QUALITY OF EUS-GUIDED LIVER BIOPSIES IS NOT ON PAR WITH THOSE OF PERCUTANEOUS AND TRANSJUGULAR METHODS: ROOM FOR IMPROVEMENT?

Background and Aims
Surgical resection is a standard treatment option for management of pancreatic cysts. The newly introduced EUS-guided ethanol ablation (EUS-ablation) has been performed as an alternative to pancreatectomy in selected patients and is expected to reduce the high rates of short-and long-term morbidity of pancreatectomy. However, no comparative studies have been carried out between EUS-ablation and pancreatectomy for pancreatic cysts.
Methods
We reviewed the prospectively collected pancreatic cyst database to analyze consecutive patients with unilocular or oligolocular pancreatic cysts requiring treatment who underwent EUS-ablation or pancreatectomy between January 2015 and July 2021. After excluding cases in which malignancies were suspected preoperatively, we performed 1:1 match using propensity score matching based on age, gender, cyst size, cyst morphology (unilocular or oligolocular), cyst location, and American Society of Anesthesiologists (ASA) physical status classification between EUS-ablation and pancreatectomy. Complete remission rates, adverse events, mortality, and occurrence of diabetes were evaluated.
Results
A total of 620 patients (310 EUS-ablation, 310 pancreatectomy) were selected after propensity score matching. During follow-up, rate of complete resolution of pancreatic cysts was lower in EUS-ablation group (77% vs. 100%; P = 0.001). However, the EUS-ablation group showed lower rates of early adverse events (EUS-ablation, 12.6% vs. pancreatectomy, 46.5%; P = 0.001), late adverse events (1% vs. 16%; P = 0.001), occurrence of diabetes (0% vs. 18%; P = 0.001), unplanned readmissions (1% vs. 15%; P = 0.001), and cumulative morbidities at 5 years (11.3% vs. 37.5%; P = 0.001) than did the pancreatectomy group. There were significant differences in the severity of the adverse events: mild (10% vs. 16%; P = 0.018); moderate (2% vs. 15%; P = 0.001); severe (0% vs. 3%; P = 0.004). The EUS-ablation group required interventions for adverse events less frequently than the pancreatectomy group did (1% vs. 10%; P = 0.001). Procedure or surgery-related mortality rates in EUS-ablation group and pancreatectomy group was 0% and 0.3%, respectively.
Conclusions
In management of unilocular or oligolocular pancreatic cyst, EUS-guided ethanol ablation might be a good alternative to pancreatectomy in selected patients regarding morbidity and mortality benefits.

Aims
Endoscopic ultrasound-guided radiofrequency ablation (EUS-RFA) is emerging as safe and effective treatment for pancreatic neoplasms. We aimed to compare EUS-RFA and surgical resection for the treatment of pancreatic insulinoma (PI).

Methods
Patients with non-metastatic PI who underwent EUS-RFA at 23 centers or surgical resection at eight high-volume pancreatic surgery institutions between 2014 and 2022 were identified and outcomes compared using a propensity-matching analysis. Patients with multiple nodules or diagnosed with MEN-1 were excluded. Primary outcome was safety. Adverse events (AEs) were defined according to international definitions and graded using the AGREE and Clavien-Dindo classifications. Secondary outcomes were treatment effectiveness, length of hospital stay, and rate of recurrence with need of additional treatment after EUS-RFA. Treatment effectiveness was defined as complete disappearance of symptoms related to insulin secretion and was assessed including possible multiple treatments in the EUS-RFA group.

Results
Overall, 328 patients were identified and 304 were included (111 EUS-RFA and 193 surgical resections). Using propensity score matching, 178 patients were selected and 89 allocated in each group (1:1). The two groups were evenly distributed in terms of age, sex, Charlson comorbidity index, ASA score, BMI, distance between lesion and main pancreatic duct, lesion site, size, and grade (Figure 1). Lesions were in the head, body, or tail in 69 (38.8%), 66 (37.1%), and 43 (24.1%) cases, respectively. Mean lesion size was 13.5 ± 4.8mm. Tumor grading was G1 in 145 (81.5%) cases, G2 in 13 (7.3%), and remained unknown in 20 (11.2%) cases. The rate of AEs was 18.0% and 61.8% after EUS-RFA and surgical resection, respectively (p<0.001). The most common AEs was pancreatic fistula in the surgical group (42.7%) and acute pancreatitis in the EUS-RFA group (10.1%) (Figure 2). No severe (AGREE ≥3) AEs were observed in the EUS-RFA group compared with 7.8% after surgery (p=0.0002). Clinical efficacy was 100% after surgery and 95.5% after EUS-RFA (p=0.160). However, the mean duration of follow-up time was shorter in the EUS-RFA group (median 23 (IQR 14-31) vs 37 (IQR 17.5-67) months, p<0.0001). Hospital stay was significantly longer in the surgical group (11.1 ± 9.7 vs 3.0 ± 2.5 days in the EUS-RFA group, p<0.0001). Fifteen (16.9%) lesions recurred after EUS-RFA after a mean time of 9.5 months, and uneventfully underwent to second EUS-RFA (n=11) or surgical resection (n=4). No distant/nodal metastases appeared during follow-up.

Conclusion
EUS-RFA is safer than surgery and highly effective for the treatment of PI. When a lesion recurs after EUS-RFA it can be safely managed with a second EUS-RFA treatment or surgical resection. If confirmed in a randomized study, EUS-RFA could be suggested as first-line therapy for sporadic PI.

Background and aims:
Pain often complicates pancreatic cancer management. EUS-CPN is usually performed using dehydrated alcohol and bupivacaine. We hypothesized that alcohol dilution by bupivacaine reduces the alcohol-induced neurolysis and subsequent pain relief, with no impact on periprocedural pain. We therefore undertook a randomized trial of EUS-guided CPN with alcohol alone vs alcohol + bupivacaine. These are interim results from our ongoing non-inferiority trial and were requested by our data safety monitoring board. The study was approved by research ethics board with approval number of 2022-9969, 21.151 and registered on clinicaltrial.gov; registration number NCT04951804.

Methods:
Patients with inoperable painful pancreatic body cancer were randomised by sealed envelope to bilateral EUS-CPN with alcohol (10cc x 2) or alcohol (10ml x2) preceded by Bupivacaine 0.5% (10ml, central). Patients and the research nurse were blinded to treatment arm. Narcotic usage was measured as mg of morphine equivalent. Follow up included a 7-point Likert pain scale, Global rating scale of change (GRSC), and narcotic use at 3, 7, 30, 60, 90 and 120 days, survival, and the incidence of adverse events (ASGE criteria).

Results:
36 patients were enrolled from November 2021 to December 2022; of whom 30 achieved complete 120 day follow up. The results are summarized in Table 1. Patients were comparable at baseline. There were no differences in intra-procedural variables. Post-procedure, the bupivacaine group showed significantly lower BP at discharge. The alcohol only group tended to have lower pain scores and higher reduction in pain and morphine use at T120 days. There were significantly more mild adverse events in the bupivacaine group.

Conclusions:
Preliminary results of this ongoing randomized trial suggest that adding bupivacaine to alcohol for EUS-CPN significantly increases mild adverse events and may reduce effectiveness in terms of pain relief and need for narcotics.

Background: Endoscopic ultrasound (EUS) guidance has become an increasingly utilized technique for liver biopsy (LB). Although the standard needle gauge (G) used for EUS-LB is 19G, some reports have used a 22G FNB device. A prospective randomized comparison of 19G FNB vs. 22G FNB for EUS-LB has not previously been done. We performed a prospective randomized trial to evaluate the specimen adequacy of a 19G Franseen tip needle compared to a 22G Franseen tip needle for EUS-LB. Methods: Forty-two consecutive (21 in each group) patients who were referred for EUS-LB were prospectively enrolled in the study, with randomization of the type of needle used (19G Franseen tip needle and 22G Franseen tip needle, Acquire, Boston Scientific, Marlborough, MA). The study was registered at ClinicalTrials.gov (NCT04806607) and received local IRB approval. In cases when the endoscopist felt that there was macroscopically inadequate specimen with the 22G FNB, an additional pass with the 19G FNB was done. Bilobar EUS-LB was performed with heparinized wet suction using the 1 pass, 3 actuations per lobe technique for both groups. Descriptive statistics were computed for all variables. Student’s t-test was used for continuous variables and Fisher Exact test was used to compute categorical variables. To avoid the family-wise error rate, a Bonferroni correction was utilized. A p-value of < 0.0125 was considered statistically significant. Results: The average age of the entire cohort was 51 +/- 14.3 years and 57.1% of patients were female. All patients were Caucasian and 95.5% of the liver biopsies were done for abnormal liver enzymes. Five patients in the 22G FNB group required an addition biopsy with the 19G FNB due to macroscopically assessed inadequacy of the biopsy. The median pre-processing length of the longest piece (LLP) was 21.5 +/- 6.3 mm with 19G FNB compared to 9.4 +/- 5.5 mm with 22G FNB (p < 0.001). Similarly, the post-processing specimen was significantly longer with 19G FNB at 17.4 +/- 5.2 mm compared to 6.8 +/- 4.2 mm with the 22G FNB (p < 0.001) (Fig 1.) The aggregate specimen length (ASL) was significantly larger with the 19G FNB at 54.9 +/- 16.3 mm compared to 22G FNB at 19.8 +/- 14.5 mm (p < 0.001) (Fig 2). There were no adverse events in either group. Incidence of post-procedure pain/discomfort was similar in both groups (14% for 19G vs 10% for 22G, p = 0.99), and no study participants had pain on post-discharge follow-up. Conclusion: Liver biopsy cores obtained with a 19G FNB needle are superior to a 22G FNB in terms of length of the longest piece and aggregate specimen length. There is considerably more fragmentation of the 22G cores during tissue processing compared to the 19G. A 19G FNB device should be preferentially used over the 22G FNB for EUS-LB.

Background
The field of endohepatology has been growing as a useful application of advanced and therapeutic endoscopy for diagnosis and management of chronic liver disease and its complications. Endoscopic ultrasound shear-wave measurement (EUS-SWM) is a currently available but underutilized technology for assessment of parenchymal stiffness, including liver stiffness.

Methods
This is a prospective study of bilobar EUS-SWM to characterize liver stiffness in a cohort of patients referred for EUS-guided liver biopsy in a tertiary liver transplant center. A push-pulse generated by echoendoscope and propagated through liver parenchyma was used to provide quantitative assessment of liver stiffness. The reliability of EUS-SWM, its diagnostic performance, and thresholds for defining stages of liver stiffness were evaluated using concomitant bilobar liver biopsy as the gold-standard.

Results
A total of 740 EUS-SWM was performed in 38 patients (Figure 1a). Measurements were performed at a median depth of 1.1 cm (range 0.9-1.4) in both left and right lobes of the liver and each measurement covered an area of 1 cm by 0.5 cm. Measurements had a median validity of 100% (IQR 83-100%) corresponding to the percentage of the target area which was used in calculating stiffness in each measurement. There was no statistically significant difference between EUS-SWM in right and left liver lobes (Wilcoxon matched-pairs p=0.096, Figure 1b). The median of all bilobar measurements with a validity of 70% or more was used as EUS-SWM in each patient for final analysis. EUS-SWM showed a statistically significant and consistent association with fibrosis on liver biopsy (Kruskal-Wallis p=0.003, Figure 2a). EUS-SWM was significantly higher in patients with endoscopic portal hypertension (p=0.001), but it was not associated with presence of inflammation on liver biopsy (p=0.877). EUS-SWM showed an accuracy of 0.77 (ROC AUC=0.77, 95%CI 0.60-0.95) for diagnosis of F0, and 0.83 (ROC AUC 0.83, 95%CI 0.70-0.96) for diagnosis of F3 or F4 fibrosis (Figure 2a). Using Youden's criteria the optimal cutoff for diagnosis of F0 fibrosis was less than 10.3 kPa (sensitivity of 82% and specificity of 75%), and the optimal cutoff for diagnosis of F3-4 fibrosis was more than 11.8 kPa (sensitivity of 100% and specificity of 74%).

Conclusion
EUS-SWM can be used to predict liver stiffness with accuracy comparable with trans-coetaneous elastography. It should be considered as part of comprehensive endoscopic evaluation of liver disease with the potential to decrease need for liver biopsy is selected patients.

Background:
US-guided liver biopsies (EUS-LB) are increasingly being performed to acquire liver tissue. However, existing data regarding the safety and efficacy of EUS-LB compared to percutaneous (PC-LB) and transjugular (TJ-LB) liver biopsies has conflicting results. In addition, most of the available data is from higher volume centers. We performed this study to compare the safety and efficacy of EUS-LB with PC-LB and TJ-LB at a center with multiple providers using variable techniques.

Methods:
We conducted a retrospective review of all liver biopsies performed at the University of Alabama at Birmingham from February 2021 to February 2022. All non-targeted liver biopsies performed in adults (>18 years) were included. Surgical biopsies and lesion-targeted liver biopsies were excluded. The extracted data included information about biopsy route, fragmentation, number of portal tracts, histologic adequacy, longest length of core (LLC), and complications. Biopsies were considered histologically adequate if labeled as such by the pathologist. Portal tracts were considered adequate when greater than 11 complete portal triads were identified, as defined by AASLD guidelines. Any new symptoms or signs of well-defined complications of liver biopsy not consistent with the clinical course prior to biopsy were considered as complications.

Results:
A total of 244 biopsies (34 EUS-LB, 140 PC-LB, and 70 TJ-LB) were included in the study. Among these, 85.6% were labeled adequate for histological evaluation, 15.2% were fragmented, and 77.5% had more than 11 portal tracts. The mean LLC was 1.67cm. Histologic adequacy, fragmentation, portal tract adequacy, LLC, and complications were significantly different among the three groups. Histologic adequacy was higher for PC-LB (92.9%, p <0.001) and TJ-LB (87.1%, p <0.001) compared to EUS-LB (52.9%). Portal tract adequacy was also significantly higher for PC-LB (84.7%, p <0.001) and TJ-LB (85 %, p <0.001) compared to EUS-LB (39.4%). Fragmentation was higher for EUS-LB (58.8%) compared to PC-LB (3.6%, p <0.001) and TJ-LB (17.1 %, p <0.001). EUS-LB had longer LLC (1.85cm) than TJ-LB (1.34cm, p<0.001) but was similar to PC-LB (1.80 cm, p=0.649). Complications with EUS-LB (17.7%) were higher compared to PC-LB (5.7%, p=0.021) but were not different when compared to TJ-LB (5.7%, p=0.053). However, it should be noted that only 17% of complications with EUS needed further intervention, while this was 50% in PC-LB and 75% in TJ-LB.

Conclusion:
At a center utilizing multiple techniques for EUS-LB, EUS-LB does not appear to perform as well as PC-LB or TJ-LB. The variable practice patterns among different providers could contribute to this discrepancy. Further studies investigating the impact of standardization of endoscopic techniques, devices, and tissue processing are warranted.