Recent Advances in Endoscopic Ultrasound-Guided Tissue Acquisition

Abstract

Endoscopic ultrasound-guided tissue acquisition (EUS-TA) constitutes a pivotal technique in the histopathological evaluation of intra-abdominal and mediastinal lesions. Recent advances enhance the diagnostic yield and clinical utility of EUS-TA. These include the advent of through-the-needle accessories, the integration of molecular and genomic profiling, the implementation of image-enhanced acquisition techniques, and the application of artificial intelligence-assisted image interpretation. Furthermore, the incorporation of digital pathology and telecytology platforms in conjunction with rapid on-site evaluation provides additional value for improving real-time diagnostic assessment. These innovations underscore the evolving role of EUS-TA as a highly adaptable and precision-oriented tool toward individualized patient care in the precision medicine era.

Keywords

endoscopic ultrasound - tissue acquisition - through-the-needle biopsy - molecular profiling - image-enhanced EUS - rapid on-site evaluation - telecytology

Introduction

Endoscopic ultrasound-guided tissue acquisition (EUS-TA) is a highly effective procedure essential for the pathological diagnosis of gastrointestinal lesions and adjacent structures. Endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) provides 80 to 95% diagnostic accuracy for intra-abdominal masses. However, EUS-FNA typically yields only cytological samples with limited cellularity, which may be insufficient for immunohistochemical and molecular analyses that require well-preserved tissue architecture and cellular morphology.

To overcome this limitation, various specialized needle designs were developed for endoscopic ultrasound-guided fine needle biopsy (EUS-FNB), including reverse-bevel, forward-bevel, Franseen, fork-tip, and three-prong asymmetric tip needles, to obtain core samples with preserved architecture, contributing diagnostic accuracy up to 82 to 97%.[1] [2] The latest needle generation enhances tissue acquisition performance, enabling advanced analyses such as immunohistochemistry and molecular profiling. Furthermore, microforceps biopsy further enhances diagnostic capability by sampling cyst walls and allowing for more accurate diagnosis beyond cyst fluid analysis, while contrast-enhanced EUS, elastography, and rapid on-site evaluation (ROSE) optimize lesion targeting and sample adequacy. Additional innovations, including telecytology and digital pathology platforms, streamline workflows and improve diagnostic reliability. These advancements enhance the precision and clinical impact of EUS-TA in diagnosing and supporting individualized cancer management.

Evolution of Needles, Devices, and Accessories

Needles

Diagnostic Performance of FNA and FNB Needles

The evolution of needle selection has shifted from FNA toward FNB. The FNB needle demonstrates superior diagnostic accuracy and enables reliable histopathological diagnosis, as evidenced by randomized controlled trials (RCTs) and meta-analyses.[3] [4] The side-fenestrated needle, Franseen needle, fork-tip needle, and three-prong asymmetric tip are examples of available needles designed for EUS-FNB ([Table 1]). Recent studies by Young Bang et al revealed that the fork-tip and Franseen needles significantly yielded a higher cellularity in a single biopsy pass compared with reverse-bevels and Menghini-tip needles, achieving 90% diagnostic accuracy for pancreatic cancer.[5] The large noninferiority RCT confirmed the superior performance of fork-tip and Franseen needles over side-fenestrated needles, even without ROSE.[6] A 2022 network meta-analysis further supported these results, highlighting higher diagnostic accuracy and sample adequacy of Franseen and fork-tip needles over reverse-bevel and FNA needles.[7] Therefore, both Franseen and fork-tip needles yielded a comparable sensitivity and accuracy approaching 97 to 98% even without ROSE based on current evidence.

Table 1 The summary of needles for endoscopic ultrasound-guided tissue acquisition

EUS-FNA		Needle type	Aspiration needle
		Commercial name	EchoTip, EZ shot, Expect, ClearView, Beacon
		Needle size	19, 22, 25 gauge
EUS-FNB	First-generation needles
		Needle type	Tru-cut needle
		Commercial name	Quick-Core
		Needle size	19 gauge
	Second-generation needles
		Needle type	Reverse bevel
		Commercial name	EchoTip ProCore
		Needle design	An opening at the side of the needle, with a hollow reverse bevel architecture, which allows cutting of the tissue during backward retraction of the needle
		Needle size	19, 22, 25 gauge
	Third-generation needles
		Needle type	Fork-tip type
		Needle design	Two protruding asymmetrical sharp points and six distal cutting edges
		Commercial name	SharkCore
		Needle size	19, 22, 25 gauge
		Needle type	Franseen type
		Needle design	Three symmetrically distributed needle points and cutting edges
		Commercial name	Acquire, SonoTip TopGain, EchoTip ClearCore, EchoTip AcuCore
		Needle size	Acquire, SonoTip TopGain: 19, 22, 25 gauge EchoTip ClearCore, EchoTip AcuCore: 22 gauge
		Needle type	Forward bevel
		Needle design	A tapered continuous cutting edge with an opening at the side of the needle and a hollow forward bevel architecture, which allows cutting of the tissue during antegrade movement of the needle
		Commercial name	ProCore
		Needle size	20 gauge
		Needle type	Three-prong asymmetric tip
		Needle design	Three-prong asymmetric tip has one tip longer than the other two
		Commercial name	Trident
		Needle size	19, 22, 25 gauge

Abbreviations: EUS-FNA, endoscopic ultrasonography-guided fine needle aspiration; EUS-FNB, endoscopic ultrasonography-guided fine needle biopsy.

Comparison of 22- and 25-Gauge FNB Needles

Current evidence indicates that the 22-gauge Franseen needle achieves 94 to 97% accuracy for pancreatic masses without ROSE.[2] [8] A prospective study found that a macroscopic visible core (MVC) >4 mm was significantly associated with tissue adequacy using the 22-gauge Franseen needle.[1] A network meta-analysis reported higher pooled sensitivity for 22-gauge compared with 25-gauge needles (94.7 vs. 87.9%), with 100% specificity of both needles. However, pairwise meta-analysis showed no significant differences, partly due to the unblinding of included studies and limited head-to-head comparison.[9] Small non-RCT reported higher diagnostic yield with the 25-gauge Franseen needle compared with the 22-gauge for diagnosing solid pancreatic lesions (SPLs) (98 vs. 88%, p = 0.105) but required more needle passes to obtain adequate cell blocks during ROSE. However, the 25-gauge needle was preferred by endoscopists for lesions in the pancreatic head and uncinate to accommodate scope torque during transduodenal access.[10] The RCT by Tomoda et al demonstrated noninferiority of the 25-gauge compared with the 22-gauge Franseen needle in tissue adequacy (70.5 vs. 78.4%, p < 0.001) and accuracy (86.4 vs. 89.8%, p = 0.180) for diagnosing SPLs. However, the 15% noninferiority margin may limit clinical applicability.[11]

Current evidence indicates a paradigm shift from FNA toward end-cutting FNB needles, particularly the Franseen and fork-tip designs. Both needle types offer high and comparable diagnostic performance in prospective studies. Among 22- and 25-gauge small needles, the diagnostic performance and adequacy were comparable, and the macroscopic on-site evaluation (MOSE) can further enhance tissue adequacy. The 25-gauge Franseen needle offers greater flexibility and improves diagnostic yield for pancreatic head and uncinate lesions via the transduodenal route. However, it requires more needle passes to obtain adequate cellularity compared with a 22-gauge needle. Therefore, balancing needle flexibility and core sampling ability is essential.

EUS-Guided Through-the-Needle Accessories

Microforceps Biopsy for Pancreatic Cystic Neoplasm

Cyst fluid analysis is recommended in various guidelines to characterize pancreatic cystic neoplasms (PCNs), as accurate diagnosis can impact surveillance or surgical decisions. A cyst fluid carcinoembryonic antigen (CEA) cutoff at 192 ng/mL distinguishes mucinous from non-mucinous cysts with 75% sensitivity and 84% specificity, whereas amylase levels <250 U/L effectively exclude pseudocysts with 98% sensitivity.[12] Although cyst fluid cytology can detect malignancy with 25 to 88% sensitivity, the cyst fluid CEA does not correlate with dysplasia grading or malignancy and yields inadequate cellularity.[12] [13] Furthermore, discrepancies between preoperative EUS findings and surgical pathology remain significant (19–34%), especially in small cysts and cystic neuroendocrine tumors.[14] To address these limitations, EUS-guided through-the-needle biopsy (EUS-TTNB) has emerged as a tool for determining cyst type and malignancy risk. EUS-TTNB is indicated when the result is expected to alter management decisions and is not recommended for cysts <15 mm or when findings would not impact treatment. Careful patient selection is essential, particularly for lesions connecting to the main pancreatic duct, including branch-duct intraductal papillary mucinous neoplasm, which are associated with higher complication rates. The adverse event ranged from 8.6 to 10.1%, including self-limited intracystic bleeding, acute pancreatitis, abdominal pain, and infection.[15]

Two commercially available microforceps for TTNB have evolved, including Moray microforceps (Steris, USA) and the Micro Bite (MTW Endoskopie Manufakture, Germany). The Moray and Micro Bite provided maximal jaw opening widths of 4.3 and 3.2 mm, respectively. Both microforceps have a sheath diameter of 0.8 mm. For the Micro Bite, the oval shape of the forceps and tooth with a spoon-shaped mouth was designed for adequate tissue sampling.[16]

EUS-TTNB involves puncturing the cyst with a 19-gauge FNA needle, through which microforceps are advanced under EUS guidance to target the cyst wall, septa, or mural nodules ([Fig. 1]). Tissue is grasped and retracted until a “tent sign” confirms adequate sampling, and the obtained specimen is submitted for histological evaluation. However, the techniques for EUS-TTNB, including forceps preloading, number of passes, number of bites per pass, number of collected specimens, specimen handling, and specimen processing, remain unstandardized. Experts recommended partial cyst fluid aspiration for biochemical, cytologic, and molecular analysis before TTNB to reduce blood contamination, create enough space for forceps manipulation inside the cyst, and decrease cyst wall tension, thereby improving tissue adequacy.[17]

Fig. 1 Through-the-needle biopsy using microforceps for diagnosis of pancreatic cystic neoplasm under endoscopic ultrasound guidance.

Recent meta-analyses reported 93.2 to 98.5% technical success rates of EUS-TTNB, defined by the presence of MVC.[15] Pooled histological adequacy and diagnostic accuracy ranged from 85.3 to 86.7% and 68.6 to 82.8%, respectively.[18] [19] A meta-analysis showed 76.6% pooled sensitivity and 98.9% specificity for diagnosing PCNs.[15] For mucinous cysts, EUS-TTNB yielded 80.0 to 88.6% pooled sensitivity and 94.7 to 95.0% specificity.[19] [20] Several meta-analyses have confirmed the superiority of EUS-TTNB over EUS-FNA with cytology in terms of sample adequacy and diagnostic accuracy, particularly in mucinous cysts.[18] [19] [21] Combining EUS-TTNB with other techniques, including needle-based confocal laser endomicroscopy (nCLE), cyst morphology assessment, and cyst fluid analysis, increased diagnostic yield up to 93% compared with TTNB or nCLE alone.[22] Furthermore, the network meta-analysis showed that EUS-TTNB and nCLE had significantly higher superiority index for differentiating mucinous cysts than cyst fluid biochemical and molecular analysis.[23]

Brush Cytology

Intracystic cytologic brushing has been explored for decades to improve the diagnostic yield for PCNs. The EchoBrush (Cook Endoscopy, United States) is designed for cytologic brushing through a 19-gauge FNA needle. After cyst wall puncture under EUS guidance, the brush is advanced from the needle tip into the cyst, placed against the inner surface wall, and rotated to collect epithelial and mucinous cells without to-and-fro movement ([Fig. 2A]). However, techniques for brush cytology vary among several studies. Initial aspiration of cyst fluid until total or 50% collapse before brushing, repeatedly to-and-fro movement for at least 30 seconds to enhance cyst wall contact, followed by final fluid aspiration for cytology until total cyst collapse was also proposed ([Fig. 2B]). The direct smears were prepared by rolling the brush on slides and fixed with alcohol. The tip of the cytologic brush was cut and placed in a sterile container or special cytologic solution and sent for cytologic evaluation. The EchoBruch yielded significant diagnostic materials, including epithelial cells, intracellular mucin, and mucinous cells, compared with conventional EUS-FNA in several studies.[24] [25] However, a multicenter RCT showed no improvement in overall accuracy compared with EUS-FNA.[26]

Fig. 2 Different techniques for through-the-needle intracystic cytologic brushing of pancreatic cystic neoplasm under endoscopic ultrasound guidance. (A) Brush is advanced from the needle tip into the cyst, placed against the inner surface wall, and rotated to collect epithelial and mucinous cells without to-and-fro movement. (B) Initial aspiration of cyst fluid until total or partial cyst collapses before brushing, repeatedly to-and-fro movement, followed by final fluid aspiration for cytological analysis.

The significant complication rate ranges from 8 to 10%, comprising self-limited pancreatitis, pseudocyst formation, or intracystic bleeding. Major bleeding is rare (0.02–0.06%) but can be severe, including fatal retroperitoneal hemorrhage and hemosuccus pancreaticus. Angioembolization was required in patients under anticoagulant therapy despite discontinuation, and the INR was normalized before the procedure.[24]

In 2023, a novel loop-shaped nitinol brush was developed for use through a 22-gauge FNA needle, featuring a 2 cm fully expanded loop.[27] A porcine RCT demonstrated promising efficacy and safety by brushing and dispersing cyst wall cells into the fluid for aspiration.[28] However, clinical validation in humans is still required.

Application of Molecular Profiling

The shift toward precision oncology has heightened demands on tissue acquisition techniques to obtain adequate specimens for molecular analysis in gastrointestinal, liver, and pancreatobiliary cancers. Comprehensive genetic profiling (CGP) is a molecular analysis of multiple genes to detect potential alterations and actionable mutations of cancers and provides insights into cancer biology and biomarkers. CGP is essential in precision oncology and personalized medicine as it contributes critical information for targeted therapy and prognosis prediction for cancer patients. Next-generation sequencing (NGS) facilitates CGP by rapidly identifying and sequencing mutations from limited tissue samples. Integration of NGS in EUS-TA benefits the treatment planning, prognosis prediction, and detection of hereditary cancers according to the National Comprehensive Cancer Network (NCCN) guidelines in 2025 ([Fig. 3]). Some centers incorporate EUS-TA with genomic profiling into molecular tumor boards for shaping individualized therapy, particularly for advanced and unresectable diseases, where targeted and personalized approaches are becoming the mainstay of treatment. EUS-TA for NGS yielded an 83.9% success rate, with no significant difference between FNA and FNB in the recent 2024 meta-analysis.[29] The use of large 19- and 22-gauge needles, sampling from pancreatic body or tail lesions, and the presence of MOSE were associated with higher success rates for NGS.[30] A comparative study demonstrated that EUS-FNB using a 22-gauge Franseen needle yielded higher nucleic acid content than EUS-FNA for SPLs, lymph nodes, gastric subepithelial tumors, and other abdominal masses.[31] Similarly, Kandel et al reported higher specimen adequacy for genomic profiling and greater DNA yield with 19- or 22-gauge fork-tip needles compared with 25-gauge FNA needles for pancreatic cancer patients.[32]

Fig. 3 Summary of actionable gene mutations in liver and pancreatobiliary malignancies and clinical implications based on the National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology 2025. MSI, microsatellite instability; dMMR, mismatch repair deficiency; TMB, tumor mutational burden; TMB-H, tumor mutational burden-high; MEN, multiple endocrine neoplasm; VHL, von Hippel-Lindau; NF, neurofibromatosis; MCN, mucinous cystic neoplasm; IPMN, intraductal papillary mucinous neoplasm.

Although NGS can sequence multiple genes from limited specimens, adequate tissue obtained by EUS-TA is essential. The adequate tissue criterion for NGS was defined based on different sequencing platforms ([Table 2]). The adequacy rates for NGS ranged from 58 to 100% in SPLs and 93 to 97% in PCNs.[33] EUS-FNB using 22-gauge Franseen or fork-tip needles with suction technique from pancreatic, lymph node, liver, and biliary tract lesions yielded a 90.8% adequacy for NGS based on the Foundation One CDx criteria.[33] In locally advanced pancreatic cancer, 22-gauge Franseen needles with stylet-slow pull and fanning technique provided 97% adequacy, and MVC length larger than 30 mm identified by MOSE was significantly associated with adequate tissue for NGS by the Foundation One CDx system (AUROC 0.74).[34] [35]

Table 2 The adequate specimen criteria for comprehensive genomic profiling in different platforms

	OncoGuide NCC Oncopanel system	Foundation One CDx system	MSK-IMPACT system
Specimen preparation	Five unstained slides (10 µm thick)	One FFPE block and one H&E-stained slide or 10 unstained slides (positively-charged and unbaked, 4–5 µm thick) and one H&E-stained slide	20 unstained slides (4–5 µm thick)
Required tumor surface area	≥16 mm2 (4 mm2 is acceptable)	≥25 mm2	≥25 mm2
Required tumor nuclei[a] (%)	≥20% DNA content ≥200 ng	≥20%	>10%

Abbreviations: FFPE, formalin-fixed paraffin-embedded; H&E, hematoxylin and eosin; MSK, Memorial Sloan Kettering Cancer Center; NCC, National Cancer Center, Japan.

^a Defined by the total number of tumor cells divided by the total number of all cells with nuclei.

The large-bore 19-gauge needles have shown feasibility for CGP in several studies. A recent prospective cohort reported comparable tissue adequacy for NGS between 22-gauge and 19-gauge FNA needles, achieving 89 to 90% in pancreatic adenocarcinoma (PDAC).[36] For FNB, the 19-gauge Franseen needle provided 63.6% adequacy by NCC Oncopanel criteria, with 100% sensitivity and specificity for PDAC. However, a higher rate of 9% adverse event rate was observed, including pancreatitis, bleeding, and infection.[37] Other studies reported 56.0 to 72.5% sample adequacy of 19-gauge Franseen needles with no statistical difference compared with surgical specimens.[38] [39] Furthermore, it also yielded higher adequacy for NGS than 22-gauge FNA and 22-gauge Franssen needles based on NCC Oncopanel criteria.[38] [39] However, the overall tissue adequacy was lower (39.2%) as the studies predominantly comprised metastatic lesions, and optimal cellularity thresholds for NGS from primary and metastatic lesions remained unclear.[39] An ongoing prospective trial has been conducted to compare tissue adequacy for NGS from primary and metastatic lesions in unresectable PDAC using the NCC Oncopanel system.[40]

For intrahepatic cholangiocarcinoma, EUS-FNA/FNB achieved 77.1% tissue adequacy for NGS. Both 22- and 19-gauge FNB needles yielded adequate tissues for NGS based on Foundation One CDx, NCC Oncopanel, and MSK-Impact criteria.[41] Furthermore, targeting the primary lesion, selecting a lesion larger than 30 mm, and performing more than three needle passes were significantly associated with adequate tissue for NGS.

In conclusion, EUS-TA is a valuable modality for obtaining CGP and supporting precision oncology for advanced cancer patients. Current evidence indicates that EUS-FNB with a large-bore 19-gauge Franseen needle provides adequate tissue for molecular profiling, especially in primary PDAC. However, the aspiration technique, the number of to-and-fro movements, and the number of needle passes remain non-standardized. Well-designed prospective trials comparing the performance of EUS-FNA and FNB are warranted to optimize and standardize EUS-TA techniques for molecular profiling.

Integration of Image Enhancement in Tissue Acquisition

EUS-TA is a valuable tool for obtaining a pathological diagnosis of various cancers. Diagnostic accuracy and adequacy can be optimized through appropriate acquisition techniques and ancillary studies, including immunohistochemistry and molecular analyses. EUS-elastography and contrast-enhanced harmonic EUS are advanced imaging modalities that further augment diagnostic yield by facilitating targeted sampling areas with the highest likelihood of malignancy, especially in cases where conventional EUS-TA may yield equivocal results.

EUS-Elastography

EUS-elastography is a real-time imaging modality that assesses tissue stiffness and enables accurate differentiation between benign and malignant lesions, particularly in the pancreas and lymph nodes. Strain elastography estimates lesion stiffness by measuring strain in response to compression and displays it in a different color pattern. The red color indicates the soft tissue part, and blue represents the hard parts within the target lesion that raise suspicion of malignancy. This facilitates targeted EUS-TA by directing sampling to the most diagnostically relevant areas ([Fig. 4]).

Fig. 4 Endoscopic ultrasound-guided tissue acquisition of a necrotic lymph node under elastography guidance. (A) Blue and red colors represent hard and soft areas, respectively. (B) Lymph node puncture with a needle trajectory toward the hard area.

Combined EUS-FNA and real-time elastography achieved 94.4% diagnostic accuracy, 93.4% sensitivity, and 100% specificity for PDAC diagnosis.[42] In a small prospective study, Ohno et al demonstrated the significant association of the fibrous stroma in core tissue histology of PDAC and the hard area visualized by strain EUS-elastography. However, the diagnostic performance of EUS-FNA/FNB of SPLs using elastography guidance showed no significant differences compared with conventional B-mode.[43] In terms of tissue cellularity, tissue quantity, and core tissue length, elastography-guided EUS-FNA also showed no significant difference between hard and soft areas in diagnosing SPLs.[43] [44] A recent large retrospective study in 2024 reported that EUS-TA from high-stiffness areas of SPLs was associated with malignant cells on histopathology. However, the diagnostic yield was comparable between high- and low-stiffness groups (93.3 vs. 92.6%). Moreover, sampling high-stiffness areas required more needle passes.[45]

In conclusion, EUS-elastography offers potential advantages in guiding tissue acquisition by targeting high-stiffness regions and enhancing diagnostic yield for malignancy. Nonetheless, its application is not mandatory, and further well-designed prospective studies are needed to compare diagnostic performance with conventional EUS-FNB.

Contrast-Enhanced Harmonic EUS

Contrast-enhanced harmonic EUS (CH-EUS) improves the visibility of targeted lesions by detecting harmonic signals generated from microbubbles after exposure to an ultrasound beam through peripheral vein contrast injection. The harmonic signals from microbubbles are stronger than those from surrounding tissues, enhancing the depiction of microvasculature and parenchymal perfusion on EUS images, and aiding in the assessment of the vascularity of the targeted lesion.

Tumor necrosis and dense desmoplastic stroma in PDAC and pancreatic neuroendocrine tumors (PNETs) can diminish the diagnostic yield and accuracy of EUS-TA, posing significant challenges for obtaining adequate tissue for pathological diagnosis. On CH-EUS, punctures at non-enhancing or avascular areas often correspond with necrotic tissue on histopathology, yielding a lower diagnostic sensitivity for PDAC compared with enhancing areas (72.9 vs. 94.3%) where the malignant cells are predominantly found.[46] CH-EUS also facilitates tumor identification in cases with chronic pancreatitis, diffusely infiltrating carcinomas, or recent acute pancreatitis, where the surrounding pancreatic parenchyma may obscure the tumor visibility. After contrast administration, heterogeneous enhancement typically indicates a tumor, whereas homogeneous enhancement suggests inflammatory changes in the peritumoral parenchyma.[47] EUS-FNA at heterogeneous enhancement area identified by CH-EUS achieved higher sensitivity and sample adequacy for PDAC than conventional EUS-FNA (84.9 vs. 68.8% and 81.2 vs. 67.3%).[48] Therefore, avoiding puncture of non-enhancing or homogeneous areas is essential to optimize diagnostic performance for SPLs.

The diagnostic sensitivity and specificity of CH-EUS-guided FNA for targeting biopsy of SPLs range from 81.6 to 90% and 100%, respectively. However, these findings are nonsignificant compared with conventional EUS-FNA.[49] Several studies reported that CH-EUS-guided FNA required significantly fewer needle passes to obtain adequate tissue samples for the diagnosis of PDAC and yielded higher tissue adequacy compared with conventional EUS-FNA.[48] [50] For malignant biliary tumors, a recent RCT in 2025 reported 83.3% sensitivity and 87.1% accuracy for CH-EUS-guided tissue acquisition, which is comparable to standard EUS-FNA.[51] For PCNs, CH-EUS-guided FNA of cyst fluid and mural nodules detected malignancy with 84.2% sensitivity and provided 76.9% diagnostic yield for conclusive cytology of high-grade dysplasia and carcinoma.[52]

A meta-analysis in 2021 demonstrated the superiority in pooled sensitivity, accuracy, and tissue adequacy of CH-EUS-guided FNA over standard EUS-FNA for diagnosing SPLs.[53] However, two recent meta-analyses in 2024 reported discordant results, attributed to the limited number of included studies, inclusion of crossover and retrospective designs, and heterogeneity in outcome measures in previous meta-analysis.[54] [55] In a meta-analysis by Engh et al, CH-EUS-guided TA showed no significant difference in diagnostic accuracy, sensitivity, tissue adequacy, or the mean number of needle passes required to achieve adequacy for diagnosing SPLs compared with conventional EUS-TA.[54] Subgroup analysis restricted to RCTs similarly revealed no difference in diagnostic performance between CH-EUS-guided and standard EUS-TA. Esposto et al. reported the non-superiority in diagnostic yield and first needle pass performance of CH-EUS-guided FNA/FNB for SPLs in a meta-analysis.[55] Similarly, Kuo et al reported comparable diagnostic accuracy and the number of passes required for diagnosis between standard EUS-FNB using the fanning technique and CH-EUS-guided FNB.[49]

In conclusion, the routine application of CH-EUS-guided TA is not mandatory and does not improve diagnostic performance and tissue adequacy for SPLs based on recent evidence. Notably, most available studies were generally conducted by experienced endosonographers. Therefore, larger prospective studies are warranted to explore the clinical utility of CH-EUS-guided TA and underscore its application in challenging scenarios, including chronic pancreatitis and PCNs, and to determine its potential benefit for less experienced endosonographers, where evidence remains limited.

Artificial Intelligence-Assisted EUS

Artificial intelligence (AI) is the ability of computer systems to simulate intelligent behavior, process information, and perform tasks associated with human intelligence with minimal human input. Machine learning (ML) is a subset of AI that employs algorithms to extract features from data to generate accurate predictions. Deep learning (DL) is a neural network that has evolved from ML, characterized by multilayered neural networks capable of automatically identifying data features, similar to the function of the human brain. A convolutional neural network (CNN) is the predominant DL architecture for 2D image and 3D video analysis. The CNN was trained by using large images and video datasets, enabling it to analyze new input images and videos effectively.

In EUS-TA, AI has been utilized to optimize diagnostic workflows by assessing the necessity of tissue acquisition and guiding procedures in real time. Qu et al evaluated the application of ML and deep neural networks (DNN) to identify SPLs at risk of yielding non-diagnostic EUS-FNA.[56] By integrating 1,447 radiomic features from contrast-enhanced CT images and EUS-FNA conclusive and inconclusive results, their DNN model achieved AUROCs of 0.821 and 0.745 in the development and validation cohorts, respectively. These findings demonstrated the promising predictive performance of the AI-assisted EUS model.

In 2023, Tang et al introduced the first AI model to guide EUS-TA.[57] The auxiliary diagnostic system, CH-EUS MASTER, was developed using a deep CNN and random forest algorithm trained on 4,638 EUS images and videos of pancreatic masses which the final diagnosis was confirmed by surgical pathology. The CH-EUS MASTER system comprised two models: Model 1 performed real-time pancreas capture and segmentation on EUS images, and Model 2 differentiated benign from malignant lesions. The CH-EUS MASTER worked with the main engine of the EUS, aiming to identify and track SPLs in real-time to guide the target area of cancerous lesions for FNA. Forty-six patients who underwent CH-EUS were randomly assigned to EUS-FNA with or without CH-EUS MASTER guidance. The CH-EUS MASTER significantly outperformed manual endoscopist labeling of CH-EUS images in terms of diagnostic accuracy, sensitivity, and specificity (92.3, 92.3, and 92.3% vs. 87.2, 88.5, and 84.6%, p < 0.05), and also achieved a higher first-pass diagnostic yield for SPLs compared with conventional EUS-FNA (80.0 vs. 33.3%, p = 0.029). Additionally, Ishikawa et al demonstrated the performance of an AI model in predicting the diagnosable material for histology using fresh specimens.[58] Using DL and contrastive learning on 173 specimens obtained by EUS-FNB from 96 patients, the AI analyzed stereomicroscopic and hematoxylin and eosin-stained images to identify core tissue. The sensitivity, specificity, and accuracy of the AI-based method were comparable with MOSE performed by EUS experts (88.97, 53.5, and 83.24% vs. 90.34, 53.5, and 84.39%). These results highlight the potential use of AI models as a novel method for evaluating diagnosable specimen material.

Based on current evidence, AI models have considerable promise for enhancing the efficacy of EUS-TA by predicting the necessity of tissue acquisition, differentiating malignant from benign lesions, guiding real-time lesion targeting by identifying regions with high malignancy suspicion, and assessing specimen adequacy. However, widespread adoption of AI-assisted EUS remains limited due to the high costs associated with acquiring quality datasets and challenges in dynamic image recognition during real-time procedures. Future research and development of practical and cost-effective AI models are essential to facilitate broader implementation and drive innovation in EUS-TA.

Rapid On-Site Evaluation and Digital Pathology

ROSE is a cytologic technique enabling real-time assessment of tissue sample quality during EUS-TA. The obtained specimens are immediately assessed using Diff-Quik or ultrafast Papanicolaou staining by an on-site cytopathologist. ROSE decreases the number of needle passes, enhances tissue adequacy, and informs management decisions based on cytologic findings. Meta-analyses have demonstrated that EUS-FNA with ROSE offered a 95% pooled sensitivity and significantly improved overall adequacy rates by ∼3.5%.[59] However, the advent of EUS-FNB raises questions about the benefit of ROSE. A recent large RCT reported the non-inferiority of EUS-FNB without ROSE compared with EUS-FNB with ROSE in diagnostic accuracy (97.5 vs. 96.4%, p = 0.396).[6] Furthermore, EUS-FNB without ROSE yielded a significantly higher tissue core rate, reduced procedure time, and comparable histologic sample quality.[6] These findings question the necessity of ROSE, especially in settings with a shortage of cytopathologists. Consequently, AI and digital pathology have been integrated into EUS-TA workflows to mitigate these limitations.

Emergence of AI-ROSE Models

In recent years, AI has been integrated with ROSE to identify malignant cells from specimens obtained by EUS-TA. Several studies proposed AI models aimed to substitute manual ROSE in centers without cytopathologists. The retrospective study in 2022 developed a deep CNN system for ROSE using 5,345 cytopathological slide images obtained from EUS-FNA with an AUROC of 0.96 in PDAC diagnosis.[60] Lin et al. developed the ROES-AI model to detect malignant cells from digitalized images of Diff-Quik-stained slides obtained by EUS-FNA.[61] The accuracy and sensitivity of the ROES-AI model were 83.4 and 79.1% in the internal validation cohort, and the external validation showed comparable results. Data augmentation using geometric transformation, color space transformation, and kernel filtering enhanced the accuracy of the ROSE-AI model in detecting PDAC up to 88.2% from specimens obtained by EUS-FNB.[62] Furthermore, AI models offer promising performance in differentiating PDAC from benign tissue during ROSE. The Mathematical Technology for Cytopathology algorithm was integrated with AI to distinguish PDAC from benign lesions obtained by EUS-FNA/FNB with a 74% accuracy.[63] The deep learning AI models using cytological images obtained from EUS-FNA/FNB through a hyperspectral imaging (HSI)-based CNN algorithm showed a 92% accuracy, 93% sensitivity, and 91.2% specificity in differentiating PDAC from benign pancreatic cells.[64] In 2025, Fang et al proposed a semi-supervised CNN system that learned from EUS-FNA cytology images to distinguish PDAC from non-PDAC tissue with 97% accuracy, 95% sensitivity, and 94% specificity, respectively, which correspondingly in the external validation dataset.[65]

All these AI-ROSE models are promising tools that could enhance the precision of PDAC diagnosis and guide the endoscopist to avoid unnecessary puncture, especially in centers without cytopathologists. Moreover, the initial cytological diagnosis could initially be established and guide the ancillary testing, including immunohistochemical or genomic profiling tests. However, larger external validation cohorts need to be addressed to ensure their performance for future clinical applications. In addition, accessibility to the ROSE-AI model worldwide is challenging due to its significant cost.

Telecytology and Digital Pathology

The implementation of telecytology and digital pathology platforms has expanded access to ROSE in centers without on-site cytopathologists. While ROSE performed by endosonographers has been explored as a potential solution to overcome this limitation, its impact on diagnostic accuracy and reduction in needle passes was inconsistent among studies. Consequently, telecytology and digital pathology represent promising adjuncts to improve diagnostic yield during EUS-TA. Telecytology typically utilizes four main platforms, including camera-based static imaging, whole slide imaging (WSI), video camera-based live streaming, and hybrid telecytology.

The efficacy and feasibility of ROSE telecytology for specimens obtained by fine-needle aspiration have been primarily demonstrated in non-EUS methods. However, limited data exist regarding the diagnostic performance of EUS-TA with ROSE telecytology. In camera-based static image telecytology, snapshot images captured by an on-site cytotechnologist may not fully represent the whole slide, limiting the accuracy of EUS-TA and making the results highly dependent on the experience of the cytotechnologist.[66] Although this limitation can potentially be addressed by the WSI technique, studies evaluating EUS-TA using WSI with ROSE telecytology remain limited.

For dynamic image telecytology, EUS-TA utilizing ROSE with video camera-based live streaming telecytology by a remote cytopathologist had a shorter evaluation time for specimen adequacy compared with conventional ROSE (0.2 vs. 0.75 hours).[67] Buxbaum et al reported 88% agreement between dynamic ROSE telecytology and final cytology and 92% agreement between conventional ROSE and final cytology, based on 25 EUS-FNA specimens.[68] Additionally, studies by Khurana et al demonstrated that EUS-FNA with ROSE via video camera-based live streaming yielded a comparable accuracy with conventional ROSE in diagnosing SPLs and reduced the non-diagnostic rate of specimens compared with those without ROSE.[69] [70] A study in 2023 further confirmed that video camera-based live streaming ROSE telecytology provided comparable diagnostic yield (96.4 vs. 94.5%, p = 0.428), albeit requiring significantly more needle passes during EUS-FNA/FNB for diagnosing SPLs.[71] A single-center retrospective study in 2024 by Cai et al implemented the Rapid On-Line Evaluation (ROLE) protocol using camera-based live streaming telecytology for diagnosing SPLs by EUS-FNA/FNB.[72] ROLE significantly improved diagnostic yield (97.3 vs. 85.5%, p = 0.023), accuracy (94.7 vs. 82.3%, p = 0.027), and sensitivity (95.7 vs. 81.1%, p = 0.011) compared with the non-ROSE group. Moreover, ROLE significantly reduced the number of needle passes (2 vs. 3, p < 0.001) and showed superior diagnostic accuracy with 22-gauge FNB needles compared with the non-ROSE group (100 vs. 93.1%, p = 0.025).

Based on current evidence, ROSE by camera-based live-streaming telecytology shows potential to improve diagnostic performance for SPLs. However, clinical data remain limited, and existing studies involve small sample sizes. Larger, well-designed prospective studies are needed to further evaluate the utility of ROSE telecytology, especially regarding the hybrid telecytology platform.

Fluorescence Confocal Microscopy

Fluorescence confocal laser microscopy (FCM) is an emerging technique for evaluating fresh tissue specimens with minimal preparation. The obtained specimens were expelled into the specific container and applied with a fluorescent dye. Using two different laser wavelengths, the system generates microscopic images via software that mimic the appearance of hematoxylin and eosin stains, closely resembling conventional histology. These images can be examined at magnifications up to 550-fold without tissue damage, distortion, or loss.[73]

Stigliano et al. demonstrated that FCM achieved 100% sensitivity, 66.7% specificity, and 97% accuracy, with strong agreement between FCM diagnosis and final histological diagnosis of SPLs specimens obtained by EUS-FNB (Cohen's k coefficient 0.95).[74] An international multicenter study in 2022 reported substantial concordance between FCM-based diagnosis and standard histological analysis from WSI of specimens obtained by EUS-FNB from 25 pancreatic lesions.[75] These findings support the potential integration of FCM with ROSE during EUS-TA to assess tissue adequacy and to advance the application of digital pathology for comprehensive histological evaluation.

Personalized Sampling Strategies

In recent years, advances in EUS-TA have significantly improved diagnostic performance and contributed to a deeper understanding of the biology and behavior of various malignancies. EUS-TA strategies are increasingly being tailored to specific lesion types, aligning with the principles of personalized treatment in the precision medicine era. [Table 3] summarizes the personalized tissue sampling strategies for various lesions based on recent evidence of advancements in EUS-TA.

Table 3 Summary of the personalized tissue sampling strategies by endoscopic ultrasound-guided tissue acquisition based on current evidences

Lesion types	Tissue sampling strategies
SPLs	Biopsy accessories and technique	The end-cutting needles (Franseen and fork-tip) yield a higher cellularity, accuracy, and adequacy compared with the reverse bevel needle[5] [6] [7]
		The diagnostic accuracy and adequacy among small 22- and 25-gauge needles were comparable, and the 25-gauge needle provides a higher diagnostic yield for head and uncinate lesions[10] [11]
		The fanning technique is recommended as it improves diagnostic accuracy
		Wet-suction and stylet-slow pull techniques were comparable for EUS-FNB in terms of tissue adequacy, integrity, and blood contamination compared with dry suction techniques
		Two needle passes are recommended with end-cutting FNB needles, three needle passes with reverse-bevel FNB needles, and at least four needle passes with FNA needles without ROSE
		MOSE is recommended to enhance diagnostic accuracy with the need for fewer needle passes
		ROSE is not necessary while performing EUS-FNB, especially with end-cutting needles, but is recommended for EUS-FNA to enhance diagnostic accuracy[6] [59]
	Molecular profiling	The 19-gauge Franseen needles provided a higher adequate tissue sampling for NGS compared with 22-gauge FNA and 22-gauge Franseen needles for PDAC. However, adverse events should be warranted[37] [38] [39]
		MOSE significantly enhances adequate tissue samples for NGS[34]
		For PDAC patients who undergo systemic therapy: consider testing potentially actionable somatic findings[a] using FDA-approved and/or validated NGS-based assay, including - Fusions: ALK, NRG1, NTRK, ROS1, FGFR2, and RET - Mutations: BRAF, BRCA1/2, KRAS, and PALB2 - Amplifications: HER2 - Immunotherapy biomarkers: MSI, dMMR, TMB
	Image enhancement	CH-EUS-guided TA was comparable with conventional EUS-FNA/FNB in accuracy, sensitivity, adequacy, mean number of needle passes, diagnostic yield, and first pass diagnosis.[54] [55] Therefore, CH-EUS-guided TA is not necessary in routine practice based on current evidence
		EUS elastography-guided TA is not mandatory as its diagnostic performance is comparable to conventional EUS-FNA/FNB, yielding comparable malignant cells from hard and soft areas, and requires more needle passes[43] [44] [45]
PCNs	Biopsy accessories and technique	EUS-TTNB outperformed EUS-FNA with cytology in sample adequacy and accuracy, particularly in mucinous cysts, and is suggested to obtain the definitive diagnosis of the pancreatic cyst, which can modify the management strategy. TTNB should be avoided in PCNs with pancreatic duct communication[15] [18] [19] [20] [21]
		The nCLE is suggested to differentiate mucinous and non-mucinous PCNs in centers with expertise, as it showed a significantly higher superiority index compared with cyst fluid biochemical markers and molecular analysis[22] [23]
		Intracystic brush cytology through a 19-gauge FNA needle provides significant diagnostic materials but does not improve diagnostic accuracy compared with conventional EUS-FNA with cyst fluid analysis based on a few small prospective studies[24] [25] [26]
	Molecular profiling	EUS-FNA with cystic fluid analysis combined with KRAS/GNAS mutation is helpful for differentiating mucinous versus non-mucinous PCNs. The use of cystic fluid cytology may help identify the presence of high-grade dysplasia and pancreatic cancer
		Positive KRAS mutation distinguishes MCN from other cystic lesions, and the sensitivity was increased when combined with CEA
		Positive GNAS mutation is highly specific for IPMN and helpful in differentiating from MCN
Lymphoma	Biopsy accessories and technique	- The 19- and 22-gauge FNA needles provide histological analysis with 80–87% sensitivity, 92–93% specificity, 83–89% accuracy, and could further enhance diagnostic accuracy when combined with flow cytometry, especially for DLBCL. The diagnostic accuracy of EUS-FNA with flow cytometry is lower for low-grade lymphomas, Hodgkin's disease, and T cell lymphomas. - The studies of EUS-FNB in the diagnosis of lymphoma are limited
	Flow cytometry	- T-cell lineage antigens: CD2, CD3, CD4, CD5, CD7, and CD8 - B-cell lineage antigens: CD10, CD19, CD20, IgM, immunoglobulin light chains (κ and λ), CD25, CD30, CD34, CD38, CD45, and CD56
	IHC	Core biopsy is essential for IHC study. - DLBCL and its variants: CD3, CD20, Ki-67, CD10, Bcl-6, MUM-1, Bcl-2, c-myc, EBV ISH - Small/low-grade B-cell lymphoma: CD3, CD20, Ki-67, Bcl-2, CD10 - T-cell/ NK-cell lymphoma: CD3, CD20, Ki-67, EBV ISH - Hodgkin lymphoma: CD3, CD20, Ki-67, CD30, CD15, PAX5, EBV ISH
	Molecular profiling	Example of proposed mandatory NGS panels for B-cell and T-cell lymphoma in diagnosing, predicting prognosis, and guiding treatment. - B-cell lymphoma: B2M, BCL10, BCL2, BCL6, BIRC3, BTG1, BTK, CARD11, CCND3, CD58, CD79B, CHD2, CREBBP, DTX1, EP300, EZH2, GNA13, KTM2D, MYC, MYD88, NOTCH1, NOTCH2, PIM1, PIM2, PLCG2, POU2AF1, PRDM1, SF3B1, SGK1, SOCS1, STAT3, STAT6, TCF3, TET2, TNFAIP3, HIST1H1E, TNFRSF14, IRF8, ID3, TP53, KLF2, UBE2A, IGHV - T cell lymphoma: CD28, DNMT3A, IDH2, JAK1, JAK3, PI3KCD, PLCG1, PRDM1, RHOA, SETD2, STAT3, STAT5B, TET1, TET2, TP53, VAV1
GISTs	Biopsy accessories and technique	EUS-FNB is preferred over conventional EUS-FNA for subepithelial lesions, and EUS-FNB is comparable with mucosal incision-assisted biopsy (MIAB) for diagnosing SEL ≥ 20 mm
	IHC	Core biopsy is essential for IHC study, including SDHB, CD117, and DOG1[a]
	Molecular profiling	- Testing for KIT and PDGFRA mutations should be performed if tyrosine kinase inhibitors are considered as part of the treatment response prediction[a] - For GISTs without KIT and PDGFRA mutations: alternative driver mutations (e.g., BRAF, NF1, NTRK, and FGFR fusions) may be assessed by NGS to identify potential targeted therapies[a]
	Image enhancement	Only a few data support the utility of CH-EUS and EUS elastography in the diagnosis of subepithelial lesions, and insufficient evidence to recommend based on the ESGE guideline
NET	IHC	Core biopsy is essential for IHC study, including 2 of 3 IHC markers of synaptophysin, chromogranin A, and INSM1[a]
		The Ki-67 proliferative index and/or mitotic rate are also required for grading[a]
	Molecular profiling	- In case of ambiguous histology that morphologically resembles NET, the broad NGS panel is recommended[a]
		Molecular profiling using NGS could guide the molecular-matched treatment. - Neuroendocrine tumor: MEN1, TSC2, PTEN, TP53, CDKN1B - Neuroendocrine carcinoma: KRAS, TP53, RB1, HRAS
Liver	Biopsy accessories and technique	EUS-LB has comparable tissue adequacy and adverse events compared with percutaneous and transjugular liver biopsy
		EUS-LB offers a comparable complete portal triad and length of the longest piece, but significantly yields a longer total specimen length compared with percutaneous liver biopsy
		A 19-gauge FNB needle demonstrated superior specimen adequacy and less tissue fragmentation compared with a 22-gauge FNB and a 19-gauge FNA needle. The 19-gauge Franseen-tip outperformed the fork-tip needle in terms of total specimen length, greater complete portal triad, and diagnostic yield
		Wet suction technique with heparin priming is preferable to improve tissue adequacy, decrease blood clotting within the needle, allowing more tissue recovery, and decreased tissue fragmentation
		The 1-pass, 3-actuation technique with a depth of at least 3 cm provided longer liver cores and more complete portal triad than a 1-pass, 1-actuation technique with an equivalent safety profile
		For the assessment of non-steatohepatitis, bilobar biopsy should be performed to decrease sampling error. Unilobar biopsy may be sufficient for other conditions
Mediastinal and intra-abdominal lymph nodes	Biopsy accessories and technique	For mediastinal lymphadenopathy: - EUS-FNA yielded 85% pooled sensitivity, 100% pooled specificity, and demonstrated the higher pooled sensitivity and specificity compared with EUS morphological evaluation alone (88.0 vs. 84.7% and 96.4 vs. 84.6%). These findings could benefit cancer diagnosis and staging. - A recent study in 2025 showed a comparable diagnostic yield of EUS-FNA and EUS-FNB. Moreover, EUS-FNA/FNB using 22-gauge needles offered a higher diagnostic yield compared with 19-gauge needles, and employing 2–4 needle passes showed a higher diagnostic yield compared with more than five needle passes
		For intra-abdominal lymphadenopathy: - The pooled sensitivity and specificity of EUS-FNA ranged from 87–94% and 98–99%, respectively. - EUS-FNA using 22- and 25-gauge needles and EUS-FNB using 20-, 22-, and 25-gauge needles yielded comparable sensitivity, accuracy, and number of needle passes. However, the specificity was significantly higher for EUS-FNB compared with EUS-FNA (100 vs. 92.62%)
		For metastatic mediastinal and intra-abdominal lymph nodes, EUS-FNA using 25-gauge needles and EUS-FNB using 22-gauge reverse-beveled needles demonstrated a comparable high sensitivity for diagnosing metastatic nodes (100 vs. 87%, p = 0.5)
	IHC and molecular profiling	The addition of IHC and molecular profiling is usually based on the provisional diagnosis that requires a combination of clinical data and images. However, EUS-FNB is preferred to procure core tissue samples with preserved architecture for further IHC studies and molecular analyses

Abbreviations: CEA, carcinoembryonic antigen; CH-EUS, contrast-harmonic EUS; DLBCL, diffuse large B cell lymphoma; dMMR, mismatch repair deficiency; ESGE, European Society of Gastrointestinal Endoscopy; EUS, endoscopic ultrasound; EUS-LB, EUS-guided liver biopsy; FDA, Food and Drug Administration; FNA, fine needle aspiration; FNB, fine needle biopsy; GISTs, gastrointestinal stromal tumors; IHC, immunohistochemistry; IPMN, intraductal papillary mucinous neoplasms; MCN, mucinous cystic neoplasm; MOSE, macroscopic on-site evaluation; MSI, microsatellite instability; nCLE, needle-based confocal laser endomicroscopy; NET, neuroendocrine tumors; NGS, next-generation sequencing; PCNs, pancreatic cystic neoplasms; PDAC, pancreatic ductal adenocarcinoma; ROSE, rapid on-site evaluation; SEL, subepithelial lesion; SPLs, solid pancreatic lesions; TA, tissue acquisition; TMB, tumor mutational burden; TTNB, through-the-needle biopsy.

^a Based on National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology 2025.

Conclusion

EUS-TA has evolved beyond conventional tissue sampling. Recent advances in technologies and innovations have been incorporated with EUS-TA to enhance diagnostic precision and support personalized care. AI integration, molecular profiling, and digital pathology are reshaping its clinical utility, reinforcing EUS-TA as a key diagnostic approach for precision medicine.

Conflict of Interest

None declared.

Authors' Contributions

Conceptualization: N.P.; literature search, data extraction: P.T., writing—original draft preparation: P.T.; writing—review and editing: N.P. All authors contributed to the manuscript revision, read, and approved of the submitted version.

• References

• 1 Pausawasdi N, Cheirsilpa K, Chalermwai W, Asokan I, Sriprayoon T, Charatcharoenwitthaya P. Endoscopic ultrasound-guided fine-needle biopsy using 22G Franseen needles without rapid on-site evaluation for diagnosis of intraabdominal masses. J Clin Med 2022; 11 (04) 1051
• 2 van Riet PA, Erler NS, Bruno MJ, Cahen DL. Comparison of fine-needle aspiration and fine-needle biopsy devices for endoscopic ultrasound-guided sampling of solid lesions: a systemic review and meta-analysis. Endoscopy 2021; 53 (04) 411-423
• 3 Khan MA, Grimm IS, Ali B. et al. A meta-analysis of endoscopic ultrasound-fine-needle aspiration compared to endoscopic ultrasound-fine-needle biopsy: diagnostic yield and the value of onsite cytopathological assessment. Endosc Int Open 2017; 5 (05) E363-E375
• 4 Oh HC, Kang H, Lee JY, Choi GJ, Choi JS. Diagnostic accuracy of 22/25-gauge core needle in endoscopic ultrasound-guided sampling: systematic review and meta-analysis. Korean J Intern Med (Korean Assoc Intern Med) 2016; 31 (06) 1073-1083
• 5 Young Bang J, Krall K, Jhala N. et al. Comparing needles and methods of endoscopic ultrasound-guided fine-needle biopsy to optimize specimen quality and diagnostic accuracy for patients with pancreatic masses in a randomized trial. Clin Gastroenterol Hepatol 2021; 19 (04) 825-835.e7
• 6 Crinò SF, Di Mitri R, Nguyen NQ. et al. Endoscopic ultrasound-guided fine-needle biopsy with or without rapid on-site evaluation for diagnosis of solid pancreatic lesions: a randomized controlled non-inferiority trial. Gastroenterology 2021; 161 (03) 899-909.e5
• 7 Gkolfakis P, Crinò SF, Tziatzios G. et al. Comparative diagnostic performance of end-cutting fine-needle biopsy needles for EUS tissue sampling of solid pancreatic masses: a network meta-analysis. Gastrointest Endosc 2022; 95 (06) 1067-1077.e15
• 8 Ishigaki K, Nakai Y, Sasahira N. et al; Acquire Study Group. A prospective multicenter study of endoscopic ultrasound-guided fine needle biopsy using a 22-gauge Franseen needle for pancreatic solid lesions. J Gastroenterol Hepatol 2021; 36 (10) 2754-2761
• 9 Facciorusso A, Wani S, Triantafyllou K. et al. Comparative accuracy of needle sizes and designs for EUS tissue sampling of solid pancreatic masses: a network meta-analysis. Gastrointest Endosc 2019; 90 (06) 893-903.e7
• 10 Siddique DA, Rahal MA, Trevino K, Wu HH, Al-Haddad MA. Endoscopic ultrasound-guided sampling of solid pancreatic lesions: a comparative analysis of 25 gauge versus 22 gauge core biopsy needles. Anticancer Res 2020; 40 (10) 5845-5851
• 11 Tomoda T, Kato H, Fujii Y. et al. Randomized trial comparing the 25G and 22G Franseen needles in endoscopic ultrasound-guided tissue acquisition from solid pancreatic masses for adequate histological assessment. Dig Endosc 2022; 34 (03) 596-603
• 12 van der Waaij LA, van Dullemen HM, Porte RJ. Cyst fluid analysis in the differential diagnosis of pancreatic cystic lesions: a pooled analysis. Gastrointest Endosc 2005; 62 (03) 383-389
• 13 Maire F, Couvelard A, Hammel P. et al. Intraductal papillary mucinous tumors of the pancreas: the preoperative value of cytologic and histopathologic diagnosis. Gastrointest Endosc 2003; 58 (05) 701-706
• 14 Salvia R, Burelli A, Nepi A. et al. Pancreatic cystic neoplasms: still high rates of preoperative misdiagnosis in the guidelines and endoscopic ultrasound era. Surgery 2023; 174 (06) 1410-1415
• 15 Gopakumar H, Puli SR. Value of endoscopic ultrasound-guided through-the-needle biopsy in pancreatic cystic lesions. a systematic review and meta-analysis. J Gastrointest Cancer 2024; 55 (01) 15-25
• 16 Vilas-Boas F, Ribeiro T, Macedo G. et al. Endoscopic ultrasound-guided through-the-needle biopsy: a narrative review of the technique and its emerging role in pancreatic cyst diagnosis. Diagnostics (Basel) 2024; 14 (15) 1587
• 17 Stigliano S, Covotta F, Di Matteo FM. A new micro-forceps for endoscopic ultrasound-guided through-the-needle biopsy in the diagnosis of pancreatic cystic lesions: single center experience. JGH Open 2021; 5 (09) 1004-1008
• 18 Facciorusso A, Del Prete V, Antonino M, Buccino VR, Wani S. Diagnostic yield of EUS-guided through-the-needle biopsy in pancreatic cysts: a meta-analysis. Gastrointest Endosc 2020; 92 (01) 1-8.e3
• 19 Tacelli M, Celsa C, Magro B. et al. Diagnostic performance of endoscopic ultrasound through-the-needle microforceps biopsy of pancreatic cystic lesions: systematic review with meta-analysis. Dig Endosc 2020; 32 (07) 1018-1030
• 20 Rift CV, Scheie D, Toxværd A. et al. Diagnostic accuracy of EUS-guided through-the-needle-biopsies and simultaneously obtained fine needle aspiration for cytology from pancreatic cysts: a systematic review and meta-analysis. Pathol Res Pract 2021; 220: 153368
• 21 Westerveld DR, Ponniah SA, Draganov PV, Yang D. Diagnostic yield of EUS-guided through-the-needle microforceps biopsy versus EUS-FNA of pancreatic cystic lesions: a systematic review and meta-analysis. Endosc Int Open 2020; 8 (05) E656-E667
• 22 Cheesman AR, Zhu H, Liao X. et al. Impact of EUS-guided microforceps biopsy sampling and needle-based confocal laser endomicroscopy on the diagnostic yield and clinical management of pancreatic cystic lesions. Gastrointest Endosc 2020; 91 (05) 1095-1104
• 23 Li SY, Wang ZJ, Pan CY, Wu C, Li ZS, Jin ZD. et al. Comparative performance of endoscopic ultrasound-based techniques in patients with pancreatic cystic lesions: a network meta-analysis. Off J Am Coll Gastroenterol 2023 ;118(2)
• 24 Al-Haddad M, Gill KR, Raimondo M. et al. Safety and efficacy of cytology brushings versus standard fine-needle aspiration in evaluating cystic pancreatic lesions: a controlled study. Endoscopy 2010; 42 (02) 127-132
• 25 Lozano MD, Subtil JC, Miravalles TL. et al. EchoBrush may be superior to standard EUS-guided FNA in the evaluation of cystic lesions of the pancreas: preliminary experience. Cancer Cytopathol 2011; 119 (03) 209-214
• 26 Lariño-Noia J, de la Iglesia D, Iglesias-García J. et al. Endoscopic ultrasound cytologic brushing vs endoscopic ultrasound - fine needle aspiration for cytological diagnosis of cystic pancreatic lesions. A multicenter, randomized open-label trial. Rev Esp Enferm Dig 2018; 110 (08) 478-484
• 27 Marques F, Baldaque Silva F, Wijngaart W, Arnelo U, Roxhed N. A loop-shaped minimally invasive brush for improved cytology sampling of pancreatic cysts during EUS-FNA. Med Devices Sens 2020; •••: 4
• 28 Marques F, Schliemann I, van der Wijngaart W, Arnelo U, Roxhed N, Baldaque-Silva F. New through-the-needle brush for pancreatic cyst assessment: a randomized controlled trial. iGIE 2023; 2 (04) 481-488
• 29 Ko SW, Jo IH, Yoon SB. Feasibility and clinical utility of endoscopic ultrasound-guided tissue acquisition for comprehensive genomic profiling in pancreatic cancer: a systematic review and meta-analysis. Pancreatology 2025; 25 (01) 89-97
• 30 Sato J, Ishiwatari H, Ishikawa K. et al. Benefits of macroscopic on-site evaluation in endoscopic ultrasound-guided tissue acquisition for comprehensive genomic profiling. Endosc Int Open 2025; 13: a25934172
• 31 Asokkumar R, Yung Ka C, Loh T. et al. Comparison of tissue and molecular yield between fine-needle biopsy (FNB) and fine-needle aspiration (FNA): a randomized study. Endosc Int Open 2019; 7 (08) E955-E963
• 32 Kandel P, Nassar A, Gomez V. et al. Comparison of endoscopic ultrasound-guided fine-needle biopsy versus fine-needle aspiration for genomic profiling and DNA yield in pancreatic cancer: a randomized crossover trial. Endoscopy 2021; 53 (04) 376-382
• 33 Nagatani T, Wani Y, Takatani M. et al. Adequacy evaluation of 22-gauge needle endoscopic ultrasound-guided tissue acquisition samples and glass slides preparation for successful comprehensive genomic profiling testing: a single institute experience. DEN Open 2025; 6 (01) e70104
• 34 Kaneko J, Ishiwatari H, Sasaki K. et al. Macroscopic visible core length can predict the histological sample quantity in endoscopic ultrasound-guided tissue acquisition: multicenter prospective study. Dig Endosc 2022; 34 (03) 622-631
• 35 Carrara S, Soldà G, Di Leo M. et al. Side-by-side comparison of next-generation sequencing, cytology, and histology in diagnosing locally advanced pancreatic adenocarcinoma. Gastrointest Endosc 2021; 93 (03) 597-604.e5
• 36 Bunduc S, Varzaru B, Iacob RA. et al. Endoscopic ultrasound-guided fine-needle aspiration pancreatic adenocarcinoma samples yield adequate DNA for next-generation sequencing: a cohort analysis. World J Gastroenterol 2023; 29 (18) 2864-2874
• 37 Hisada Y, Hijioka S, Ikeda G. et al. Proportion of unresectable pancreatic cancer specimens obtained by endoscopic ultrasound-guided tissue acquisition meeting the OncoGuide™ NCC Oncopanel System analysis suitability criteria: a single-arm, phase II clinical trial. J Gastroenterol 2022; 57 (12) 990-998
• 38 Okuno N, Hara K, Mizuno N. et al. Clinical utility of endoscopic ultrasound-guided tissue acquisition for comprehensive genomic profiling of pancreatic cancer. Clin Endosc 2023; 56 (02) 221-228
• 39 Ikeda G, Hijioka S, Nagashio Y. et al. Fine-needle biopsy with 19G needle is effective in combination with endoscopic ultrasound-guided tissue acquisition for genomic profiling of unresectable pancreatic cancer. Dig Endosc 2023; 35 (01) 124-133
• 40 Takeshita K, Hijioka S, Nagashio Y. et al. Study protocol for a prospective self-controlled trial on success in meeting comprehensive genomic profiling analysis criteria for specimens obtained by endoscopic ultrasound-guided tissue acquisition using a 19G needle from primary and metastatic lesions in pancreatic cancer with metastatic lesions: the PRIMATE study. Diseases 2024; 12 (08) 182
• 41 Yanaidani T, Hara K, Okuno N. et al. Clinical utility of endoscopic ultrasound-guided tissue acquisition for comprehensive genomic profiling of patients with biliary tract cancer, especially with intrahepatic cholangiocarcinoma. Clin Endosc 2024; 57 (03) 384-392
• 42 Facciorusso A, Martina M, Buccino RV, Nacchiero MC, Muscatiello N. Diagnostic accuracy of fine-needle aspiration of solid pancreatic lesions guided by endoscopic ultrasound elastography. Ann Gastroenterol 2018; 31 (04) 513-518
• 43 Ohno E, Kawashima H, Ishikawa T. et al. The role of EUS elastography-guided fine needle biopsy in the histological diagnosis of solid pancreatic lesions: a prospective exploratory study. Sci Rep 2022; 12 (01) 16603
• 44 Gheorghiu M, Sparchez Z, Rusu I. et al. Direct comparison of elastography endoscopic ultrasound fine-needle aspiration and B-mode endoscopic ultrasound fine-needle aspiration in diagnosing solid pancreatic lesions. Int J Environ Res Public Health 2022; 19 (03) 1302
• 45 Lee TS, Lee SH, Kim J. et al. Increased needle passes for comparable diagnostic yield in endoscopic ultrasound-guided tissue acquisition for pancreatic stiff lesions measured by elastography. Pancreatology 2024; 24 (07) 1192-1198
• 46 Kamata K, Takenaka M, Omoto S. et al. Impact of avascular areas, as measured by contrast-enhanced harmonic EUS, on the accuracy of FNA for pancreatic adenocarcinoma. Gastrointest Endosc 2018; 87 (01) 158-163
• 47 Itonaga M, Ashida R, Kitano M. Endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) with image enhancement. Diagnostics (Basel) 2020; 10 (11) 888
• 48 Itonaga M, Kitano M, Kojima F. et al. The usefulness of EUS-FNA with contrast-enhanced harmonic imaging of solid pancreatic lesions: a prospective study. J Gastroenterol Hepatol 2020; 35 (12) 2273-2280
• 49 Kuo YT, Chu YL, Wong WF. et al. Randomized trial of contrast-enhanced harmonic guidance versus fanning technique for EUS-guided fine-needle biopsy sampling of solid pancreatic lesions. Gastrointest Endosc 2023; 97 (04) 732-740
• 50 Lai J-H, Lin C-C, Lin H-H, Chen M-J. Is contrast-enhanced endoscopic ultrasound-guided fine needle biopsy better than conventional fine needle biopsy? A retrospective study in a medical center. Surg Endosc 2022; 36 (08) 6138-6143
• 51 Seicean A, Samarghitan A, Bolboacă SD. et al. Contrast-enhanced harmonic versus standard endoscopic ultrasound-guided fine-needle aspiration in solid pancreatic lesions: a single-center prospective randomized trial. Endoscopy 2020; 52 (12) 1084-1090
• 52 Olar MP, Bolboacă SD, Pojoga C. et al. Clinical utility of the contrast-enhanced endoscopic ultrasound guided fine needle aspiration in the diagnosis of pancreatic cyst. Diagnostics (Basel) 2022; 12 (09) 2209
• 53 Facciorusso A, Mohan BP, Crinò SF. et al. Contrast-enhanced harmonic endoscopic ultrasound-guided fine-needle aspiration versus standard fine-needle aspiration in pancreatic masses: a meta-analysis. Expert Rev Gastroenterol Hepatol 2021; 15 (07) 821-828
• 54 Engh MA, Teutsch B, Schulze Wenning A. et al. Contrast-enhanced endoscopic ultrasound likely does not improve diagnostic adequacy during endoscopic ultrasound guided tissue acquisition: a systematic review and meta-analysis. Pancreatology 2024; 24 (04) 649-660
• 55 Esposto G, Massimiani G, Galasso L. et al. Endoscopic contrast-enhanced ultrasound and fine-needle aspiration or biopsy for the diagnosis of pancreatic solid lesions: a systematic review and meta-analysis. Cancers (Basel) 2024; 16 (09) 1658
• 56 Qu W, Yang J, Li J. et al. Avoid non-diagnostic EUS-FNA: a DNN model as a possible gatekeeper to distinguish pancreatic lesions prone to inconclusive biopsy. Br J Radiol 2023; 96 (1151) 20221112
• 57 Tang A, Tian L, Gao K. et al. Contrast-enhanced harmonic endoscopic ultrasound (CH-EUS) MASTER: a novel deep learning-based system in pancreatic mass diagnosis. Cancer Med 2023; 12 (07) 7962-7973
• 58 Ishikawa T, Hayakawa M, Suzuki H. et al. Development of a novel evaluation method for endoscopic ultrasound-guided fine-needle biopsy in pancreatic diseases using artificial intelligence. Diagnostics (Basel) 2022; 12 (02) 434
• 59 Matynia AP, Schmidt RL, Barraza G, Layfield LJ, Siddiqui AA, Adler DG. Impact of rapid on-site evaluation on the adequacy of endoscopic-ultrasound guided fine-needle aspiration of solid pancreatic lesions: a systematic review and meta-analysis. J Gastroenterol Hepatol 2014; 29 (04) 697-705
• 60 Zhang S, Zhou Y, Tang D. et al. A deep learning-based segmentation system for rapid onsite cytologic pathology evaluation of pancreatic masses: a retrospective, multicenter, diagnostic study. EBioMedicine 2022; 80: 104022
• 61 Lin R, Sheng LP, Han CQ. et al. Application of artificial intelligence to digital-rapid on-site cytopathology evaluation during endoscopic ultrasound-guided fine needle aspiration: a proof-of-concept study. J Gastroenterol Hepatol 2023; 38 (06) 883-887
• 62 Fujii Y, Uchida D, Sato R. et al. Effectiveness of data-augmentation on deep learning in evaluating rapid on-site cytopathology at endoscopic ultrasound-guided fine needle aspiration. Sci Rep 2024; 14 (01) 22441
• 63 Yamada R, Nakane K, Kadoya N. et al. Development of “mathematical technology for cytopathology,” an image analysis algorithm for pancreatic cancer. Diagnostics (Basel) 2022; 12 (05) 1149
• 64 Qin X, Zhang M, Zhou C. et al. A deep learning model using hyperspectral image for EUS-FNA cytology diagnosis in pancreatic ductal adenocarcinoma. Cancer Med 2023; 12 (16) 17005-17017
• 65 Fang D, Huang Y, Li S. et al. A semi-supervised convolutional neural network for diagnosis of pancreatic ductal adenocarcinoma based on EUS-FNA cytological images. BMC Cancer 2025; 25 (01) 495
• 66 Chung MJ, Park SW, Kim SH. et al. Clinical and technical guideline for endoscopic ultrasound (EUS)-guided tissue acquisition of pancreatic solid tumor: Korean Society of Gastrointestinal Endoscopy (KSGE). Gut Liver 2021; 15 (03) 354-374
• 67 Marotti JD, Johncox V, Ng D, Gonzalez JL, Padmanabhan V. Implementation of telecytology for immediate assessment of endoscopic ultrasound-guided fine-needle aspirations compared to conventional on-site evaluation: analysis of 240 consecutive cases. Acta Cytol 2012; 56 (05) 548-553
• 68 Buxbaum JL, Eloubeidi MA, Lane CJ. et al. Dynamic telecytology compares favorably to rapid onsite evaluation of endoscopic ultrasound fine needle aspirates. Dig Dis Sci 2012; 57 (12) 3092-3097
• 69 Khurana KK, Graber B, Wang D, Roy A. Telecytopathology for on-site adequacy evaluation decreases the nondiagnostic rate in endoscopic ultrasound-guided fine-needle aspiration of pancreatic lesions. Telemed J E Health 2014; 20 (09) 822-827
• 70 Khurana KK, Rong R, Wang D, Roy A. Dynamic telecytopathology for on-site preliminary diagnosis of endoscopic ultrasound-guided fine needle aspiration of pancreatic masses. J Telemed Telecare 2012; 18 (05) 253-259
• 71 Kouanda A, Mclean R, Faggen A. et al. Telecytology versus in-room cytopathologist for EUS-guided FNA or fine-needle biopsy sampling of solid pancreatic lesions. Gastrointest Endosc 2023; 97 (03) 466-471
• 72 Cai Y, Rao X, Zhang J. et al. A Rapid On-Line Evaluation (ROLE) protocol in the diagnostic performance improvement in endoscopic ultrasound-guided tissue acquisition for solid pancreatic lesions. Diagnostics (Basel) 2024; 14 (06) 597
• 73 Krishnamurthy S, Cortes A, Lopez M. et al. Ex vivo confocal fluorescence microscopy for rapid evaluation of tissues in surgical pathology practice. Arch Pathol Lab Med 2018; 142 (03) 396-401
• 74 Stigliano S, Crescenzi A, Taffon C. et al. Role of fluorescence confocal microscopy for rapid evaluation of EUS fine-needle biopsy sampling in pancreatic solid lesions. Gastrointest Endosc 2021; 94 (03) 562-568.e1
• 75 Amendoeira I, Arcidiacono PG, Barizzi J. et al. New digital confocal laser microscopy may boost real-time evaluation of endoscopic ultrasound-guided fine-needle biopsy (EUS-FNB) from solid pancreatic lesions: data from an international multicenter study. EBioMedicine 2022; 86: 104377

Publication History

Article published online:
31 October 2025