Endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) has become the standard method for the diagnosis of solid pancreatic lesions and is necessary to determine tailored therapeutic plans.¹ Its reported accuracy ranges from 65% to 96%, and the diagnostic accuracy is particularly high for pancreatic cancer.^2,3 In addition, EUS-FNA of solid pancreatic lesions is a safe procedure, with a low overall complication rate of approximately 2.5%, including an extremely low risk of tumor seeding.^4–6 Due to its high diagnostic accuracy and safety, EUS-FNA has surpassed percutaneous sampling techniques such as transabdominal ultrasonography or computed tomography-guided biopsy. However, EUS-FNA has several limitations. It often provides only a cytological sample for diagnosis, but in some pancreatic neoplasms, such as stromal tumors and lymphomas, histological specimens are necessary to confirm the diagnosis.⁷ Additionally, EUS-FNA has a relatively low negative predictive value of 65% to 72% for the diagnosis of pancreatic malignancy.⁸ Therefore, even if the result of EUS-FNA is negative, it is not certain that the lesion is benign. False-negative results have the potential to delay patient care and exert negative impacts on patient outcomes.⁹ A variety of factors can affect the diagnostic yield of EUS-FNA, including the nature, size, and location of the lesion, the presence of necrosis or inflammation, the size and type of the needle, the number of needle passes, the use of a stylet, the use and type of suction, the FNA technique, the presence of rapid on-site cytologic evaluation (ROSE), the experience of the endosonographer and cytologist, and the method of cytopathology preparation. To overcome these limitations, several attempts have been made to improve the diagnostic performance of EUS-FNA. The main objective of this review is to focus on how to improve the diagnostic yield of EUS-FNA from a technical point of view and to discuss relevant clinical applications.

ROSE is a technique in which cytology samples from EUS-FNA are rapidly stained and screened for diagnostic material in the endoscopy suite during the procedure.¹⁰ In the presence of ROSE, the diagnostic yield of EUS-FNA in solid pancreatic lesions is improved by up to 10%–15%, and the diagnostic accuracy exceeds 90% in most studies.^11,12 Direct communication and immediate feedback with a cytopathologist on the adequacy of the sample obtained during the procedure may confirm optimal tissue acquisition, thereby increasing the diagnostic yield of EUS-FNA and reducing the need for repeated or alternative procedures for a definitive diagnosis,^11,13 as well as potentially reducing procedure-related complications and time by minimizing the total number of needle passes.¹¹

However, ROSE is not available at many centers because of its high cost and limited medical resources. A recent systematic review and meta-analysis demonstrated that the implementation of ROSE did not improve the diagnostic yield for malignancies or the proportion of patients with adequate specimens.¹⁴ ROSE did not have a beneficial effect on cellular yield; thus, a similar cellular yield may lead to comparable diagnostic efficacy between cases with and without ROSE.^14,15 Therefore, Kong et al¹⁴ stated that the routine application of ROSE may not affect outcomes in clinical practice at tertiary care centers. The Korean Society of Gastrointestinal Endoscopy (KSGE) guideline also suggests that the routine application of ROSE cannot guarantee superior diagnostic accuracy and performance in terms of sensitivity and specificity.¹⁵ Nevertheless, for a given sample adequacy and number of needle passages, the application of ROSE is expected to achieve higher per-case diagnostic accuracy than non-application.¹⁵

The fanning technique, which was first introduced by Bang et al¹⁶ in 2013, has been defined as a procedure in which multiple different areas within a mass are targeted while the needle undergoes to-and-fro movement using the up/down knob of the endoscope during each needle passage.¹⁵ The accurate placement of the needle into the lesion alone does not always guarantee adequate tissue acquisition. In particular, larger tumors are more likely to have necrosis in the center, and even the periphery may exhibit reactive desmoplasia and inflammatory debris induced by the tumor.¹⁷ Due to the extensive central necrosis with displacement of inflammatory cells and fibrotic stroma encompassing the tumor cells, the cytohistologic yield of EUS-FNA may be insufficient and lead to false-negative outcomes despite good needle placement.^17,18 Therefore, the application of the fanning technique could theoretically increase the likelihood of achieving a true diagnosis, thereby reducing the risk of inconclusive results.¹⁵ Indeed, a randomized controlled trial (RCT) conducted by Bang et al¹⁶ demonstrated that the fanning technique of EUS-FNA was superior to the standard approach because fewer passes were required to establish an accurate diagnosis, without significant differences in technical failure, the complication rate, or diagnostic accuracy. Furthermore, the fanning technique does not require additional financial costs. The KSGE guideline also suggests that the fanning technique for EUS-guided tissue acquisition (EUS-TA) offers technically acceptable feasibility and superior diagnostic outcomes, requiring fewer passes to establish the definite diagnosis, than the standard technique.¹⁵

Recently, another maneuver called the “torque technique,” which is applied by twisting the body of the echoendoscope to the right (clockwise) or left (counterclockwise) without using the left/right control knob, was introduced.¹⁹ Similar to the fanning technique, which targets multiple vertical points, this technique aims at multiple horizontal target points within the same needle passage without additional needle puncture to reduce false-negative results or additional adverse events.¹⁹ A prospective randomized study revealed that the torque technique during EUS-guided fine-needle biopsy (EUS-FNB) for solid pancreatic lesions showed superior diagnostic performance, with optimal histologic core procurement and acceptable technical feasibility, compared with the standard technique.¹⁹

To date, studies have reported variable diagnostic performance of sampling techniques during EUS-TA of solid pancreatic lesions based on needle size (gauge [G]) (19G vs. 22G vs. 25G) and type (FNA vs. FNB) (Fig. 1).⁹ Although, the optimal needle size for EUS-FNA depends on the nature and location of the pancreatic mass lesions, most previous studies have shown similar diagnostic yields in discriminating malignant from benign lesions and have failed to show superiority among these FNA needles.^20–23 However, two meta-analyses comparing 22G and 25G FNA needles for diagnosing pancreatic malignancies have suggested that the 25G FNA needle is more sensitive than the 22G.^24,25 The better diagnostic yield of the smaller 25G FNA needles in these studies may be attributed to the more cellular and less bloody aspirate with less tissue injury, which can lead to better cytologic interpretation.¹ The 25G needle is also more flexible, making it technically easier to use for lesions in the head portion and uncinate process of the pancreas.¹

Figure 1. Needle-tip designs for fine-needle aspiration (FNA) and fine-needle biopsy (FNB). (A) Expect FNA needle (Boston Scientific, Natick, MA, USA); (B) Franseen FNB needle (Acquire; Boston Scientific); (C) EchoTip Ultra FNA needle (Cook Medical, Limerick, Ireland); and (D) EchoTip Procore FNB needle (Cook Medical).

Meanwhile, it is often difficult to use EUS-FNA cytology alone to diagnose some conditions, such as autoimmune pancreatitis, neuroendocrine tumors (NETs), and well-differentiated adenocarcinoma, where histology with preserved tissue architecture is essential for diagnosis.²⁶ To overcome the limitations of FNA needles associated with cytology, FNB needles uniquely designed to obtain a tissue core have been introduced. The first FNB needle was a 19G Quick-core (also called the Tru-Cut core biopsy needle; Cook Medical, Limerick, Ireland). The larger 19G FNB needle has the potential to acquire larger tissue samples, potentially improving diagnostic accuracy. However, it is difficult to access through the transduodenal route and has a higher rate of technical failure and complications in pancreatic head lesions due to its lack of flexibility and scope angulation.⁷ For these reasons, despite the acquisition of larger samples of core tissue, the 19G Tru-cut core biopsy needle did not show improved diagnostic accuracy over FNA needles.^27,28 Subsequently, a new FNB needle called ProCore (Echo Tip ProCore; Cook Medical), which has a reverse-bevel design for cutting the tissue core, has been developed. Ideally, EUS-TA of pancreatic tumors should routinely include a histological evaluation by biopsy in addition to cytology.¹⁵ FNB needles have the theoretical advantage of improving sample procurement with preservation of the intact tissue architecture, which allows immunohistochemistry or special stains required for critical differential diagnoses, as well as obtaining results in fewer passes and thus potentially improving the efficiency and costs, compared to those of standard FNA needles.^9,29 However, one RCT reported a similar diagnostic yield between FNB and FNA needles in discriminating malignancies.³⁰ A meta-analysis failed to show superiority of the ProCore in terms of diagnostic accuracy over standard FNA needles, but the ProCore needle established the diagnosis with fewer passes.^31,32 A recent systematic review with a network meta-analysis also demonstrated that there were no significant differences in diagnostic accuracy, sample adequacy, and histologic core procurement between EUS-FNA and FNB needles, accounting for different needle gauges.⁹ These results add credence to the recently published European Society of Gastrointestinal Endoscopy (ESGE) guideline, which equally recommends both 22G and 25G needles for routine sampling of solid pancreatic masses, regardless of the needle type (both FNA and FNB needles).³³ The KSGE guideline also recommends FNA and FNB needles equally in routine EUS-TA for solid pancreatic lesions, but suggests that FNB is better when the primary aim of sampling is to obtain a histologic core tissue specimen (e.g., autoimmune pancreatitis or NETs).¹⁵ Two new core biopsy needles with a novel tip, opposing bevel, and sheath design were recently introduced. One has a fork-tip design with two leading sharp tips on the opposite side of the lumen (SharkCore; Medtronic, Minneapolis, MN, USA), and the other is a Franseen tip design with three symmetric cutting edges (Acquire; Boston Scientific, Natick, MA, USA).⁹ Several studies have demonstrated significant or comparable performance in terms of diagnostic and histologic yields for both new FNB needles.³⁴ A recent RCT comparing the diagnostic yield of EUS-FNB for pancreatic masses between four needle types for a single pass (reverse-bevel, Menghini-tip, Franseen, or fork-tip) demonstrated that the new FNB needles (Franseen and fork-tip) yielded higher cellularity and achieved high diagnostic accuracy, exceeding 90% on a single pass, which was significantly higher than was possible with the older-generation needles.³¹ Larger prospective studies comparing these new FNB needles with previous devices to improve the accuracy and histologic core procurement are warranted.³⁵

The use of a stylet in the needle during EUS-TA prevents blockage of the needle or contamination by the intestinal mucosa when passing through the gastrointestinal wall and allows more adequate acquisition of the target tissue.³⁶ The stylet is also able to maintain the stiffness of the needle, allowing puncture of a fibrotic lesion.³⁷ However, using a stylet can prolong the procedure and the risk of unintentional needle injuries, especially when multiple passes are required.^36,38

Previous studies have shown that stylet use did not lead to improvements in diagnostic yield and specimen adequacy for solid pancreatic lesions.^36,39 Thus, the KSGE and ESGE technical guidelines stated that the use of a stylet does not appear to guarantee any advantages with regard to the adequacy of the specimen and diagnostic yield, and there is insufficient evidence to recommend for or against using a stylet during EUS-TA.^15,40 However, because all commercially available FNA and FNB needles are already preloaded with a stylet, based on the results of previous studies, clinicians can consider a first pass with a stylet and subsequent passes without a stylet.⁴¹ EUS-TA performed in this way may not improve the risk of bloodiness or contamination of the specimen, but may shorten the procedure time. Moreover, in special cases such as fibrotic or hard lesions, the use of a stylet should be considered because the stylet may stiffen the needle, making it easier to puncture the lesion.³⁷ Conversely, when puncturing from the bulb with the echoendoscope in the long position, using a stylet may actually make EUS-TA more difficult, as the stylet may be challenging to advance or remove due to the tortuosity and tip angulation of the echoendoscope.³⁶ If it becomes possible to select a fine-needle aspiration/biopsy (FNA/B) needle with or without a stylet, it might result in a cost reduction for the EUS-TA system.³⁶

Theoretically, negative pressure from the application of suction is expected to increase cellularity and improve diagnostic yield. Standard suction is generally performed using a 10-mL syringe and may be applied constantly, intermittently, or manually.¹ Several previous trials comparing the presence or absence of suction for EUS-FNA of pancreatic lesions have shown that application of suction resulted in not only superior diagnostic outcomes in terms of higher cellularity and sensitivity, but also a higher blood contamination rate resulting in decreased sample quality, which can lead to reduced diagnostic yield.⁴² When EUS-FNA was used to detect malignant lymphadenopathy, the conventional suction method worsened specimen bleeding and did not improve the diagnostic accuracy compared to that of FNA without suction.⁴³ The need for suction during EUS-TA depends on the characteristics of the lesion and the quality of the initial aspirate. If the initial aspirate is bloody, suction aspiration should not be performed in the subsequent passes.^18,44 Similarly, if the initial aspirate is scant or insufficient, application of suction is preferred in the next puncture.⁴⁴ The ESGE technical guideline recommends applying continuous suction for EUS-FNA of solid masses to increase sensitivity, but no suction for lymph nodes.⁴⁰ The KSGE guideline recommends the routine application of suction in cases of poor cellularity, such as fibrotic lesions in patients with chronic pancreatitis, whereas it does not recommend suction in non-fibrotic lesions that may contain necrosis and blood to minimize contamination of the cellular sample.¹⁵

Furthermore, a new technique, called the “slow-pull-back technique,” which uses only capillary aspiration with minimal negative pressure achieved by slow and continuous withdrawal of the stylet during to-and-fro needle movement, was recently introduced for EUS-FNA or FNB of solid pancreatic lesions. A prospective study comparing the diagnostic yield of EUS-FNA samples obtained with the slow-pull-back technique or the standard suction technique in patients with suspected malignant pancreatic lesions demonstrated that the diagnostic yield, the cellularity of smears, and the rate of acquiring sufficient histological material were similar between two groups, but the pathological diagnosis was faster and more cost-effective using the slow-pull-back technique due to less bleeding of samples and a lower number of slides.⁴⁵ A recent RCT comparing the diagnostic yield among patients with suspected pancreatic malignancy who underwent EUS-FNB using a 22G ProCore needle with the slow-pull-back, standard suction, or non-suction technique after stylet removal found that the standard suction technique showed higher blood contamination, but did not increase the rate of core-tissue acquisition compared with other methods.⁴⁶ Instead, the slow-pull-back technique provided greater cellularity with less blood contamination over other methods, while guaranteeing comparable results in terms of the rate of adequate core-tissue acquisition.⁴⁶ The KSGE guideline also suggested that the slow-pull-back technique is more effective in terms of adequate tissue acquisition and requires fewer needle passes for solid pancreatic lesions.¹⁵

In the “wet-suction technique,” another novel technique, the needle is flushed with 5 mL of saline solution to replace the air column within the needle lumen with saline before aspiration, and then a 10-mL suction syringe is attached in the “locked” position to the needle. In a prospective randomized trial, the wet-suction technique using 22G FNA needles resulted in significantly better cellularity and specimen adequacy than the standard suction technique.⁴⁷ Another technique, high-negative pressure suction using a 60-mL syringe with 50-mL of high negative pressure, was also found to be superior to the standard suction technique during EUS-FNA procedures in terms of tissue acquisition.⁴⁸ Further multicenter prospective studies are warranted to compare the aforementioned two novel methods with the slow-pull-back technique, along with validation according to different pancreatic lesions, needle type, and needle size.

An optimal number of needle passes would be particularly useful at institutions where ROSE is not routinely available during EUS-TA.⁴⁹ Making more needle passes than necessary not only lengthens the procedure, but may also increase the risk of procedure-related complications such as bleeding and pancreatitis. Conversely, a suboptimal number of needle passes may increase the rate of false-negative results and can lead to unnecessary expenses due to the need for repeated and subsequent procedures arising from nondiagnostic specimens, especially in the presence of chronic pancreatitis.^15,50

Recent prospective trials have reported that at least 5 to 7 passes with a standard FNA needle are required for solid pancreatic lesions in the absence of ROSE,^12,49 and fewer needle passes are required to obtain an accurate diagnosis with an FNB needle than with an FNA needle.⁵¹ Moreover, making more than 4 needle passes significantly improved the diagnostic yield even for small tumors less than 2 cm in diameter.¹² In Korea, where ROSE is usually not possible, the KSGE guidelines suggest that making 4 needle passes during EUS-TA would be suitable for diagnosing pancreatic tumors, although more needle passes may be needed if the tumor is smaller than 2 cm.¹⁵ Fewer needle passes have been reported to be required for the EUS-FNB procedure.¹⁵

Movement of the needle within the lesion may also be important. In general, quick back-and-forth jabbing or “staccato” movements (~5 to 10) are recommended instead of slow movements.¹ Short “ice pick” movements versus long strokes within the lesion may break up tissue to improve the cytologic yield.¹

If the initial EUS-FNA/B is inconclusive and nondiagnostic, surgery is the best option when the clinical and imaging suspicion for malignancy is very high, the lesion appears to be resectable, and the patient is a good surgical candidate.⁵⁰ When there is persistent clinical suspicion of malignancy, but the patient’s health status is marginal and resectability is intermediate, so that a histologic diagnosis is necessary to establish a treatment plan, several alternative diagnostic tools can be chosen, such as endoscopic retrograde cholangiopancreatography (ERCP) with bile duct brushing and/or endobiliary forcep biopsy, transabdominal ultrasonography or computed tomography-guided percutaneous biopsy, or laparoscopic exploration. However, each of the above-mentioned methods has disadvantages of postprocedural complications, such as post-ERCP pancreatitis and bleeding, risk of tumor seeding, and invasiveness. Meanwhile, several studies have revealed that repeated EUS-TA can yield a definitive diagnosis in 59% to 93% of cases of suspected pancreatic malignancies that are inconclusive on initial EUS-TA.^52–55 Repeated EUS-TA provided a conclusive diagnosis in the majority of cases; therefore, it is the best course of action and should be strongly considered ahead of other modalities.^55,56 However, repeating the EUS-FNA/B procedure more than three times is reported not to enhance the diagnostic yield.⁵⁷

Despite the high sensitivity of EUS for the detection of solid pancreatic masses, it is difficult to distinguish pancreatic cancer from other diseases on EUS imaging alone.⁵⁸ Contrast-enhanced harmonic endoscopic ultrasonography (CEH-EUS) involves intravenous injection of an ultrasound contrast agent that facilitates the visualization of tissue perfusion patterns and assessment of the microvascularization within pancreatic masses.^1,59,60 Pancreatic adenocarcinoma has a distinct hypovascular appearance compared to that of other lesions such as NETs or chronic and autoimmune pancreatitis; thus, CEH-EUS can depict most pancreatic cancers as a solid lesion with hypoenhancement.¹ Therefore, CEH-EUS has been used for the differential diagnosis of focal pancreatic masses, which distinguishes pancreatic adenocarcinoma (distinct hypovascularity with hypo-enhancement) from NETs (hypervascularity with strong arterial hyper-enhancement), and pseudotumoral chronic pancreatitis and autoimmune pancreatitis (isovascularity or hypervascularity with isoenhancement or hyperenhancement).^50,59 In a recent meta-analysis, hypoenhanced lesions on CEH-EUS had a pooled sensitivity of 94% and specificity of 89% for the differential diagnosis of pancreatic adenocarcinoma in patients with pancreatic mass lesions.⁶¹

CEH-EUS plays a complementary role and assists in identifying target lesions for EUS-TA,⁵⁸ and there have been several studies on whether FNA guidance by CEH-EUS would increase diagnostic accuracy relative to conventional EUS-FNA.^62,63 One prospective study reported that the combination of CEH-EUS with FNA achieved greater accuracy with fewer needle passes.⁶² Another prospective randomized trial showed that CEH-EUS with FNA enabled the collection of sufficient biopsy samples with fewer needle passes than required for conventional EUS-FNA, resulting in improved safety and diagnostic efficacy compared with the conventional procedure.⁶³ In contrast, two recent RCTs showed that the use of CEH-EUS did not improve the diagnostic performance of EUS-FNA/B sampling or reduce the number of needle passes required for obtaining sufficient tissue samples.^64,65 However, those studies mainly involved FNA sampling, and the usefulness of CEH-EUS with FNB in the diagnosis of pancreatic mass lesions according to the type, size, and sampling technique of FNB needles was not confirmed. Further large studies evaluating the value of CEH-EUS during FNA/B are needed, but the combination of CEH-EUS and FNA/B should be considered in some challenging cases, such as those associated with chronic pancreatitis or severe necrosis, to avoid mistargeting.⁵⁵ Moreover, CEH-EUS can be easily performed without complications during FNA/B procedures.⁵⁵

Needle-based confocal laser endomicroscopy (nCLE) is a novel imaging technology that can provide real-time magnified endoscopic images at the cellular level at 1,000-fold magnification.⁶⁶ An nCLE miniprobe can be inserted through a 19G EUS-FNA needle; thus, EUS-guided nCLE (EUSnCLE), which allows real-time optical biopsies, during EUS-TA has been introduced.^66,67 Real-time optical biopsy can further improve the diagnostic yield by reducing sampling error and providing real-time feedback during the procedure when ROSE is not available.²⁶ A few studies have demonstrated the feasibility and safety of EUSnCLE in solid pancreatic lesions.^10,66 A recent pilot study presented distinctive EUSnCLE findings of malignant solid pancreatic lesions, including dark clumping with or without dilated vessels (> 40 μm), and characteristic findings of benign lesions, which have white fibrous bands and normal acinar cells, and asserted that the EUSnCLE criteria used are simple and likely applicable in daily practice with good inter-observer agreement.⁶⁶ The authors concluded that real-time histology with EUSnCLE is safe and potentially reduces the procedure time and unnecessary resources required for the EUS-TA procedure, including ROSE.⁶⁶ Another study aiming to define EUSnCLE criteria for the characterization of pancreatic masses with histopathological correlations described pancreatic adenocarcinoma as being characterized by dark cell aggregates and irregular vessels with leakages of fluorescein, chronic pancreatitis as characterized by residual regular glandular pancreatic structures, and NETs as characterized by black cell aggregates surrounded by vessels and fibrotic areas, which correlated well with the pathologic findings of the corresponding lesions.¹⁰ The authors argued that EUSnCLE is feasible and safe for diagnosing solid pancreatic masses, and given the low negative predictive value of EUS-FNA, EUSnCLE might help rule out malignancy after previously inconclusive EUS-FNA findings.¹⁰ However, the current price of nCLE probes is too high for it to be a cost-effective method, and larger systematic studies are needed to establish the role of EUSnCLE in the diagnosis of pancreatic masses.⁶⁶

Although many factors influence the diagnostic accuracy of EUS-TA in solid pancreatic lesions, the most important factor appears to be the experience and ability of the endosonographer.¹ The diagnostic yield of EUS-TA for solid pancreatic lesions may vary considerably among endosonographers.¹ Guidelines for EUS-FNA state that the diagnostic accuracy of EUS-FNA for pancreatic mass lesions increases with experience after a minimum of 75 to 100 cases for interpretative competence and 25 to 50 cases for FNA.^40,68–70 The American Society for Gastrointestinal Endoscopy guideline recommends at least 150 supervised EUS procedures, which should include a minimum of 75 pancreaticobiliary system cases and a minimum of 50 EUS-FNA cases.^4,71 A recent report demonstrated that the minimum number of EUS procedures for achieving competency was approximately 225 cases.⁷² Based on these study results, the KSGE guideline recommends that the average trainee must perform at least 225 EUS examinations, with a total of 50 EUS-TA procedures, to achieve competency in EUS-FNA or FNB. However, a recent meta-analysis emphasized that there was no clear number of EUS procedures;⁷³ thus, it is reasonable to assume that endoscopists performing a large number of EUS-TA cases are more likely to achieve successful results because the procedure is highly operator-dependent with significant inter-operator variability.¹

EUS-FNA is a key procedure for diagnosing solid pancreatic lesions, but multiple variables affect its diagnostic yield. Awareness of these potential variables, including advances in EUS-guided fine-needle sampling, and the ability to apply, modify, and combine sampling techniques in a given situation will undoubtedly improve the diagnostic and therapeutic outcomes in patients with solid pancreatic lesions.

None.

Hong Joo Kim has been an editor (Deputy Editor) of the International Journal of Gastrointestinal Intervention (IJGII) since 2017; however, Hong Joo Kim has not been involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.