three doses of pethidine daily for severe abdominal pain improvement before EMPI treatment for OCP. Despite the technical failure, the expanded APD diameter was significantly retracted, which resulted in pancreatic ductal decompression and consequently rapid pain improvement, without the need for any pain medications. Moreover, Kaplan-Meier analysis showed a long-term cumulative clinical success of 55.3% and 41.5% at 3 and 7 years, respectively, in 22 patients during a follow-up period of a median 57.5 (range 7-266) months (Figure 3), indicating good therapeutic efficacy for pancreatic diseases, especially OCP, and thus providing more therapeutic options for OCP. For example, we applied minor papillotomy to treat OCP when deep cannulation via the major papilla was unsuccessful during 1996-1999. Among the 29 patients who received endoscopic treatment for CP during the period of time, minor papillotomy was performed in 6 patients with symptomatic OCP due to unsuccessful cannulation via the major papilla, and achieved pain improvement in all these 6 patients after a follow-up period of 10-18 months[14].

Pancreatic duct stones develop in approximately 50%-90% of patients with CP[21], and are common reasons of pancreatic duct obstruction with severe consequences if untreated. Indeed, 60%-70% of patients with obstructing pancreatic stones may be accompanied by an MPD stricture[22], and management of pancreatic stones through minor papilla for endotherapy is challenging. In the present study. EMPI procedures achieved complete and partial success of stone clearance, respectively, in seven (43.8%) and 2 (12.5%) of the 16 cases with pancreatic stones. Immediate clinical success (pain improvement) was achieved in 12 (75.0%) all these 16 patients. In 4 cases (25%), stones impacted in the downstream MPD failed to be cleared because the basket could not reach to the site of stones in the downstream MPD via the minor papilla. However, 3 (18.8%) with stenting to bypass the stones via the minor papilla into the upstream MPD and 1 with EMP alone resulted in pain improvement due to decompression of the pancreatic duct. These findings indicate that EMPI is a very efficacious modality to relieve abdominal pain in most cases of obstructive pancreatic stones with unsuccess via the major papilla. Previous studies have demonstrated that pancreatic stenting prior to ESWL not only reduces the amount of shock waves, shortens the duration of therapy, and decreases the incidence of post-ESWL pancreatitis, but also increases the clearance rate of stone fragments, resulting in ductal decompression and consequently pain improvement[27,28]. In particular, ERCP alone or combined with ESWL is recommended as first-line treatment for painful obstructive pancreatic duct stones[4,28]. It has been reported that the technical success and clinical success rates of ERCP combined ESWL (for radiopaque stone) in patients with chronic calcific pancreatitis are 83% and 80%, respectively[29], and the complete stone clearance rate is 72.6%[30]. An early meta-analysis reported that ESWL achieved a complete clearance rate of 70.7% for pancreatic duct stones[25]. However, a nasopancreatic drainage tube is required for visualizing the radiolucent stone[30]. Besides, ESWL is expensive, and not widely available, and ESWL alone does not address the issue of the accompanying MPD stricture, which has been associated with a high stone recurrence rate[31]. Electrohydraulic lithotripsy or laser lithotripsy under SpyGlass significantly improves the success of the endoscopic treatment of large stones[4,23,31,32], and SpyGlass direct visualization overcomes the limitations of ERCP in relying on contrast for diagnosis[9,33].

Therefore, appropriate application of imaging, ERCP, SpyGlass, and other methods will help achieve complementary advantages, and precise individualized treatment can reduce the pains of patients without affecting the treatment effect and, at the same time, reduce the recurrence rate and complications of PDS and maximize the benefits of patients.

Noticeably, complex PBM accompanied with stones in both the dorsal and common bile ducts was observed in 1 case with complete pancreas divisum, and EMPI successfully removed the stones and achieved pain improvement. PBM is a rare congenital anomaly with the junction of the pancreatic and bile ducts located outside the duodenal wall, which is classified into types A-D. Of these types, type D (or complex type) is the rarest, accounting for less than 5% of PBM cases[34]. This case of complex PBM represented the second we had ever observed; the first one we reported previously was also classified as type D with complete pancreas divisum accompanied with a choledochal cyst[35]. These 2 cases accounted for 0.12% of our total ERCP cases over a 35-year period. The case in the present study was diagnosed based on ERCP and MRCP as having pancreas divisum with absent ventral duct[16]. In addition, both MRCP (Figure 4A) and ERCP (Figure 4H) showed a narrowed and relatively short upstream dorsal duct, suspected of agenesis of the dorsal pancreas (i.e. the absence of the body and tail of the pancreas). However, computed tomography scans for the pancreas can rule out this condition (Figure 4I).

In the present study, early AEs were observed in 6 of 43 patients undergoing DMPC alone (n = 28) or with additional free-hand NKPMP (n = 11), with PEP being the most common AE related to EMPI (n = 5, 11.6%). It should be emphasized that the incidence appeared to be higher in the cases undergoing free-hand NKPMP after unsuccessful DCMP than in those with EMP after successful DCMP (27.3% vs 7.1%) although the difference was not statistically significant. In addition, the only severe PEP case was related to free-hand NKPMP. Thus, free-hand NKPMP is associated with a potential risk to the development of PEP or even severe PEP. Late AEs were observed in 2 cases. In 1 case with PP-type pancreas divisum, severe narrowing of the minor papilla, leading to the development of CP, occurred 11 years after EMP although a PPS was placed to prevent PEP and minor papilla stricture after the procedure. In another case in which double 7 Fr stents were placed side by side for 6 months after EMP, upstream MPD long-segment narrowing developed 6 months after the procedure, most likely due to the patient's continuous alcohol abuse and stent blockage leading to secondary inflammatory fibrosis causing the stricture.

In the present study, 5 Fr stents were used for the prevention of PEP and early papilla stricture, or ductal decompression when 7 Fr stents were unable to pass beyond the severe stricture, 7-8.5 Fr stents were used for the ductal decompression, stenting to bypass the stones for pain improvement, or pseudocyst drainage therapy, and 10 Fr stents were used for stricture dilation in patients with OCP. It has been reported that endoscopic stenting achieves pain improvement in 93% of patients with obstructive pain due to pancreatic ductal adenocarcinoma[36]. However, there are few reports on the relief of abdominal pain due to tumorous obstruction through stenting of APD. In the present study, all 3 cases of pancreatic duct adenocarcinoma with APD stenting achieved immediate clinical success, and they were not further followed up after discharge from the hospital.

Previous studies have shown that using a balloon to expand the papilla is beneficial for extracting large pancreatic stones[24,37,38]. Due to the small size of the minor papilla and the absence of a sphincter (only the muscular coat of the duodenum)[39], and the APD being narrower than the MPD, balloon dilation of the minor papilla is associated with a high risk of AEs including PEP, even with PPS placement[12]. However, there is no consensus on the safe diameter of the balloon for expansion. Until now, there are few reports on balloon expansion of the minor papilla in no pancreas divisum[9,22,40,41]. One study on the safety of pancreatic duct balloon expansion in ex vivo animal experiments[42] provided valuable data; however, the resistance of inflammatory pancreatic duct strictures or fibrotic thickening of the duct wall to balloon expansion force is unclear. In the present study, 7 patients (77.8%) used 4-8 mm balloons for stone extraction, whereas 1 case (2.4%) experienced self-limiting bleeding from the pancreatic duct due to dilation stricture with a 6 mm balloon, and a 7 Fr stent was placed without further intervention. No perforations or severe AEs related to the EMPI procedure were observed in this study.

It has been reported that endoscopic ultrasound-guided pancreatic duct drainage (EUS-PDD) procedure can relieve abdominal pain through pancreatic duct decompression, with technical and clinical success rates of 81.4% and 84.6%, respectively, as reported in a meta-analysis[43]. These rates are similar to that (88.4% and 79.1%, respectively) observed in the present study. However, EUS-PDD requires the assistance of pancreatoscopy[44,45] or a double-balloon enteroscope[46] and is associated with a high incidence of AEs (21%) of AEs. Thus, this technique is not widely applied; well-trained gastroenterologists are required to perform the procedure in order to reduce the incidence of AEs[43].

It was of note in this study that, of the 4 patients undergoing diagnostic EMPI procedures, including pancreatoscopy or IDUS examination, 3 showed abnormalities in the minor papilla, suggestive of IPMN. In the first patient with a fish eye-like appearance in both the minor and major papilla, ultra-thin pancreatoscopy only showed some mucus in MPD, and pancreatography via the minor papilla showed branch duct cystic dilation and communicating with MPD. This patient was diagnosed with branch duct-type IPMN. In the second patient, IDUS showed papillary protruding lesions in MPD. Subsequently (10 months later), a large amount of viscous mucus flowed out of the major papilla, ruling out complete pancreas divisum. In the third case, the minor papilla was swollen, with less mucus at the papilla orifice, while the major papilla appeared normal, typically indicating IPMN with pancreas divisum, which is very rare[47]. Histologically, tumors with villous papillary architecture, low-grade and intestinal type IPMN all were shown (Figure 5D and E) in the second and third patients, which all were diagnosed by imaging with main duct-type IPMN. Since it is difficult to distinguish the tumor origin among ampulla, distal common bile duct and pancreas, especially for advanced ampullary carcinoma[48], differential diagnosis is required between IPMN and ampullary villous adenoma or ampullary carcinoma due to their pathological findings. It has been reported that papillary tumors with a height of ≥ 3 mm in IPMN are detected by IDUS with a detection rate of 100%, compared to 83% with peroral pancreatoscopy[49]. Pancreatic ductography is valuable for diagnosing cystic masses communicating with the MPD, while pancreatoscopy is unable to evaluate tumors located in a side branch[50]. However, a definitive diagnosis of IPMN with pancreas divisum requires a combination of pancreatoscopy, MRCP, and pancreatography. In the third case, both MRCP and ERCP did not show the ventral duct, with significant dilation of the dorsal duct throughout (13 mm), consistent with pancreas divisum with absent ventral duct[16]. The histological examination of operative specimens showed IPMN with associated invasive carcinoma (colloid carcinoma) without local lymph node metastasis in this case. All of these 3 patients with IPMN were pain-free after EMPI procedures including EMP, or mucus removal with a balloon. They all survived at the last follow-up. Particularly, in the third case, multiple tumors were presented inside the pancreatic duct, and appeared to be malignant with atypical vascular images[33] as shown by ultra-thin pancreatoscopy, with the dorsal pancreatic duct dilation of more than 12 mm as shown by pancreatography, indicative of surgical therapy[19]. This patient underwent pancreatectomy 15 years later and was still alive 4.5 years after pancreatectomy. Our findings indicate that IPMN progresses slowly and patients with IPMN have a favorable prognosis if appropriately treated.

The present study had several limitations. First, it was conducted as a single center and retrospective analysis, which may introduce inherent biases associated with retrospective data collection, and thus limits the generalizability of the results. Second, the sample size of 43 studied patients was small. Noticeably, no sample size calculation was performed for the present study. As EMPI is a very specialized procedure and can only be operated by experienced endoscopists under ERCP, and thus is applied in the treatment of symptomatic pancreatic diseases only at specialized hospitals, the number of patients indicated for EMPI is still small due to its unclear feasibility, clinical efficacy, and safety. That is why we carried out this retrospective study. Ideally, patients diagnosed with symptomatic pancreatic diseases and exocrine pancreatic insufficiency should be included. In the present study, 43 patients with symptomatic pancreatic diseases were included, most with OCP. Among these patients, in 10 cases only pancreatic function test were done. Due to this relatively small number as well as the retrospective nature of the study, we decided not to include “exocrine pancreatic insufficiency” in the analysis of the present study and would include it in a future prospective study. Third, the follow-up period of 57.5 (7-266) months was relatively short, and thus extending the follow-up period would provide crucial data on the durability of EMPIs and their sustained effectiveness in managing symptomatic pancreatic diseases. Fourth, there was no comparison group or control arm, which hinders the ability to fully assess the relative effectiveness of EMPIs compared to alternative treatments. Finally, the utilization of pancreatoscopy (n = 3) and IDUS (n = 1) via the minor papilla for diagnostic intervention and therapeutic intervention (i.e. pancreatic stone therapy) was uncommon in clinical practice, and thus these diagnostic EMPI procedures need further evaluation; indeed, there is limited literature[9,22,40,41,47] on these specific interventions. Therefore, a further multicenter, prospective study with a larger and more diverse patient cohort (including those with exocrine pancreatic insufficiency), a longer follow-up duration, and a control group is needed to allow for a more rigorous evaluation of the benefits and potential drawbacks of EMPIs and thus enhance clinical understanding and management strategies of these interventions for symptomatic pancreatic diseases.

CONCLUSION

In conclusion, EMPI is a feasible, effective, and safe modality for the management of symptomatic pancreatic diseases, in terms of the technical and clinical success, stone clearance, and incidence and severity of AEs. NKPMP appears to enhance the technical success rate. However, potential risks of developing PEP and late AEs should be kept in mind. Additionally, while the present study provides valuable insights into the technical success, clinical efficacy, and safety of EMPI, further research is necessary to address the limitations mentioned. Prospective studies with longer follow-up periods and comparison groups would offer a more comprehensive understanding of the effectiveness and durability of EMPI in the treatment of symptomatic pancreatic diseases.