Combined versus Separate Sessions of Endoscopic Ultrasound and Endoscopic Retrograde Cholangiopancreatography for the Diagnosis and Management of Pancreatic Ductal Adenocarcinoma with Biliary Obstruction

Abstract

Background:

A single procedure session combining EUS and ERCP (EUS/ERCP) for tissue diagnosis and biliary decompression for pancreatic duct adenocarcinoma (PDAC) is technically feasible. While EUS/ERCP may offer expedience and convenience over an approach of separate procedures sessions, the technical success and risk for complications of a combined approach is unclear.

Aims:

Compare the effectiveness and safety of EUS/ERCP versus separate session approaches for PDAC.

Methods:

Study patients (2010–2015) were identified within our ERCP database. Patients were analyzed in three groups based on approach: Group A: Single session EUS-FNA and ERCP (EUS/ERCP), Group B: EUS-FNA followed by separate, subsequent ERCP (EUS then ERCP), and Group C: ERCP with/without separate EUS (ERCP +/− EUS). Rates of technical success, number of procedures, complications and time to initiation PDAC therapies were compared between groups.

Results:

200 patients met study criteria. EUS/ERCP approach (Group A) had a longer index procedure duration (median 66 min, p=0.023). No differences were observed Group A versus sequential procedure approaches (Groups B and C) for complications (p=0.109) and success for of EUS-FNA (p=0.711), ERCP (p=0.109). Subgroup analysis (>2 months of follow up, not referred to hospice, n=126) was performed. No differences were observed for stent failure (p=0.307) or need for subsequent procedures (p=0.220). EUS/ERCP (Group A) associated with a shorter time to initiation of PDAC therapies (mean, 25.2 vs 42.7 days, p=0.046).

Conclusions:

EUS/ERCP approach has comparable rates of success and complications compared to separate, sequential approaches. An EUS/ERCP approach equates to shorter time interval to initiation of PDAC therapies.

Introduction

Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer death in the United States. The five-year survival for PDAC is less than 7% due to late stage at the time of diagnosis (1–3). Only 15–20% of the patients with PDAC are candidates for a curative resection at presentation.

The majority of symptomatic patients present with jaundice (>50%), caused by an obstructing tumor localized to the uncinate, head and/or genu of the pancreas (1, 4). Cholestasis, defined as a bilirubin >2.5mg/dL, is a barrier to the initiation of chemotherapy for PDAC. Symptoms of biliary obstruction such as pruritus, fatigue and fat malabsorption also compromise quality of life for patients with PDAC. (3)

Endoscopic retrograde cholangiopancreatography (ERCP) with stent insertion has a high rate of technical success (>90%), low rate of complications (5%) and is considered the standard of care for decompression of bile duct obstruction from PDAC. (3) Deployment of self-expanding metallic stents (SEMS) at index ERCP provides a further interval benefit in terms of durability of decompression and need for fewer subsequent endoscopic interventions for patients with PDAC. While SEMS are more expensive than plastic biliary stents, they are superior when longitudinal medical costs related to hospitalizations due to stent failure and subsequent procedures necessary to maintain biliary decompression are taken into consideration. SEMS offer superior durability at a similar overall healthcare expense to plastic biliary stents. (5, 6)

However, selecting the optimal approach for decompression for biliary obstruction can be a complex decision, predicated on clinical information available at the time of ERCP. Confirmatory tissue diagnosis, surgical candidacy and life expectancy are paramount among these considerations. Endoscopic ultrasonography with fine needle aspiration (EUS-FNA) is the standard of care for tissue acquisition for PDAC with a sensitivity of 85–90% and specificity of 100%. (7) EUS and ERCP are now considered complimentary diagnostic and therapeutic procedures for patients with obstructive jaundice from a pancreatic malignancy. Given SEMS have significant expense and the permanence of uncovered SEMS, pre-ERCP tissue acquisition is ideal to select the most cost effective, durable option for decompression (plastic stent vs SEMS, uncovered vs covered SEMS). Consequently, EUS-FNA ideally occurs prior to ERCP.

Combining EUS and ERCP (EUS/ERCP) into a single endoscopic session is technically feasible for endoscopy units that offer both services (8). The majority of therapeutic endoscopy training programs now graduate physicians with both skill sets. However, this approach increases anesthesia time, may create circumstance where peri-ampullary tissue trauma from FNA passes impacts ERCP technical success rates and has an unclear interval risk for complications (9). While promptly providing a tissue diagnosis and biliary decompression in a single session is convenient for patients, combined EUS/ERCP is of unclear clinical benefit and interval risk for complications. To date, no comparative data exists on the ideal initial approach for timing and coordination of EUS-FNA and ERCP for PDAC.

In the present study, we aimed to evaluate the efficacy and safety of combined, same session EUS/ERCP in patients with distal biliary obstruction from PDAC. We compared the combined, single session approach to the sequential use of EUS-FNA and ERCP modalities in the management and diagnosis of PDAC.

Methods

Inclusion and Exclusion Criteria:

The Indiana University School of Medicine Institutional Review Board (IRB) approved this project. Consecutive inpatients and outpatients that underwent ERCP for obstructing PDAC at Indiana University Hospital between 2010 and 2015 for malignant biliary obstruction were identified within our prospectively maintained ERCP database. All EUS-FNA and ERCP records were reviewed retrospectively in these patients. Inclusion criteria were: that the first ERCP and EUS procedures were performed at our institution, biliary obstruction as defined by a total bilirubin >2.5 mg/dL and/or biliary dilation, cross-sectional imaging with PDAC localized to the uncinate, head, or genu of the pancreas, and an ultimate, final tissue confirmation of PDAC. Exclusion criteria were: lack of evaluable records for pre-procedure clinical data, anesthesia and EUS/ERCP procedures; uncertainty if prior EUS or ERCP procedures were performed outside of our institution; prior tissue acquisition with radiographic or surgical maneuvers; prior percutaneous trans-hepatic biliary drainage (PTBD) or surgery for biliary decompression; surgically altered foregut anatomy that would impact decision making for EUS/ERCP approach (Billroth II, Roux-en-Y gastric bypass); lack of final tissue diagnosis confirming PDAC; and malignant obstruction due to lesions other than PDAC localized to the pancreatic bed.

Data Acquisition:

Electronic medical records were reviewed for demographics, EUS and ERCP maneuvers, type and duration of anesthesia, complications relating to EUS/ERCP procedures and/or anesthesia. Index imaging findings such as the size of mass, location, presence of metastases, and overall stage of pancreatic cancer (Stage I/II, Stage III and Stage IV) were also reviewed and recorded. Follow up data were evaluated including need for subsequent procedures for index tissue diagnosis and biliary decompression, management of biliary stent failure, admission for cholangitis and time to first therapy for PDAC (chemotherapy or resection).

The outcomes measured were duration of the first (index) EUS and/or ERCP procedure session, technical success of each respective procedure and complications due to study procedures and anesthesia. Technical success was defined for EUS-FNA as successful tissue acquisition and for ERCP as satisfactory stent placement for biliary drainage. Secondary outcomes include: biliary stent failure (defined by new or persistent jaundice, cholangitis, biliary stent migration or occlusion necessitating a repeat procedure), need for and number of subsequent procedures (e.g. ERCP, EUS, PTBD, surgery) for tissue acquisition and/or biliary decompression, length of hospital stay when inpatient, and time to initiation of therapy for PDAC.

Study procedures

EUS-FNA is performed at our institution using linear array echoendoscopes (Olympus 140 Series), either 22 or 25-gauge needles and cytological examination is performed at bedside with the assistance of cytotechnicians with or without cytopathologists. Adequacy of the samples and an on-site diagnosis is conveyed to the endosopist(s) during the EUS-FNA procedure. During the study period, EUS-FNA was performed by five physicians (I.E., S.S., J.E., L.M., J.D, MA). ERCP was performed using Olympus JF / TJF 160 / 180 series duodenoscopes by seven experts (I.E., S.S., J.E., L.M., J.W., E.F., G.L.). Combined EUS/ERCP was performed either as dual operator (e.g., J.D. followed by G.L.) or single operator (e.g., J.E.) format. Intraductal sampling at ERCP was performed using a cytobrush and/or standard/pediatric biopsy forceps when deemed necessary. ERCP sampling is often performed if ERCP was the first procedure or if ERCP followed a preceding EUS- FNA with on-site interpretation as indeterminate for malignancy. ERCP biliary decompression is performed at our institution with or without sphincterotomy. While stent selection is ultimately at the discretion of the ERCP operator, we generally deploy 10 Fr plastic biliary stents if the tissue diagnosis is unavailable/indeterminate and SEMS (covered or uncovered) if malignancy is confirmed prior to ERCP. A covered SEMS stent may be placed with a high pre-ERCP suspiscion (e.g. mass on imaging, elevated CA 19–9) if a tissue diagnosis is unavailable.

Statistical Analysis

For the purpose of statistical analysis, patients were divided into three groups by index approach for endoscopic tissue acquisition and biliary decompression (Figure 1): Group A: Same session EUS-FNA and ERCP (EUS/ERCP); Group B: EUS-FNA followed by sequential subsequent ERCP (EUS then ERCP); and Group C: ERCP with or without sequential EUS (ERCP +/− EUS).

Figure 1: Allocation for Analysis of patients that met study criteria (n=200).

Approach for analysis of patients that met study criteria: Inclusion criteria: index ERCP and EUS procedures performed at our institution, biliary obstruction (total bilirubin >2.5 mg/dL and/or biliary dilation), PDAC localized to the uncinate, head, or genu of the pancreas, and final tissue confirmation of PDAC. Exclusion criteria: lack of evaluable records for pre-procedure data, anesthesia and EUS/ERCP procedures; prior EUS or ERCP procedures were performed outside of our institution; prior tissue acquisition with radiographic or surgical maneuvers; prior percutaneous trans-hepatic biliary drainage (PTBD) or surgery for biliary decompression; surgically altered foregut anatomy (Billroth II, Roux-en-Y gastric bypass); lack of final tissue diagnosis confirming PDAC; and malignant obstruction due to lesions other than PDAC.

Data analysis was performed using Statistical Package for Social Sciences (SPSS) version 24, 25 software (SPSS Inc., Chicago, IL, United States). Descriptive data was reported using parameters such as frequency (%), median (interquartile range, IQR), mean (standard deviation). We used chi-square or Fisher exact test for categorical data. We used Student’s T-test or Wilcoxon rank-sum test for continuous variables as appropriate. The level of significance was considered as < 0.05.

Results

Baseline Characteristics:

Ultimately, a total of 200 patients managed with EUS-FNA for tissue acquisition and ERCP for biliary decompression during the study period met study criteria and were analyzed. General anesthesia was utilized for ERCP in the majority of cases (93%). Rates of on-site diagnosis for EUS-FNA were 85% and final diagnostic cytopathology were 90%. The overall rate of technical success for index ERCP was high (97%). Partial gastric outlet obstruction was present in 7 patients (3.5%), with ultimate technical success of ERCP in all of these patients. A duodenal stent was placed in n=6 patients (3%). There was no difference in rates of technical success for EUS-FNA (p=0.711) and ERCP (p=0.109) for combined EUS/ERCP (Group A) versus a sequential approach (Groups B+C).

Examining cohorts managed through each of these approaches, patients managed with an EUS then ERCP (Group B) approach tended to be outpatients (78% vs 63%, p=0.08), younger (60 vs 70 years, p= 0.003) and have a lower bilirubin level prior to index procedure (total bilirubin 6.3 vs_12.3 mg/dL, median, p=0.008) when compared to the other groups (A+C) (Table 1). Patients managed with an ERCP +/− EUS (Group C) more often had intraductal sampling (87% vs 18%, p <0.0001) when compared to EUS first approaches (A+B). Of patients sampled with intra-ductal accessories in the ERCP +/− EUS (Group C) approach, 34% of patients had positive bile duct pathology, obviating the need for EUS-FNA as a subsequent procedure. SEMS were less often placed at index ERCP in patients with an ERCP +/− EUS (Group C) approach (48% vs 72%, p=0.001) when compared to EUS first approaches (A+B). Of the ERCP +/− EUS patients (Group C), 60% did not have a subsequent EUS-FNA procedure. Of these patients, a significant proportion were managed with surgery after ERCP (40%) and received a tissue diagnosis at the time of surgery (36%).

Table 1:

Baseline Patient Characteristics, Endoscopic, Anesthesia and Pathology Data (n=200)

Group	A		B		C
Variable	EUS/ERCP Combined	% or (IQR)	EUS Then ERCP	% or (IQR)	ERCP +/− EUS-FNA	% or (IQR)	EUS/ERCP Combined (A) vs Sequential (B+C)
							p value
Total	88		37		75
Age, Median (Years)	70	(59,78)	60	(54,67)	71	(61,80)	0.608
Female Sex	49	55.7	17	45.9	36	48	0.342
Age Adjusted Charlson Comorbidity Index, Median	4	(2,5)	2	(2,4)	4	(2,5)	0.453
Inpatient Status, n (%)	33	37.5	8	21.6	27	36	0.309
Length of Stay Prior, Median	1	(1,2)	6	(4,11)	2	(1,4)	0.043
Mass Size, Median	3	(2.5,3.9)	3.5	(2.6,4)	3.1	(2.4,3.9)	0.359
Stage, n (%)
I/II	39	44.3	15	40.5	25	33.3	0.199
III	18	20.5	9	24.3	19	25.3
IV	26	29.5	11	29.7	15	20
Pre-Procedure T Bili, Median	12.2	(6.9,17.5)	6.3	(2.1,14.2)	13.9	(8.7,19.6)	0.407
General Anesthesia, n (%) * for ERCP component of the two procedures	84	95.5	34*	91.9	67*	89.3	0.023
Index Procedure Duration, Minutes, Median	66	(54.5,79)	29	(25,42)	34.5	(28,51)	<0.001
Duration EUS If Performed Separate, Minutes, Median			29	(25,42)	27.5	(19,46)
Intraductal Sampling, n (%)	17	19.3	5	13.5	65	86.7	<0.001
Positive, n (%)	10	11.4	0	0	22	33.8	0.038
Index Stent Type, n (%)							0.022
Plastic	25	28.4	9	24.3	37	49.3
Any SEMS	62	70.5	26	70.3	34	45.3
fCEMS	6	6.8	2	5.4	3	4.0
EUS-FNA, n (%)	88	100	37	100	30	40
No of Passes, Median	4	(3,5)	4	(3,5)	4.5	(3.7)
On-Site Cytopathology Pos, n (%)	77	87.5	33	89.2	22	73.3	0.291
Final Cytopathology
Diagnostic, n (%)	80	90.9	34	91.9	25	83.3	0.771
Suspicious, n (%)	1	1.1	0	0	1	13.3
Non-Diagnostic/False Negative, n (%)	7	8	3	8.1	4	3.3
Any Complications, n (%)	2	2.3	3	8.1	5	6.7	0.275
Anesthesia Related, n (%)	0	0	1	2.7	1	1.3	0.21
Procedure Complications, n(%)	2	2.3	2	5.4	4	5.3	0.275
Pancreatitis, n (%)	2	2.3	0	0	3	4
Bleeding, n (%)	0	0	0	0	1	1.3
Failed ERCP, n (%)	1	1.1	2	5.4	4	5.3	0.109
PTC	0	0	0	0	1	25
Surgical Resection	1	100	1	50	0	0

No significant differences were otherwise found between the groups regarding their baseline characteristics and maneuvers performed at index EUS and ERCP procedures.

A combined EUS/ERCP approach (Group A) equated to a longer first procedure session (median 66 min, p =0.023), with greater utilization of general anesthesia (96% vs 90%, p=0.023) compared to patients managed with sequential, separate procedure sessions (Groups B + C). However, there was no difference in rates of endoscopic or anesthesia related complications (p= 0.275) with an EUS/ERCP approach (Group A) compared to sequential approaches (Group B+ C). Rates of on-site (88% vs 82%, p=0.291) and final cytopathology diagnosis (91% vs 88%, p= 0.771) for EUS-FNA and technical success for ERCP (99% vs 95%, p=0.109) were the similar with an EUS/ERCP approach (Group A) vs sequential, separate procedure approaches (Groups B + C).

Post-procedure Outcomes and Follow Up:

To evaluate the impact of index endoscopic approaches on the clinical course of our cohort subsequent to the index procedures, we performed a subgroup analysis on patients that were not immediately transitioned to hospice after endoscopic procedures and had more than 60 days of follow up after the index diagnostic and therapeutic procedures were complete (mean 232 days) (Table 2).

Table 2:

Outcomes in Patients with Greater than 60 Days Follow-Up Data (n=126)

	A		B		C
Variable	EUS/ERCP Combined	% or (IQR)	EUS Then ERCP	% or (IQR)	ERCP, +/− EUS	% or (IQR)	EUS Combined (A) vs Sequential (B+C))
							p-value
Total	60		22		44
Age Adjusted Charlson Comorbidity Index, Median	3	(0,2)	2	(0,2)	3	(0,2)	0.793
Inpatient Status, n (%)	21	35	3	13.6	11	25	0.116
Stage, n (%)							0.429
I/II	27	45	10	45.5	23	52.3
III	10	16.7	5	22.7	11	25
IV	20	33.3	7	31.8	7	15.9
Length of Follow Up, Days, Mean (STDV)	294.2	(215.1)	320	(273)	389.7	(266.8)	0.103
PDAC Therapy
Chemotherapy	53	88.3	20	90.9	39	88.6	0.229
Surgery	27	45	8	36.4	28	63.6	0.292
Stent Failurea, n (%)	20	33.3	5	22.7	11	25	0.307
Repeat ERCP, n (%)	18	30	4	18.2	11	25
Total Procedures for Adequate Biliary Drainage, Median	1	(1,2)	1	(1,2)	1	(1,2)	0.220
Post Procedure Total Inpatient days for Biliary Drainage and Diagnostic Sampling, Mean (STDV) b	4.6	(4.9)	10	(9.5)	3.4	(2.6)	0.858
Time to Initiation of PDAC Therapy, Mean (STDV)c	25.2	(15.6)	50.6	(100.7)	38.5	(38.3)	0.046

^aDefined as failure for resolution of jaundice, new jaundice, cholangitis or stent occlusion on follow up after biliary stent placement.

^bAggregate inpatients days attributable to procedures for diagnostic sampling and/or maintenance of biliary decompression following index procedure.

^cTime interval from index procedure to first therapy directed toward PDAC (chemotherapy or resection)

For patients managed with a combined EUS/ERCP (group A), rates of stent failure (33% vs 24%, p= 0.307) and total procedures required for tissue acquisition and maintain biliary decompression (p=0.220) were similar to separate, sequential approaches (Groups B + C).

The average time period between index endoscopic procedure and initiation of the PDAC therapy was shorter (25.2 vs 42.7 days, p=0.046) with a combined EUS/ERCP (Group A) vs sequential approaches (Groups B+ C).

Discussion:

We offer the first comparative study examining the effectiveness and outcomes for a combined, same session EUS-FNA/ERCP approach for tissue acquisition and biliary decompression in patients with obstructive jaundice from PDAC.

In this retrospective study, we examined rates of technical success for EUS-FNA and ERCP with combined versus sequential approaches. Rates of confirmatory EUS-FNA tissue diagnosis across all of the groups in our cohort was consistent with the literature for EUS-FNA diagnosis of PDAC (7, 10). Rates of on-site and final cytopathology diagnosis from FNA tissue acquisition during a combined EUS/ERCP (Group A) procedure were the same as sequential, separate procedure approaches (Groups B+C). Examining technical success of ERCP immediately following EUS-FNA is of particular interest given the potential for increased difficulties with cannulation and stent insertion. An ERCP may be technically challenging after EUS-FNA due to edema from FNA needle punctures, mechanical trauma to ampullary and peri-ampullary mucosa, hematoma and intra-luminal bleeding. However case series that evaluate a combined approach for EUS-FNA/ERCP describe rates of cannulation and stent insertion consistent with that expected for ERCP alone (4,5,6). Our comparative study demonstrates the same rate of technical success for ERCP biliary drainage with a combined EUS-FNA/ERCP when compared to sequential approaches.

Fisher et al. postulated that the increased duration of anesthesia necessary for a combined EUS-FNA/ERCP procedure would place patients at increased risk for cardiopulmonary complications. (11) However, Ross et al. reported rates of complications for a combined EUS/ERCP to be consistent with that expected for ERCP alone in their cohort (8, 9). This is also true in other case series on the subject. (7–10). While we demonstrate that the index procedure duration is significantly longer for combined EUS-FNA/ERCP, we found no difference in anesthesia or procedure related complications. We should note that the majority of our ERCP procedures are performed under general anesthesia with endotracheal intubation (93% cohort, 96% combined EUS/ERCP). This is an important consideration when assessing for anesthesia related complications. A prospective study evaluating rates of anesthesia complications during ERCP published by Berzin et al. revealed procedure duration to be a predictor for such events. However, Berzin’s cohort was predominantly anesthetized using a sedation approach without intubation (89%). The majority of events in Berzin’s study were respiratory (e.g. hypoxia to less than <85%) and more often occurred in non-intubated patents. (12) We prefer anesthesia with endotracheal intubation for combined EUS/ERCP sessions for several reasons. First, we wish to decrease the risk of unplanned procedure interruptions or terminations from the respiratory events that can occur during lengthy endoscopic procedures. Second, in an effort to optimize procedure time and enhance efficiency for our combined EUS/ERCP cases, we also perform both segments of a combined EUS/ERCP in the prone position on the fluoroscopy table. This strategy eliminates the extra time required to reposition the patient from a traditional EUS-FNA position (left lateral) to an ERCP position (prone on the fluoroscopy table) during a session. Control of the patient’s airway with an endotracheal tube is that much more desirable for our teams when it is anticipated that the patent will remain prone for an extended period of time during a procedure. In summary, while duration appears to be a risk for anesthesia related complications during advanced endoscopic procedures, our study lends support to the existing literature that suggests that the incremental increase in procedure duration required to combine an EUS and ERCP for management of PDAC does not increase risk for complications.

Outcomes were similar between a combined EUS/ERCP approach (Group A) and sequential approaches (Groups B+C). Specifically, rates of stent failure, number of procedures required to maintain biliary decompression and inpatient days following index procedure were the same between combined and sequential approaches. While metallic stents (>70%), were more often placed in patients managed with a combined EUS/ERCP (Group A) or an EUS then ERCP (Group B) approach, there was no difference in rates of stent failure comparing combined EUS/ERCP and sequential groups (p=0.31). It is possible that our study is underpowered to detect any difference between groups for stent failure. Another explanation for the similar rates of stent failure between these groups in spite differing rates of SEMS placement may be due to our center’s approach for recall of PDAC patients for exchange of plastic biliary stents. We tend to proscribe a relatively early follow up interval for exchange of plastic biliary stents to a SEMS for these patients. We will often schedule an ERCP stent exchange for PDAC patients at intervals of less than 8-weeks if chemotherapy is planned. This is to decrease the likelihood that an interval event such as stent failure or ascending cholangitis will interrupt a patient’s chemotherapy treatment.

When examining time from index procedure to initiation of PDAC therapies (chemotherapy, radiation or surgery), combined EUS-FNA/ERCP appeared to expedite transition of patients to PDAC directed therapies. We analyzed this metric in the subset of patients that were not ultimately referred to hospice, as these are patients in whom an expedited strategy for tissue diagnosis and decompression is of greatest import. While this difference in time to therapy is small, it should be noted that our institution has carefully aligned resources (highly trained support staff, dedicated pancreatic cancer patient navigator RN, weekly multidisciplinary pancreatic cancer conference, same day oncology, surgery appointments). This system equates to the efficient deployment of consultations, procedures and staging imaging for PDAC patients following a diagnosis.(3) This is demonstrated in our study through the short average time interval between first diagnostic/therapeutic endoscopic procedure and delivery of PDAC therapies (mean 30.5 days) for our entire cohort. In this context, the impact of combined EUS/ERCP in expediting patient care may be understated.

Specifically, a combined EUS/ERCP may be of greater impact for expediting care at institutions and/or in care systems that lack the resources to efficiently coordinate and integrate multi-disciplinary care for patients with PDAC.

Performing combined EUS-FNA/ERCP efficiently, effectively and safely requires careful coordination of resources within any single endoscopy unit. Either a dual trained operator or readily available separate EUS and ERCP operators are required to perform both segments of the case in a single session. Given that as of 2014 more that 66 advanced endoscopy training positions were available across 58 USA training programs and that the majority offer both EUS and ERCP training, it is unlikely that a lack of dual trained EUS/ERCP operators will be a significant barrier in the years to come. (13) The true challenge is aligning resources such as on-site cytopathology, cross-trained, experienced nursing staff, anesthesia support and recovery room resources to accommodate this approach.

Expeditious care is desirable for all of our patients anticipating a cancer diagnosis. The ability to provide a jaundiced patients with a confirmatory diagnosis, durable biliary drainage and ready them for PDAC therapies within a single procedure session cannot be overstated. While ERCP and/or EUS-FNA may not be necessary in patients with an obstructing pancreatic head mass that is clearly resectable on pre-procedure imaging, this scenario describes a minority of patients. (3, 14) The majority of patients with PDAC are either borderline resectable or unresectable at the time of diagnosis. Patients deemed borderline resectable and of adequate functional status will receive 6–8 weeks of neoadjuvant chemotherapy prior to resection. Finally, a proportion of patients with resectable disease are deemed unresectable at staging laparoscopy. Consequently, the majority of patients with biliary obstruction from PDAC require durable biliary drainage and based on current literature SEMS are the optimal stent for decompression. Comparative studies designed to evaluate the efficacy of SEMS versus plastic stents reveal that SEMS are associated with lower rates of stent failure, unplanned ERCP, fewer procedures and are cost effective. (6, 15) EUS-FNA tissue diagnosis prior to ERCP triages patients for placement of SEMS.

A hypothetical benefit of a combined EUS/ERCP approach is cost savings generated through the elimination of charges associated with a second, discrete session of endoscopy. Based on prior cost analysis studies, EUS or ERCP procedure sessions can generate charges of up to $1000-$6200. (16, 17). Combining EUS/ERCP procedures result in a single session of facility and anesthesia fees. This approach would hypothetically eliminate a second set of expenses otherwise generated by at least one additional, subsequent procedure session. Consequently, combining EUS/ERCP into a single session should offer some degree of relative cost savings when comparing the number of index diagnostic and therapeutic drainage procedures required to manage PDAC with obstructive jaundice. However, we should emphasize that a more detailed cost effectiveness analysis that weighs the expense and propensity for deployment of different types of biliary stents, sensitivity of sampling procedures and the expected rates of stent failure is required to evaluate this question with precision. A nuanced cost analysis of the decision tree for endoscopic management of PDAC would also offer insight into the degree of the expected differences in expense between approaches across a population of patients. While outside the scope of our study, a cost effectiveness study may offer further support for a combined EUS/ERCP index approach if a significant savings in healthcare expenditures is clearly identified.

Additional benefits to a combined approach include the ability to augment the diagnostic yield of the overall procedure session by adding intraductal sampling during the ERCP if EUS-FNA initially fails to be confirmatory. There is also the added capability to convert to EUS assisted biliary access (e.g rendezvous or transduodenal stent placement) if ERCP fails.

Finally, while we do not offer data to this end, it is also our experience that the satisfaction of outpatients with a combined EUS/ERCP is quite high for reasons beyond expeditious throughput to PDAC therapies. The need to arrange transportation and work leave for a single outpatient procedure with a combined approach offers convenience over the logistics required to accommodate multiple procedure sessions.

Limitations:

Study limitations include a retrospective design and a potential element of selection bias. Patients that underwent index EUS-FNA or ERCP procedures at other institutions were intentionally excluded from analysis. Consequently, our data represents experience from a single high volume tertiary referral center where index sampling and decompression were all performed by our team. We excluded patients with index EUS-FNA or ERCP procedures performed outside our center to control for the impact of variables such as delays in outside referrals between procedures, operator experience and/or differences in institutional resources on technical success and outcomes. We recognize that this may limit the generalizability of the findings in our study. Moreover, patients included in our study all had a confirmatory diagnosis of PDAC. This also limits generalizability of conclusions regarding combined EUS-FNA/ ERCP for tissue acquisition and decompression for non-PDAC malignant tumors (e.g. lymphoma, pancreatic neuroendocrine tumor, cholangiocarcinoma, metastases…) and benign conditions (chronic pancreatitis) of the paraduodenal pancreas.

Summary:

In conclusion, combined EUS/ERCP offers similar rates of complications, diagnostic and therapeutic success compared to sequential approaches. SEMS are more often deployed during a combined EUS/ERCP owing to a pre-procedure tissue diagnosis. Combined EUS/ERCP may expedite delivery of PDAC therapies. We believe the option to combine EUS-FNA and ERCP as a single procedure should figure prominently in decision making for the best approach for a patient with a pancreatic mass and obstructive jaundice. (Figure 2). Further studies are needed to assess cost effectiveness and quantify patient satisfaction with this approach.

Figure 2: Proposed decision algorithm for sequence of EUS-FNA and ERCP for patients with a pancreatic mass, obstructive jaundice and no immediate plan for resection.

* Poor ECOG status, life expectancy <3 months, extensive comorbidities, patient and family preferences should be heavily considered

^{^} Either dual trained or separate, but available operators for same session EUS/ERCP, anesthesia support for longer procedure >60 minutes, on-site cytopathology

** Age >60, elevated CA 19–9, highly suspicious imaging (upstream pancreatic duct dilation, pancreatic atrophy, metastatic liver/lung lesions), cachexia/weight loss >5–10%, absence of chronic pancreatitis

Abbreviations: SEMS: self-expanding metallic biliary stents, uncovered or covered, cSEMS: covered self-expanding metallic biliary stents, FNA/B: EUS- fine needle aspiration or biopsy