Abstract
Background:
Endoscopic ultrasound (EUS) and ultrasound (US) are the two primary imaging modalities used to guide pancreatic needle biopsy. This study aimed to compare the diagnostic performance and complications associated with EUS- and US-guided pancreatic biopsies.
Methods:
A total of 2517 consecutive patients who underwent 2583 cases of EUS- or US-guided pancreatic biopsy for focal pancreatic disease between January 2017 and December 2023 at 7 university teaching hospitals were included. The endpoints evaluated were diagnostic inaccuracy, complications, and repeat biopsies for each method. The propensity score matching (PSM) method was used for the analysis.
Results:
For the entire cohort, diagnostic inaccuracies were observed in 7.0% of cases in the EUS-guided biopsy group compared to 3.5% in the US-guided biopsy group, representing a statistically significant difference (P = 0.001). The rates of major (P = 1.000) and minor complications (P = 0.309) were similar between the two groups. The rate of repeat biopsies was significantly higher in the EUS-guided biopsy group than in the US-guided biopsy group (5.0% vs. 2.8%, P = 0.024). However, after balancing lesion characteristics using PSM, no significant differences were observed between the EUS- and US-guided biopsies in diagnostic inaccuracy (7.2% vs. 8.5%, P = 0.600), major complications (0.5% vs. 0%, P = 0.499), minor complications (1.7% vs. 0.2%, P = 0.069), or repeat biopsy rate (5.5% vs. 6.5%, P = 0.656). Subgroup analysis revealed that among patients with exophytic and backward growth pancreatic lesions, both diagnostic inaccuracy (7.1% vs. 33.3%, P = 0.001) and the repeat biopsy rate (10.7% vs. 27.8%, P = 0.029) were significantly lower in the EUS-guided biopsy group compared to the US-guided biopsy group.
Conclusions:
This study confirmed that both EUS-guided and US-guided pancreatic biopsies are safe and effective for patients with focal pancreatic disease. After PSM, diagnostic inaccuracy, repeat biopsy and complication rates were similar, but EUS-guided biopsy was preferred for lesions with exophytic and backward growth morphology.
Keywords: diagnosis, endoscopic ultrasound-guided fine needle aspiration, image-guided biopsy, pancreatic neoplasms, propensity score
HIGHLIGHTS
This multicenter study compared the diagnostic performance and complications of EUS- and US-guided pancreatic biopsies in 2583 cases from 2517 patients with focal pancreatic disease.
Overall, though EUS-guided biopsies initially appeared to demonstrate higher diagnostic inaccuracies (7.0% vs. 3.5%, P = 0.001) and repeat biopsy rates (5.0% vs. 2.8%, P = 0.024) compared to US-guided biopsies, these differences were not observed after propensity score matching (diagnostic inaccuracies: 7.2% vs. 8.5%, P = 0.600; repeat biopsy: 5.5% vs. 6.5%, P = 0.656). Additionally, the complication rates were comparable between the two methods (P > 0.05).
For patients with exophytic and backward growth pancreatic lesions, EUS-guided biopsy significantly reduced both diagnostic inaccuracy (7.1% vs. 33.3%, P = 0.001) and repeat biopsy rates (10.7% vs. 27.8%, P = 0.029) than US-guided biopsy, underscoring its potential clinical advantage in this subgroup.
Introduction
The acquisition of pancreatic tissue is a critical step in diagnosing pancreatic diseases. This can be achieved through an intraoperative biopsy, image-guided percutaneous biopsy, or endoscopic biopsy. It is essential to confirm malignancy in focal pancreatic lesions, particularly when these lesions are borderline resectable or unresectable, prior to initiating nonsurgical or neoadjuvant treatment[1,2]. Currently, endoscopic ultrasound (EUS)- and ultrasound (US)-guided pancreatic biopsy are recommended as first-line methods for sample acquisition in patients with focal pancreatic disease. Both methods demonstrate good diagnostic performance and low incidence of adverse events. EUS-guided pancreatic biopsy, which includes fine-needle aspiration (FNA) and fine-needle biopsy (FNB), is the preferred technique due to its high patient tolerance and low risk of complications[3,4]. Image-guided percutaneous biopsy, such as FNA or core-needle biopsy (CNB), is recommended for obtaining larger tissue specimens, particularly in patients with unresectable pancreatic lesions. These techniques can complement each other when one is not feasible[5]. There has been no head-to-head randomized controlled trial directly comparing the diagnostic efficacy and complications of the two methods. Caymaz et al suggested that the diagnostic yield for solid pancreatic lesions was significantly higher in the US-CNB group than in the EUS-FNA group[6]. Conversely, Volmar et al reported in their observational study that EUS-FNA exhibited a slightly higher sensitivity and accuracy than US- or computed tomography (CT)-guided FNA for pancreatic lesions. After controlling for lesion size, EUS demonstrated greater accuracy than US or CT when evaluating lesions smaller than 3 cm[7]. The emergence of this conflicting evidence, which is of low strength, necessitates further validation in a large-scale, real-world cohort. Therefore, we aimed to compare EUS- and US-guided pancreatic biopsies in terms of the risks of diagnostic inaccuracy, complications, and need for repeat biopsy in a propensity-matched, large-scale cohort with a follow-up period of 1 year.
This cohort study has been reported in line with the STROCSS guidelines[8].
Materials and methods
Patients
Between January 2017 and December 2023, 2663 consecutive patients with focal pancreatic disease underwent 2729 procedures for pancreatic tissue acquisition from focal pancreatic lesions at 7 university teaching hospitals in China. Patients were excluded from the study if they had less than 12 months of follow-up, were lost to follow-up, had incomplete follow-up data, or underwent intraoperative biopsy or fluid aspiration. Ultimately, 2517 patients (with 2583 cases) who underwent EUS- or US-guided pancreatic biopsy were included in the final analysis. All enrolled patients had clear indications for pancreatic biopsy, either to establish a definitive diagnosis or to provide histopathological confirmation for non-surgical treatment. The choice of pancreatic biopsy method was primarily guided by factors such as the characteristics of the pancreatic lesion, adjacent anatomical structures, patient’s body mass index (BMI), the feasibility of a percutaneous approach, and the available equipment and expertise at each center.
Assessment
Baseline information, including age, sex, and BMI, was documented and gathered from electronic medical records. The research also incorporated risk factors such as tumor size, location, and morphology, which were evaluated and recorded by operators and their assistants. The tumor size was measured using prebiopsy US images, presenting the maximum dimension in centimeters. Six radiologists, each with over 10 years of experience in abdominal imaging, independently measured the tumor sizes, and the values were subsequently averaged. The exophytic and backward growth morphology of pancreatic lesions was assessed using CT or MRI. This terminology denotes a tumor originating at the most medial edge of the pancreas, extending rearward beyond it, without extending beyond the anterior margin of the pancreas (Fig. 1). Such lesions may be easily overlooked due to atypical features, such as the absence of typical duct obstruction and distal gland atrophy, or they may present as a perivascular soft tissue cuff encasing the superior mesenteric artery or celiac trunk rather than as a distinct mass[9,10]. The locations of the pancreatic lesions were divided into five categories: head, uncinate process, neck, body, and tail. For pancreatic tumors situated at junctions, such as between the head and neck or the neck and body, classification was determined based on the site selected for biopsy.
Figure 1.
Two examples demonstrating the exophytic and backward growth morphology of pancreatic lesions. (A) Contrast-enhanced CT image, red arrow, and hypovascular lesion and (B) gray-scale ultrasound image, measurement markers, and hypoechoic lesion illustrate the exophytic and backward growth morphology of a pancreatic lesion located in the neck of the pancreas. (C) Gray-scale ultrasound image, red arrow, and hypoechoic lesion and (D) contrast-enhanced ultrasound image, red arrow, and hyperenhanced lesion display a lesion in the pancreatic uncinate process, which exhibits an exophytic and backward growth morphology.
Definitions
Pancreatic lesions were classified as true-positive if diagnosed as pancreatic ductal adenocarcinoma (PDAC), pancreatic neuroendocrine tumor (PNET), solid pseudopapillary tumor of the pancreas (SPTP), other definite malignancies, or benign pancreatic tumors based on histopathologic analysis and consistent with the results of at least 12 months of follow-up. Cases in which the initial pathological findings indicated inflammation, but clinical assessments and imaging strongly suggested malignancy, or where an indeterminate diagnosis was made due to the non-representative nature of the acquired tissue or cells, were categorized as having “undefined results.” A multidisciplinary consultation was conducted to determine the necessity for repeat sample acquisition. Patients without indications for repeat sampling were monitored for a minimum of 12months, taking into account clinical features, imaging (CT or MRI), and tumor marker measurements. Lesions that remained stable in size (defined as ≤20% increase in diameter) with benign imaging characteristics and no suspicious changes in tumor markers were classified as benign. Conversely, lesions exhibiting a >20% increase in diameter and an absolute increase of 5 mm or more, or a suspicious rise in tumor markers, were deemed malignant[11,12]. Following postpuncture examination after at least 12 months of follow-up, pancreatic lesions ultimately diagnosed as benign were categorized as true negative results, whereas those diagnosed as malignant were classified as false negative results.
The diagnostic inaccuracy of both EUS- and US-guided biopsy for focal pancreatic lesions was defined as the ratio of the sum of false-negative and false-positive results to the total number of biopsies. Complications occurring postbiopsy were rated as major or minor according to the Society of Interventional Radiology clinical practice guidelines, which include immediately visible hematomas or bleeding, vasovagal reactions, infections, biopsy-related pancreatitis, organ perforation (duodenum, colon), and arteriovenous fistula. All complications related to pancreatic biopsy that occurred within 30 days were included in the assessment[13].
Statistical analysis
The characteristics of the patients were described by group (EUS-guided biopsy and US-guided biopsy groups). Non-normally distributed continuous variables were reported as medians with interquartile ranges (IQRs) and compared using nonparametric tests. Categorical variables were expressed as percentages and analyzed using the chi-square test. A random forest-based interpolation technique was used to estimate the missing values, which comprised 0–12.7% of the baseline data. PSM analysis was used to reduce bias by equating the two groups based on the following variables: tumor size (<2 cm), tumor location, and exophytic and backward growth morphology. PSM was conducted using the nearest-neighbor 1:1 matching method, with a caliper width set at 0.02. This specific caliper width was chosen based on recommendations from statistical literature[14], indicating that it effectively balances optimal bias reduction while preserving a substantial sample size for subsequent analysis. The diagnostic inaccuracy rate for focal pancreatic lesions was compared between the EUS- and US-guided pancreatic biopsy groups. Additionally, the incidence rates of major and minor complications, as well as the rate of repeat biopsies, were analyzed and compared. Both groups were evaluated using a general model and a PSM model. Subsequently, subgroup analyses were performed to assess the association between biopsy methods and outcomes (diagnostic inaccuracy and repeat biopsy) across various subgroups defined by tumor size, morphology, and location. Statistical significance was set at a P-value of <0.05. Statistical analyses were performed using the SPSS software (version 26.0; IBM SPSS) and R software (version 4.2.3; R Development Core Team).
Results
Baseline characteristics
A total of 2583 procedures were analyzed, with 444 procedures of EUS-guided biopsy and 2139 procedures of US-guided biopsy in the general model. Following PSM, 804 patients were analyzed using 402 EUS-guided biopsy and 402 US-guided biopsy. A flowchart illustrating the patient selection and PSM process is shown in Figure 2. After PSM, the median age of all patients was 64.0 years (IQR: 12.0 years), and 59.0% (474/804) were male. The median BMI was 21.1 kg/m2 (IQR: 4.3 kg/m2). Tumors smaller than 2 cm were observed in 25.1% (202/804) of cases. Pancreatic lesions exhibiting exophytic and backward growth morphology were noted in 13.7% (110/804) of cases. Tumors were located in various regions of the pancreas: 28.4% (228/804) in the head, 16.4% (132/804) in the uncinate process, 10.7% (86/804) in the neck, 15.9% (128/804) in the body, and 28.6% (230/804) in the tail. The characteristics of the entire cohort are presented in Table 1.
Figure 2.
Flowchart of cohort selection.
EUS, endoscopic ultrasound; US, ultrasound; PS, propensity score.
Table 1.
Baseline characteristics and US characteristics of pancreatic lesions-entire cohort (general model and PSM model)1
| Characteristics | General model | PSM model | ||||
|---|---|---|---|---|---|---|
| EUS-guided biopsy (n = 444) | US-guided biopsy (n = 2139) | P value | EUS-guided biopsy (n = 402) | US-guided biopsy (n = 402) | P value | |
| Baseline characteristics | ||||||
| Ages, median (IQR), y | 65 (14) | 65 (13) | 0.921 | 64 (14) | 64 (11) | 0.252 |
| Sex, male | 208/444 (63.1%) | 1249/2139 (58.4%) | 0.071 | 245 (60.9%) | 229 (57.0%) | 0.282 |
| BMI, median (IQR), kg/m2 | 20.97 (4.52) | 21.48 (3.97) | 0.02 | 21.05 (4.54) | 21.07 (4.18) | 0.922 |
| US characteristics of the pancreatic lesion | ||||||
| Tumor size < 2 cm, No (%) | 137 (30.9%) | 154 (7.2%) | <0.001 | 100 (24.9%) | 102 (25.4%) | 0.935 |
| Exophytic and backward growth morphology, No (%) | 82 (18.5%) | 80 (3.7%) | <0.001 | 56 (13.9%) | 54 (13.4%) | 0.918 |
| Tumor location | <0.001 | 0.564 | ||||
| Head, No (%) | 129 (29.1%) | 585 (27.3%) | 118 (29.4%) | 110 (27.4%) | ||
| Uncinate process, No (%) | 63 (14.2%) | 417 (19.5%) | 58 (14.4%) | 74 (18.4%) | ||
| Neck, No (%) | 49 (11.0%) | 239 (11.2%) | 47 (11.7%) | 39 (9.7%) | ||
| Body, No (%) | 70 (15.8%) | 588 (27.5%) | 64 (15.9%) | 64 (15.9%) | ||
| Tail, No (%) | 133 (30.0%) | 310 (14.5%) | 115 (28.6%) | 115 (28.6%) | ||
EUS: endoscopic ultrasound, US: ultrasound, PSM: propensity score matching, IQR: interquarterly range, BMI: body mass index.
In our study, the EUS-guided biopsy techniques included FNA (25G, 22G, and 19G) and FNB (25G and 22G), with the aim of obtaining histological specimens from the lesions. At least three passes were performed for each case, with additional passes determined by the macroscopic evaluation of the specimens. Due to the limited availability of FNB devices, the number of FNB cases was relatively low, accounting for 4.5% (20 out of 444). In US-guided biopsy, which comprised both FNA (18G, 19G, and 21G) and CNB (18G), and with the increasing clinical demand for high-quality tissue specimens from pancreatic lesions, CNB accounted for the majority of cases, making up 82.6% (1766 out of 2139). In these cases, at least 2 passes were performed, with further passes determined based on the macroscopic assessment of specimen quality.
Diagnostic inaccuracy
The pathology report following pancreatic biopsy revealed a range of findings, including PDAC, PNET, SPTP, inflammation, other malignancies, benign tumors, and undefined results. After a minimum follow-up period of 12 months , the final diagnoses for all cases were established. Details of the pathological and final diagnoses in both the general and PSM models are presented in Table 2.
Table 2.
The final diagnosis of pancreatic lesions-entire cohort (general model and PSM model)2
| General model | PSM model | |||
|---|---|---|---|---|
| EUS-guided biopsy (n = 444) | US-guided biopsy (n = 2139) | EUS-guided biopsy (n = 402) | US-guided biopsy (n = 402) | |
| Biopsy pathology | ||||
| PDAC | 324 | 1826 | 289 | 328 |
| PNET | 12 | 59 | 11 | 15 |
| SPTP | 7 | 4 | 7 | 1 |
| Inflammation | 21 | 87 | 20 | 15 |
| Other malignancies | 9 | 69 | 8 | 8 |
| Benign tumors | 0 | 12 | 0 | 0 |
| Undefined results | 71 | 82 | 67 | 35 |
| Final diagnosis | ||||
| PDAC | 349 | 1898 | 312 | 363 |
| PNET | 15 | 65 | 14 | 15 |
| SPTP | 7 | 4 | 7 | 1 |
| Inflammation | 43 | 87 | 42 | 13 |
| Other malignancies | 12 | 69 | 11 | 8 |
| Benign tumors | 13 | 16 | 11 | 2 |
| Undefined results | 5 | 0 | 5 | 0 |
EUS: endoscopic ultrasound, US: ultrasound, PSM: propensity score matching, PDAC: pancreatic ductal adenocarcinoma, PNET: pancreatic neuroendocrine tumor, SPTP: solid pseudopapillary tumor of the pancreas.
In the general model, diagnostic inaccuracies occurred in 31 cases (7.0%) in the EUS-guided biopsy group and 75 cases (3.5%) in the US-guided biopsy group, revealing a significant difference between the 2 groups (P = 0.001). The odds ratio (OR) for inaccurate diagnostic events in the EUS-guided biopsy group, compared to the US-guided biopsy group, was 2.066, with a 95% confidence interval (CI) of 1.342–3.181. Following PSM, diagnostic inaccuracies were observed in 29 cases (7.2%) in the EUS-guided biopsy group and 34 cases (8.5%) in the US-guided biopsy group, showing no significant difference in the incidence of diagnostic inaccuracies after PSM (P = 0.600) (Table 3). The OR for inaccurate diagnostic events in the EUS-guided biopsy group relative to the US-guided biopsy group was 0.842, with a 95% CI of 0.502–1.410.
Table 3.
Detailed overview of major and minor complications in the general and PSM models3
| Characteristics | General model | PSM model | ||||
|---|---|---|---|---|---|---|
| EUS-guided biopsy | US-guided biopsy | P value | EUS-guided biopsy | US-guided biopsy | P value | |
| Major complication, No (%) | 2 (0.5%) | 11 (0.5%) | 1.000 | 2 (0.5%) | 0 (0%) | 0.499 |
| 1 case of duodenal perforation | 1 case of delayed gastroduodenal artery pseudoaneurysm | 1 case of duodenal perforation | ||||
| 1 case of significant abdominal blood accumulation | 5 cases of significant abdominal or pelvic blood accumulation | 1 case of significant abdominal blood accumulation | ||||
| 5 cases of puncture-related pancreatitis | ||||||
| Minor complication, No (%) | 7 (1.6%) | 21 (1.0%) | 0.309 | 7 (1.7%) | 1 (0.2%) | 0.069 |
| 7 cases of postoperative pain | 11 cases of postoperative pain | 7 cases of postoperative pain | 1 case of mild postoperative intra-abdominal hematomas | |||
| 9 cases of mild postoperative intra-abdominal hematomas or hepatic subcapsular hematomas | ||||||
| 1 case of vasovagal reaction | ||||||
EUS: endoscopic ultrasound, US: ultrasound, PSM: propensity score matching
Major complications
In the entire cohort, the incidence of major complications was 0.5% (13/2583). These complications included one case of duodenal perforation requiring surgical repair, one case of delayed gastroduodenal artery pseudoaneurysm formation managed with interventional embolization, and six cases of significant abdominal or pelvic blood accumulation necessitating blood transfusions. Additionally, puncture-related pancreatitis occurred in five cases, resulting in prolonged hospital stays. No puncture-related life-threatening events, disabling events, or patient deaths were reported. Major complications were observed in 2 cases (0.5%) in the EUS-guided biopsy group and 11 cases (0.5%) in the US-guided biopsy group in the general model, indicating no significant difference between the 2 groups (P = 1.000). Compared with the US-guided biopsy group, the OR for major complications in the EUS-guided biopsy group was 0.875, with a 95% CI of 0.193–3.963. After PSM, the incidence of major complications was recorded in two cases (0.5%) in the EUS-guided biopsy group and zero cases (0.0%) in the US-guided biopsy group. There was no significant difference in the incidence of major complications after PSM (P = 0.499) (Table 3). Compared with the US-guided biopsy group, the OR for major complications in the EUS-guided biopsy group was 1.005, with a 95% CI of 0.998–1.012.
Minor complications
In the entire cohort, the incidence of minor complications was 1.1% (28/2583). These complications included 18 patients experiencing postoperative pain, 9 patients developing mild postoperative intra-abdominal hematomas or hepatic subcapsular hematomas, and 1 patient experiencing a vasovagal reaction, all of which improved with conservative treatment. Minor complications occurred in 7 cases (1.6%) in the EUS-guided biopsy group and in 21 cases (1.0%) in the US-guided biopsy group in the general model, showing no significant difference between the 2 groups (P = 0.309). Compared with the US-guided biopsy group, the OR of minor complications in the EUS-guided biopsy group was 1.616, with a 95% CI of 0.683–3.824. After PSM, minor complications occurred in seven cases (1.7%) in the EUS-guided biopsy group and in one case (0.2%) in the US-guided biopsy group. No significant difference in the incidence of minor complications was observed after PSM (P = 0.069) (Table 3). Compared with the US-guided biopsy group, the OR of minor complications in the EUS-guided biopsy group was 7.106, with a 95% CI of 0.870–58.025.
Repeat biopsy
In the general model, the occurrence of repeat biopsy was noted in 22 cases (5.0%) in the EUS-guided biopsy group and 59 cases (2.8%) in the US-guided biopsy group, demonstrating a significant difference between the 2 groups (P = 0.024). When comparing the US-guided biopsy group, the OR for repeat biopsy in the EUS-guided biopsy group was 1.838, with a 95% CI of 1.114–3.032. Following PSM, the incidence of repeat biopsy was recorded in 22 cases (5.5%) in the EUS-guided biopsy group and in 26 cases (6.5%) in the US-guided biopsy group. No significant difference in the incidence of repeat biopsy was observed after PSM (P = 0.656) (Table 3). In this matched analysis, the OR for repeat biopsy in the EUS-guided biopsy group compared with the US-guided biopsy group was 0.837, with a 95% CI of 0.466–1.503.
Subgroup analyses
Subgroup analysis results are presented in Table 4. When stratified by lesion location (head, uncinate process, neck, body, and tail) and tumor size (<2 cm and ≥2 cm), no significant differences in diagnostic inaccuracy or repeat biopsy risk were observed between the EUS-guided and US-guided biopsy groups. However, in the subgroup of patients with exophytic and backward growth morphology, a significant interaction was observed between the biopsy method and outcomes, with diagnostic inaccuracy rates of 7.1% for EUS-guided biopsy versus 33.3% for US-guided biopsy (P = 0.001) and repeat biopsy rates of 10.7% versus 27.8%, respectively (P = 0.029). These findings indicate that, in this subgroup, EUS-guided biopsy is associated with a significantly lower risk of diagnostic inaccuracy and repeat biopsy compared to US-guided biopsy. Figure 3 illustrates the relationship between the biopsy method and the rates of diagnostic inaccuracy and repeat biopsy across the various subgroups.
Table 4.
Subgroup analysis based on lesion location, size, and morphology4
| OR | CI lower | CI higher | P value | |
|---|---|---|---|---|
| Diagnostic inaccuracy vs. biospy method | ||||
| Head | 0.438 | 0.158 | 1.209 | 0.140 |
| Uncinate process | 0.878 | 0.312 | 2.469 | 1.00 |
| Neck | 1.721 | 0.298 | 9.937 | 0.685 |
| Body | 0.543 | 0.151 | 1.954 | 0.530 |
| Tail | 2.791 | 0.721 | 10.801 | 0.215 |
| Exophytic and backward growth morphology | 0.154 | 0.048 | 0.493 | 0.001 |
| Non-exophytic and backward growth morphology | 1.616 | 0.847 | 3.083 | 0.151 |
| Tumor size < 2 cm | 0.719 | 0.277 | 1.871 | 0.631 |
| Tumor size ≧ 2 cm | 0.900 | 0.487 | 1.664 | 0.757 |
| Repeat biopsy vs. biospy method | ||||
| Head | 0.564 | 0.179 | 1.780 | 0.397 |
| Uncinate process | 0.611 | 0.175 | 2.139 | 0.549 |
| Neck | 1.429 | 0.319 | 6.396 | 0.724 |
| Body | 0.644 | 0.173 | 2.402 | 0.744 |
| Tail | 4.108 | 0.452 | 37.330 | 0.370 |
| Exophytic and backward growth morphology | 0.312 | 0.111 | 0.879 | 0.029 |
| Non-exophytic and backward growth morphology | 1.485 | 0.679 | 3.249 | 0.334 |
| Tumor size < 2 cm | 0.435 | 0.146 | 1.302 | 0.192 |
| Tumor size ≧ 2 cm | 1.133 | 0.555 | 2.313 | 0.856 |
EUS: endoscopic ultrasound, US: ultrasound, PSM: propensity score matching, OR: odds ratio, CI: confidence interval.
The bold values in Table 4 indicate statistically significant results (p < 0.05). They are highlighted to emphasize the findings that met our predefined threshold for significance.
Figure 3.
Subgroup analysis evaluating the relationship between biopsy method and diagnostic inaccuracy and repeat biopsy.
OR, odds ratio; CI, confidence interval; EB, exophytic and backward.
Discussion
This large-scale, propensity score-matched cohort study investigated the risks of diagnostic inaccuracy, complications, and repeat biopsies associated with EUS- and US-guided biopsies for focal pancreatic lesions. In the general model, the rates of diagnostic inaccuracy and repeat biopsy for EUS-guided biopsy were found to be significantly higher than those for US-guided biopsy. However, no significant differences in the incidence of major and minor complications were observed between EUS- and US-guided pancreatic biopsies. After PSM, no differences were detected between EUS- and US-guided pancreatic biopsy in terms of diagnostic inaccuracy, major complications, minor complications, and repeat biopsy. In subgroups of patients with lesions showing exophytic and backward growth morphology, the incidence rates of diagnostic inaccuracy and repeat biopsy for US-guided biopsy were higher than those for EUS-guided biopsy.
EUS-guided needle biopsy was first proposed by Vilmann et al in the 1990s[15]. Since then, it has rapidly developed into a method for diagnosing submucosal lesions of the digestive tract and extramural lesions adjacent to the digestive tract, including those in the retroperitoneum, pancreas, and adrenal gland. Various biopsy methods are commonly employed[16,17]. EUS-FNA has become an indispensable technique for sample acquisition, especially for the diagnosis of focal pancreatic lesions, as recommended by multiple practice guidelines[2,18]. Previous studies have demonstrated that EUS-FNA possesses high diagnostic efficiency for solid pancreatic tumors, with reported diagnostic accuracy ranging from 65% to 96%, sensitivity from 54% to 95%, and specificity from 71% to 100%[19,20]. With the advancement of precision medicine, the demand for pathology and requirements for pancreatic specimens are increasing. Due to the fact that EUS-FNA is primarily utilized to obtain cells or fluid, and the volume of tissue retrieved is limited, it does not adequately meet the current clinical needs for pathology. To address the limitations of EUS-FNA, various new biopsy needles have been developed in recent years to procure more histological specimens for biopsy under EUS, referred to as EUS-FNB[21]. The distinguishing feature of EUS-FNB is the specialized design of the needle tip, which may include grooves or alterations in shape. Reports indicate that the diagnostic sensitivity of EUS-FNB for PDAC ranges from 70% to 100%, with a specificity between 86% and 100%, highlighting its advantages over EUS-FNA[22,23]. The National Comprehensive Cancer Network (NCCN) guidelines recommend the use of EUS-FNB for pancreatic lesion biopsy. When a pathological diagnosis is achieved, it is advisable to perform additional punctures to preserve tissue samples for further analyses, such as genotype analysis, single-cell sequencing, and new drug sensitivity testing[1].
Percutaneous pancreatic biopsy is most commonly performed under CT or US guidance. Su et al demonstrated that the diagnostic accuracy of CT-guided percutaneous pancreatic biopsy was 93.4%, with a sensitivity of 93.2% and a specificity of 100%. Additionally, this study reported an overall complication rate of 8.7% with no instances of serious bleeding or needle tract implantation events[24]. In contrast, another single-center retrospective study indicated that the diagnostic accuracy of CT-guided percutaneous pancreatic biopsy was 89.9%, with a sensitivity of 88.8%, specificity of 100%, and overall complication rate of 20.2%[25]. Compared to CT-guided biopsy, US-guided pancreatic biopsy offers advantages such as real-time monitoring, complete visibility of the needle insertion process, and absence of radiation exposure, along with a lower complication rate than that associated with CT guidance[26]. A previous study reported that the overall diagnostic accuracy, sensitivity, and specificity of US-CNB/FNA for focal pancreatic lesions were 94.4%, 93.8%, and 100%, respectively. Specifically, the accuracy, sensitivity, and specificity of US-FNA were 93.7%, 93.1%, and 100%, respectively, whereas those of US-CNB were 95.4%, 94.7%, and 100%, respectively[7,27–29]. Although the diagnostic performance of US-FNA was marginally lower than that of US-CNB, the difference was not statistically significant. The European Federation of Societies for Ultrasound in Medicine and Biology (EFSUMB) recommends that US-guided biopsy should be prioritized for locally advanced pancreatic tumors without surgical indication. If percutaneous approach is not feasible, EUS-guided needle biopsy should be considered. For resectable and borderline resectable pancreatic tumors with atypical imaging characteristics, EUS-guided needle biopsy is advised[5]. Our study compared the diagnostic inaccuracy rates of two commonly used pancreatic needle biopsy techniques for focal pancreatic lesions that shared equal diagnostic efficacy.
This study confirms that in solid pancreatic lesions, the application of pancreatic needle biopsy, guided by US and EUS, is safe. The incidence of both major and minor complications following EUS- and US-guided pancreatic biopsy was lower, with no significant difference observed between the two groups. The literature indicates that the primary complications associated with pancreatic biopsy include acute pancreatitis, bleeding, infection, intestinal perforation, needle tract seeding and metastasis, and pancreatic leakage[30]. The serious complication rate for US-CNB is estimated to be between 2.6% and 21.0%, whereas that for US-FNA ranges from 1.5% to 20.0%. The overall complication rate for EUS-FNA is approximately 0.98–3.4%[31,32]. Previous studies have reported that the incidence of needle tract seeding and metastasis in US-FNA and EUS-FNA of pancreatic tumors ranges from approximately 0.15% to 3.4%, with a higher prevalence observed in US-FNA[33]. Micames et al examined 89 patients who underwent pancreatic FNA under US guidance and found a needle tract implantation rate of 16.3%[34]. The potential for implantation metastasis after percutaneous pancreatic biopsy remains a significant concern. Some studies suggest that EUS-FNA can approach the lesion more closely and has a shorter puncture path, allowing the biopsy needle to quickly retract into the endoscope’s forceps channel after obtaining the sample, thereby reducing the tissue residence time. Consequently, the likelihood of needle tract seeding is considered lower in EUS-FNA than in US-FNA[35]. Conversely, other studies suggest that the use of coaxial needles may mitigate the risk of needle tract seeding after percutaneous biopsy; however, high-quality evidence to support this claim is currently lacking[22,36]. In our study, which had a follow-up period of 12 months, no cases of needle tract seeding following pancreatic biopsies were observed. Although this finding aligns with the low incidence reported in the literature, it is crucial to address this aspect more thoroughly. We did not actively monitor for needle tract seeding as a primary outcome, as the retrospective design of our study and the relatively limited follow-up period may have hindered the detection of such rare events. We recommend conducting prospective, randomized controlled trials with standardized protocols in future research to validate these findings. This approach would help reduce variability in biopsy techniques and outcomes, providing more robust data on the safety and efficacy of different biopsy methods.
Our study suggested that in the general model, the rate of repeat biopsy was significantly higher in the EUS-guided pancreatic biopsy than in the US approach; however, after balancing the size, morphology, and tumor location of the lesions between the two groups through PSM, the difference in the probability of repeat biopsy between the groups was not statistically significant. Several factors contributed to the inconsistency between the pathological diagnosis obtained from EUS-guided needle biopsy and the clinical diagnosis, which include the absence of tumor cells, as well as cases where abnormal cells were present but could not be definitively diagnosed. Results from a recent meta-analysis support the recommendation for repeat EUS-FNA in cases of pancreatic masses following initial non-diagnostic or inconclusive biopsy results. This study reported a specificity of 98% and a low false-positive rate, indicating that repeat EUS-FNA significantly enhances the diagnostic yield, and the use of rapid on-site evaluation (ROSE) was also recommended[37]. A case-control study suggested that repeated EUS-FNB sampling is required in patients with solid pancreatic lesions and a first negative EUS-FNB[38].
The pancreatic lesion demonstrated an exophytic and backward growth morphology, extending posteriorly within the pancreatic parenchyma and easily adhering to surrounding tissues such as the duodenal wall, adjacent vessels, and other structures. Previous research has indicated that imaging-guided biopsies targeting the isolated perivascular soft tissue cuff in patients suspected of having pancreatic carcinoma are both feasible and safe, aligning with the characteristics of exophytic and backward growth morphology[39]. Our study emphasizes the clinical benefits of using EUS-guided biopsy over US-guided percutaneous biopsy for pancreatic lesions exhibiting exophytic and backward growth morphology. Due to the EUS-guided approach’s lower diagnostic inaccuracy and decreased need for repeat biopsies, it has the potential to be prioritized in clinical guidelines for managing such cases. This preference is largely due to the challenges faced with percutaneous visualization and localization of these lesions via transabdominal US imaging, where the puncture pathway is often complex and lengthy. The presence of the intestinal tract, pancreatic parenchyma, and a dilated pancreatic duct pose significant obstacles to percutaneous needle puncture. In contrast, EUS provides a more direct and definitive approach to accessing these lesions, enhancing the efficiency of sample acquisition. This suggests that EUS-guided biopsy could improve diagnostic precision and efficiency, reducing patient discomfort and risks associated with repeat procedures. Integrating these findings into clinical decision-making frameworks could enhance patient management strategies and optimize resource use in health care settings. Further research could solidify these recommendations, potentially leading to updates in guidelines that acknowledge these advantages in the management of pancreatic lesions with exophytic and backward growth morphology.
This study had several limitations. First, it was a short-term study; although the follow-up period of 1 year is the longest reported on this issue to date, a longer study duration would allow for more definitive conclusions. Second, this study was conducted retrospectively without randomization, which may have introduced selection bias. To address this limitation, we employed PSM analyses. However, we acknowledge that even with PSM, some residual confounding factors may persist. In this study, the choice of puncture technique was influenced by various factors, resulting in varied proportions of pancreatic sampling methods under EUS and US guidance, with differences observed in needle type, gauge, and number of passes. Nevertheless, all these techniques shared the common objective of obtaining tissue strips from pancreatic lesions for pathological diagnosis. To overcome this limitation and explore the impact of biopsy technique on diagnostic performance, further prospective, randomized controlled studies need to be designed. Finally, this retrospective multicenter study included an unbalanced number of cases from each center. The uneven case volume across the seven centers may introduce center-specific biases, potentially skewing the overall study results. Moreover, the disproportionate case distribution might affect the statistical power of the PSM, potentially leaving residual confounding from centers with fewer cases.
Conclusion
This study confirms that both EUS- and US-guided pancreatic biopsies are safe and effective for patients with focal pancreatic disease. Initially, EUS-guided biopsy demonstrated a higher risk of diagnostic inaccuracies and a greater need for repeat biopsies; however, after PSM, the rates of diagnostic inaccuracies, repeat biopsies, and complications were comparable between the two methods. Notably, EUS-guided biopsy was preferred for lesions exhibiting exophytic and backward growth morphology. These findings provide compelling evidence for selecting the optimal pancreatic biopsy method for patients with focal pancreatic lesions requiring tissue sampling.
Footnotes
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
Contributor Information
Weilu Chai, Email: 1516059@zju.edu.cn.
Weina Wan, Email: 694417372@qq.com.
Tian’an Jiang, Email: tiananjiang@zju.edu.cn.
Ethical approval
This study was approved by the Institutional Review Board of the First Affiliated Hospital of the Zhejiang University School of Medicine [(2025B) IIT Ethics Approval No. 0080].
Consent
The informed consent was waived by the Institutional Review Board of the First Affiliated Hospital of the Zhejiang University School of Medicine due to its retrospective design [(2025B) IIT Ethics Approval No. 0080].
Sources of funding
This work was supported by the Joint Fund of Zhejiang Provincial Natural Science Foundation (LKLY25H180002), and the Medical and Health Research Project of Zhejiang Province (2024KY1002).
Author contributions
T.J.: conceptualization; W.C., Q.L., D.X., K.L., W.W., L.Y., R.L., T.J.: data curation; W.C.: formal analysis; W.C.: funding acquisition; W.C., Q.L., D.X., K.L., W.W., L.Y., R.L., T.J.: investigation; W.C., T.J.: methodology; W.C., Q.L., D.X., K.L., W.W., L.Y., R.L., T.J.: project administration; W.C., Q.L., D.X., K.L., W.W., L.Y., R.L., T.J.: resources; W.C.; software; T.J.: supervision; W.C.: visualization; W.C.: writing – original draft; and T.J.: writing – review and editing.
Conflicts of interest disclosure
The authors declare that they have no competing interests.
Guarantor
All authors.
Research registration unique identifying number (UIN)
Provenance and peer review
Not commissioned, externally peer-reviewed.
Data availability statement
All patient image and text data are not publicly available due to ethical restrictions, but may be made available to qualified researchers on reasonable request from the corresponding author.
Acknowledgements
We would like to thank Yunlu Bai, Wuyongga Bao, Na Feng, Qingyang Kong, and Xitong Cao for their invaluable support in data collection.
References
- [1].Tempero MA, Malafa MP, Al-Hawary M, et al. Pancreatic adenocarcinoma, version 2.2021, NCCN clinical practice guidelines in oncology. J National Compr Cancer Network 2021;19:439–57. [DOI] [PubMed] [Google Scholar]
- [2].D’Onofrio M, de Sio I, Mirk P, et al. SIUMB recommendations for focal pancreatic lesions. J Ultrasound 2020;23:599–606. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].García JI, Noia JL, Iglesias HL, et al. EUS-guided sampling for the cyto-histological diagnosis of solid pancreatic tumors: EUS-FNA with ROSE vs EUS-FNB. analysis of a prospective registry. Endoscopy 2023;55:S279. [Google Scholar]
- [4].Aadam AA, Amick AE, Shah JN, et al. A multicenter prospective randomized controlled cross-over trial comparing endoscopic ultrasound (EUS)-guided fine needle aspiration (FNA) and fine needle biopsy (FNB) for pancreatic and non-pancreatic masses. Gastrointestinal Endoscopy 2014;79:AB188–AB189. [Google Scholar]
- [5].Sidhu PS, Brabrand K, Cantisani V, et al. EFSUMB guidelines on interventional ultrasound (INVUS), Part II. diagnostic ultrasound-guided interventional procedures (Short Version). Ultraschall Med 2015;36:566–80. [DOI] [PubMed] [Google Scholar]
- [6].Caymaz I, Afandiyeva N. Diagnostic evaluation of solid pancreatic lesions: endoscopic ultrasound-guided fine needle aspiration versus percutaneous ultrasound-guided core needle biopsy. Cardiovasc Intervent Radiol 2023;46:1596–602. [DOI] [PubMed] [Google Scholar]
- [7].Volmar KE, Vollmer RT, Jowell PS, Nelson RC, Xie HB. Pancreatic FNA in 1000 cases: a comparison of imaging modalities. Gastrointest Endosc 2005;61:854–61. [DOI] [PubMed] [Google Scholar]
- [8].Agha RA, Mathew G, Rashid R, et al. Revised strengthening the reporting of cohort, cross-sectional and case-control studies in surgery (STROCSS) guideline: an update for the age of artificial intelligence. Prem J Sci 2025;10:100081. [Google Scholar]
- [9].Al-Hawary MM, Francis IR, Chari ST, et al. Pancreatic ductal adenocarcinoma radiology reporting template: consensus statement of the society of abdominal radiology and the american pancreatic association. Gastroenterology 2014;146:291–304. [DOI] [PubMed] [Google Scholar]
- [10].Gong XH, Xu JR, Qian LJ. Atypical and uncommon CT and MR imaging presentations of pancreatic ductal adenocarcinoma. Abdom Radiol (NY) 2021;46:4226–37. [DOI] [PubMed] [Google Scholar]
- [11].Eisenhauer EA, Therasse P, Bogaerts J, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009;45:228–47. [DOI] [PubMed] [Google Scholar]
- [12].Wu W, Jing X, Xue GQ, et al. A multicenter randomized controlled study of contrast-enhanced US versus US-guided biopsy of focal liver lesions. Radiology 2022;305:721–28. [DOI] [PubMed] [Google Scholar]
- [13].Tam AL, Findeiss L, Dake MD, et al. Standard 2.0: methodology Update. J Vasc Interv Radiol 2018;29:1347–49. [DOI] [PubMed] [Google Scholar]
- [14].Noseworthy PA, Attia ZI, Behnken EM, et al. Artificial intelligence-guided screening for atrial fibrillation using electrocardiogram during sinus rhythm: a prospective non-randomised interventional trial. Lancet 2022; 400:1206–12. [DOI] [PubMed] [Google Scholar]
- [15].Vilmann P, Jacobsen GK, Henriksen FW, Hancke S. Endoscopic ultrasonography with guided fine needle aspiration biopsy in pancreatic disease. Gastrointest Endosc 1992;38:172–73. [DOI] [PubMed] [Google Scholar]
- [16].Dumonceau JM, Polkowski M, Larghi A, et al. Indications,results, and clinicalimpact of endoscopic ultrasound (EUS)-guided sampling in gastroenterology: European Society of Gastrointestinal Endoscopy (ESGE) Clinical Guideline. Endoscopy 2011;43:897–912. [DOI] [PubMed] [Google Scholar]
- [17].Rastogi A, Wani S, Gupta N, et al. A prospective, single-blind,randomized, controlled trial of EUS-guided FNA with and without a stylet. Gastrointest Endosc 2011;74:58–64. [DOI] [PubMed] [Google Scholar]
- [18].Wang W, Shpaner A, Krishna SG, et al. Use of EUS-FNA in diagnosing pancreatic neoplasm without a definitive mass on CT. Gastrointest Endosc 2013;78:73–80. [DOI] [PubMed] [Google Scholar]
- [19].Eloubeidi MA, Chen VK, Eltoum IA, et al. Endoscopic ultrasound-guided fine needle aspiration biopsy of patients with suspected pancreatic cancer: diagnostic accuracy and acute and 30-day complications. Am J Gastroenterol 2003;98:2663–68. [DOI] [PubMed] [Google Scholar]
- [20].Gress F, Gottlieb K, Sherman S, Lehman G. Endoscopic ultrasonography guided fine-needle aspiration biopsy of suspected pancreatic cancer. Ann Intern Med 2001;64:459. [DOI] [PubMed] [Google Scholar]
- [21].Chen YI, Chatterjee A, Berger R, et al. Endoscopic ultrasound (EUS)-guided fine needle biopsy alone vs. EUS-guided fine needle aspiration with rapid onsite evaluation in pancreatic lesions: a multicenter randomized trial. Endoscopy 2021;54:4–12. [DOI] [PubMed] [Google Scholar]
- [22].Moutinho-Ribeiro P, Iglesias-García J, Gaspar R, Macedo G. Early pancreatic cancer — the role of endoscopic ultrasound with or without tissue acquisition in diagnosis and staging. Digestive Liver Dis 2019;51:4–9. [DOI] [PubMed] [Google Scholar]
- [23].Mohamed WT, Jahagirdar V, Jaber F, et al. Endoscopic ultrasound-guided fine-needle biopsy versus aspiration for tissue sampling adequacy for molecular testing in pancreatic ductal adenocarcinoma. Cancers 2024;16:761–761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Su YY, Liu YS, Chao YJ, Chiang NJ, Yen CJ, Tsai HM. Percutaneous computed tomography-guided coaxial core biopsy for the diagnosis of pancreatic tumors. J Clin Med 2019;8:1633. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Hsu MY, Pan KT, Chen CM, et al. CT-guided percutaneous core-needle biopsy of pancreatic masses: comparison of the standard mesenteric/retroperitoneal vs the trans-organ approaches. Clin Radiol 2016;71:507–12. [DOI] [PubMed] [Google Scholar]
- [26].Bhatti I, Ojo D, Dennison AR, Rees Y, Elabassy M, Garcea G. Percutaneous pancreatic biopsies-still an effective method for histologic confirmation of malignancy. Surg Laparosc Endosc Percutan Tech 2016;26:334–37. [DOI] [PubMed] [Google Scholar]
- [27].Xin Y, Yang Y, Chen Y, Wang Y, Cao XJ, Zhou X. Safety and efficacy of ultrasound-guided percutaneous coaxial core biopsy of pancreatic lesions: a retrospective study. J Ultrasound 2021;24:269–77. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Huang Y, Shi J, Chen YY, Li K. Ultrasound-guided percutaneous core needle biopsy for the diagnosis of pancreatic disease. Ultrasound Med Biol 2018;44:1145–54. [DOI] [PubMed] [Google Scholar]
- [29].D’Onofrio M, De Robertis R, Barbi E, et al. Ultrasound-guided percutaneous fine-needle aspiration of solid pancreatic neoplasms: 10-year experience with more than 2,000 cases and a review of the literature. Eur Radiol 2016;26:1801–07. [DOI] [PubMed] [Google Scholar]
- [30].Wang KX, Ben QW, Jin ZD, et al. Assessment of morbidity and mortality associated with EUS guided FNA: a systematic review. Gastrointest Endosc 2011;73:283–90. [DOI] [PubMed] [Google Scholar]
- [31].Matsubara J, Okusaka T, Morizane C, Ikeda M, Ueno H. Ultrasound guided percutaneous pancreatic tumor biopsy in pancreatic cancer: a comparison with metastatic liver tumor biopsy, including sensitivity, specificity, and complications. J Gastroenterol 2008;43:225–32. [DOI] [PubMed] [Google Scholar]
- [32].Katanuma A, Maguchi H, Yane K, et al. Factors predictive of adverse events associated with endoscopic ultrasound-guided fine needle aspiration of pancreatic solid lesions. Dig Dis Sci 2013;58:2093–99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [33].Caturelli E, Rapaccini GL, Anti M, Fabiano A, Fedeli G. Malignant seeding after fine-needle aspiration biopsy of the pancreas. Diagn Imaging Clin Med 1985;54:88–91. [PubMed] [Google Scholar]
- [34].Micames C, Jowell PS, White R, et al. Lower frequency of peritoneal carcinomatosis in patients with pancreatic cancer diagnosed by EUS-guided FNA vs. percutaneous FNA. Gastrointestinal Endoscopy 2003;58:690–95. [DOI] [PubMed] [Google Scholar]
- [35].Kosugi C, Furuse J, Ishii H, et al. Needle tract implantation of hepatocellular carcinoma and pancreatic carcinoma after ultrasound-guided percutaneous puncture: clinical and pathologic characteristics and the treatment of needle tract implantation. World J Surg 2003;28:29–32. [DOI] [PubMed] [Google Scholar]
- [36].Fotiadis N, De Paepe KN, Bonne L, et al. Comparison of a coaxial versus non-coaxial liver biopsy technique in an oncological setting: diagnostic yield, complications and seeding risk. Eur Radiol 2020;30:6702–08. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [37].Lisotti A, Frazzoni L, Fuccio L, et al. Repeat EUS-FNA of pancreatic masses after nondiagnostic or inconclusive results: systematic review and meta-analysis. Gastrointestinal Endoscopy 2020;91:1234–41. [DOI] [PubMed] [Google Scholar]
- [38].Camus B, Pellat A, Rouquette A, et al. Diagnostic yield of repeat endoscopic ultrasound-guided fine needle biopsy for solid pancreatic lesions. Cancers 2023;15:3745–3745. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [39].O’Connell MJ, Paulson EK, Jaffe TA, Ho LM. Percutaneous biopsy of periarterial soft tissue cuffs in the diagnosis of pancreatic carcinoma. Abdom Imaging 2004;29:115–19. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
All patient image and text data are not publicly available due to ethical restrictions, but may be made available to qualified researchers on reasonable request from the corresponding author.