Methods improving endoscopic retrograde cholangiopancreatography brush cytology in malignant biliary strictures

Abstract

Accurate diagnosis of biliary strictures is crucial as malignant causes are associated with high morbidity and mortality, and unnecessary surgery for benign disease carries substantial risk. Endoscopic retrograde cholangiopancreatography with brush cytology remains the most widely used tissue sampling technique, but its sensitivity in distinguishing benign from malignant strictures is limited. This review summarized current and emerging strategies aimed at improving the diagnostic yield of endoscopic retrograde cholangiopancreatography brush cytology. Technical refinements such as rapid on-site evaluation, sheath-rinse collection, and increased brushing passes have shown incremental improvements. Combining brushing with bile aspiration or intraductal biopsy further enhances sensitivity. In parallel, molecular approaches including fluorescence in situ hybridization, next-generation sequencing, and microRNA profiling offer promising advances in early and accurate detection. A multimodal approach that integrates optimized cytological sampling with molecular diagnostics appears most effective in improving accuracy. Standardization and validation of these combined methods in prospective studies are needed to establish their role in clinical practice and to improve outcomes for patients with indeterminate biliary strictures.

Key Words: Biliary stricture; Pancreatic cancer; Cholangiocarcinoma; Next-generation sequence; Fluorescence in situ hybridization; Endoscopic retrograde cholangiopancreatography; Brushing; MicroRNA

Core Tip: Endoscopic retrograde cholangiopancreatography brush cytology remains a first-line sampling technique for indeterminate biliary strictures but suffers from limited sensitivity despite high specificity. Technical refinements including increased brushing passes, rapid on-site cytopathology evaluation, sheath-rinse techniques, and combined bile aspiration or fluoroscopic-guided biopsy improve cellular yield and diagnostic accuracy. Integration of advanced molecular diagnostics such as fluorescence in situ hybridization, next-generation sequencing, and microRNA profiling further enhances detection rates and enables identification of actionable genomic alterations. A multimodal diagnostic strategy combining optimized sampling techniques with molecular analysis represents the most effective current approach for improving early diagnosis and clinical decision making in malignant biliary strictures.

INTRODUCTION

Biliary strictures (BSs) represent a heterogeneous clinical condition arising from a wide spectrum of benign inflammatory processes and malignant neoplasms. In a substantial proportion of cases, conventional imaging and endoscopic assessment fail to conclusively determine etiology, creating a diagnostic gray zone with significant therapeutic implications. Because management strategies differ dramatically between benign and malignant disease, establishing tissue diagnosis is fundamental to avoid both delayed oncologic treatment and unnecessary high-risk surgery. Despite the high prevalence of malignancy among BSs, accurate differentiation between benign and malignant causes remains critical as clinical presentation and imaging characteristics frequently overlap[1,2]. Approximately 15%-20% of patients who undergo surgical resection for presumed malignant indeterminate strictures are ultimately diagnosed with benign disease with postoperative mortality rates reported up to 10%, largely reflecting procedural complexity[3]. In addition, patients with malignant BSs often require surgical resection or palliative biliary drainage and have a poor prognosis if it is not diagnosed at the early stage. Therefore, establishing histological diagnosis is the key in determining further management of these patients since consequences are harmful when a malignant cause of a BS is missed or diagnosed at a later stage precluding surgery.

The histological diagnosis of BS without an associated mass in which endoscopic ultrasound-guided sampling can be performed remains a major challenge in clinical practice. When cross-sectional imaging reveals a discrete mass lesion, endoscopic ultrasound-guided fine-needle aspiration is often the most straightforward method for obtaining a tissue diagnosis. In the absence of a clearly identifiable mass, endoscopic retrograde cholangiopancreatography (ERCP) is typically employed as the primary diagnostic and therapeutic procedure for the evaluation and management of BSs.

Currently ERCP is the standard of care for indeterminate BSs either by brush cytology and biopsy or by direct cholangioscopy. However, although ERCP combined with brush cytology provides valuable diagnostic information, its sensitivity and accuracy can be limited particularly in detecting malignancies[4]. At present, concern has been expressed about improving the accuracy of ERCP brushing, and several methods have emerged with the intent to addressing this issue. Attempts of advanced molecular diagnostics of brushing samples offer the potential to significantly enhance diagnostic yield. Several adjunctive techniques have been explored to enhance the diagnostic performance of brush cytology specimens, including fluorescence in situ hybridization (FISH), rapid on-site cytopathologic assessment, evaluation of tumor-associated microRNA (miRNA) expression profiles, and various DNA-based molecular analyses[5]. Because no unified diagnostic framework currently exists for BSs, this review outlined contemporary management strategies and examined practical approaches to enhance the diagnostic accuracy of ERCP brush cytology, aiming to support earlier and more reliable clinical decision-making.

Etiology of BSs

A BS is an abnormal narrowing of the bile duct that impairs bile flow from the liver to the duodenum, potentially leading to cholestasis, jaundice, and biliary cirrhosis. This reduction in the luminal diameter of the bile duct can be due to intrinsic or extrinsic factors such as fibrosis, inflammation, or malignancy[1,2]. The clinical spectrum of BSs ranges from inflammatory narrowing to tumor-related obstruction. Lesions may originate within the ductal epithelium or result from compression by adjacent malignancies, most commonly cholangiocarcinoma (CCA) and pancreatic carcinoma, although other hepatobiliary and metastatic tumors may also be implicated[3,4]. Most BSs are malignant with pancreatic adenocarcinoma remaining the most common cause of malignant stricture of the distal common bile duct. CCA is the second most common cause of malignant BSs but is more likely in patients with mid or proximal stricture. Malignant transformation may also develop in the setting of chronic benign inflammatory disorders, particularly primary sclerosing cholangitis[6,7].

Benign BSs may result from iatrogenic injury, chronic pancreatitis, choledocholithiasis, primary sclerosing cholangitis, eosinophilic cholangitis, IgG4-related sclerosing cholangitis, and infectious, vascular, and other causes such as trauma and chemoradiation[7-10]. Differentiating benign etiologies of BSs can be challenging, particularly because certain non-malignant conditions carry an inherent oncologic risk. For instance, patients with primary sclerosing cholangitis have an estimated risk of developing CCA that is several hundred times higher than that of the general population[9].

Indeterminate BSs are defined as ductal narrowings for which cross-sectional imaging such as CT or magnetic resonance cholangiopancreatography (MRCP) does not reveal a distinct mass, and in which conventional diagnostic techniques, including ERCP with fluoroscopic evaluation, brush cytology, and/or intraductal forceps biopsy, are unable to conclusively determine whether the lesion is benign or malignant[11]. These strictures are implicated in a broad spectrum of pathological entities, encompassing both benign and malignant conditions.

Initial evaluation and diagnostic methods

A comprehensive first-line evaluation combines liver function testing with detailed cross-sectional imaging to define the level and nature of obstruction. Patients with BSs often present with symptoms related to bile flow obstruction and hyperbilirubinemia. As a result jaundice, pruritus, malaise, weight loss, abdominal pain, anorexia, nausea and vomiting are the most usual symptoms, and they typically progress surreptitiously as the bilirubin level rises. Biochemical evaluation should encompass a full hepatic panel, including aminotransferases (aspartate aminotransferase, alanine aminotransferase), cholestatic markers such as alkaline phosphatase and gamma-glutamyl transferase, and total bilirubin levels. In obstructive patterns elevations in alkaline phosphatase typically predominate over transaminases, and progressively higher bilirubin concentrations have been correlated with an increased probability of underlying malignancy[12,13]. Thomasset et al[14] demonstrated that both isolated hyperbilirubinemia and concurrent liver enzyme abnormalities significantly increased the probability that a BS was malignant.

In patients with indeterminate BSs, assessment for benign etiologies should include measurement of serum IgG4, as elevated levels may suggest IgG4-related sclerosing cholangitis when interpreted alongside clinical and imaging findings[15]. Tumor markers may help to differentiate BSs although no single marker is fully reliable. The most frequently used tumor markers for pancreaticobiliary malignancies are carbohydrate antigen 19-9 (CA19-9) and carcinoembryonic antigen. Serum CA19-9 levels may increase in malignancy and in benign conditions associated with cholestasis, including biliary inflammation, infectious processes, and even metabolic disorders such as diabetes. Carcinoembryonic antigen can be associated with a variety of pathologies apart from CCA and pancreatic carcinoma. It has lower sensitivity and specificity for CCA compared with CA19-9 (30%-68% and 75%-95%, respectively)[16]. It is typically used in combination with CA19-9 to improve diagnostic accuracy. Although CA125 is classically associated with ovarian malignancy, elevated levels are also observed in approximately 65% of patients with pancreaticobiliary cancers, especially in the presence of peritoneal dissemination[17].

Noninvasive evaluation of BSs relies on abdominal ultrasonography and advanced cross-sectional modalities, particularly contrast-enhanced CT and magnetic resonance imaging with MRCP. Contrast-enhanced MRCP provides high-resolution visualization of the biliary tree and accurately defines the level of obstruction with a reported sensitivity and specificity approaching 90%. It may also differentiate between benign and malignant strictures with a sensitivity of 38%-90% and a specificity of 70%-85%. Given its noninvasive nature and comparable diagnostic performance in delineating strictures, MRCP is generally favored as the initial imaging modality over ERCP[4]. Contrast-enhanced CT plays a central role in detecting mass-forming lesions, assessing local tumor extension, and evaluating vascular involvement. These features are critical not only for establishing the diagnosis but also for determining resectability and guiding operative planning[4].

Noninvasive imaging modalities are crucial to determine the presence and location of the stricture, but invasive methods are needed to obtain a conclusive histological diagnosis for an adequate differentiation of the etiology of BS. This field of invasive imaging is rapidly expanding and includes the more traditional ERCP, endoscopic ultrasound-guided fine needle aspiration, cholangioscopy, intraductal ultrasound, and confocal laser endomicroscopy that requires specialized equipment and trained personnel.

When therapeutic biliary decompression is indicated, ERCP serves as the principal invasive modality, enabling tissue acquisition through brushing, scraping, or intraductal biopsy. Endoscopic ultrasound-guided fine needle aspiration is generally favored when cross-sectional imaging identifies a mass associated with a BS or when ERCP fails to establish a definitive diagnosis. In contrast, cholangioscopy allows direct endoluminal inspection of the stricture and enables site-specific biopsy under visual guidance[3]. Nonetheless, ERCP continues to serve as a valuable diagnostic tool for tissue acquisition to determine the nature of these strictures. Choledochoscopy presents several inherent challenges, including its dependence on procedural skill, significant financial burden, and equipment-related technical limitations. Its clinical utility should also be carefully compared with standard approaches when evaluating economic justification. Furthermore, in patients with distal BSs, cholangioscopy demonstrates reduced diagnostic performance, primarily because maintaining adequate scope stability in this region can be technically demanding.

Updated recommendations from the American Society for Gastrointestinal Endoscopy emphasize three core strategies to optimize the endoscopic assessment of malignancy in BSs of uncertain origin[18]. It advocates for the integration of fluoroscopic-guided biopsy sampling and brush cytology during ERCP, particularly for hilar strictures. Cholangioscopy is recommended for patients with non-distal BSs or those who have undergone nondiagnostic ERCP without cholangioscopy. The guideline also supports combining endoscopic ultrasound with ERCP when initial ERCP sampling is nondiagnostic, particularly in patients with distal strictures or radiologic evidence of lymph node enlargement or metastatic disease.

ERCP SAMPLES AND COMBINING TECHNIQUES TO IMPROVE ACCURACY

ERCP enables precise delineation of a BS, including its anatomical location and longitudinal extent while also facilitating the acquisition of tissue specimens from the affected segment for cytological analysis. ERCP is associated with several adverse events like pancreatitis, bleeding, cholangitis, cholecystitis, and perforation. A meta-analysis documented adverse events in < 5% of patients and procedure-related mortality < 0.2%[19,20]. Although interpretation of cholangiographic results during ERCP is the first step, various types of samples are taken for cytological examination, namely bile aspirate, bile duct lavage, forceps for intraductal biopsies under fluoroscopic guidance and most often brush biopsy for cytology evaluation from the site of stenosis. Brush cytology is the most frequently utilized sampling technique during ERCP, primarily due to its technical simplicity and its relatively low risk of procedure-related complications (Figure 1).

Figure 1 Schematic illustration of endoscopic retrograde cholangiopancreatography brush cytology technique for sampling biliary strictures. Endoscopic retrograde cholangiopancreatography showing a cytology brush advanced through the endoscope into the common bile duct for tissue sampling at the site of the biliary stricture. The pancreatic duct is shown for anatomical reference. This schematic demonstrates the brushing technique used to obtain cellular material for cytological and molecular analysis. Illustration created by the authors. CBD: Common bile duct; PD: Pancreatic duct.

Osnes et al[21] first described brush cytology performed during ERCP in 1975. It is the most common sampling technique used in daily practice. While ERCP brush cytology offers high specificity in diagnosing malignant BSs, its sensitivity can be variable. The sensitivity of brush cytology varies significantly across studies ranging from 23%-86%[21,22] (Table 1).

Table 1 Diagnostic methods of malignant biliary strictures: Sensitivity and specificity.

Diagnostic method	Study design	Number of patients	Sensitivity (%)	Specificity (%)	Ref.
Brushing cytology	Review	1556	41.6 ± 3.2		Burnett et al[24]
	Prospective	109	40	100	Boldorini et al[47]
	Retrospective	74	67	98	Dudley et al[62]
	Retrospective	1017	65	78	Khan et al[48]
	Review		23-56	95	Singh et al[6]
	Prospective	102	36	85	Zoundjiekpon et al[50]
	Prospective	101	26-41	100	Kushnir et al[40]
FISH	Prospective	109	70	100	Boldorini et al[47]
	Retrospective	74	55	94	Dudley et al[62]
	Retrospective	1017	72	67	Khan et al[48]
	Prospective	102	20.0-42.9	82	Zoundjiekpon et al[50]
	Prospective	101	39	100	Kushnir et al[40]
	Review + meta-analysis	2516	57.6	87.8	Aggarwal et al[54]
NGS (MP)	Retrospective	74	74	98	Dudley et al[62]
	Prospective	101	50	97	Kushnir et al[40]
Brushing cytology + FISH	Retrospective	74	76	92	Dudley et al[62]
	Retrospective	1017	85	42	Khan et al[48]
	Prospective	101	50-60	95	Kushnir et al[40]
	Prospective	102	50	74	Zoundjiekpon et al[50]
Brushing cytology + NGS (MP)	Retrospective	74	85	96	Dudley et al[62]
	Prospective	101	56	95	Kushnir et al[40]
Brushing cytology + NGS (MP) + FISH	Retrospective	74	85	90	Dudley et al[62]
	Prospective	101	66-69	95	Kushnir et al[40]
Brushing cytology + fluoroscopic biopsy	Review		60-70	95	Singh et al[6]
Brushing cytology + bile fluid	Prospective	111	84	85	Roth et al[37]
Brushing cytology + miRNA	Prospective	57	54-85	95	Le et al[71]
Brushing cytology + miR-196a	Prospective	57	92.9	100	Le et al[71]
Brushing cytology + 2 positive of miR16, miR196a, miR221	Prospective	57	88.5	88.9	Le et al[71]

FISH: Fluoroscope in situ hybridization; MP: Molecular profiling; NGS: Next generation sequencing; miRNA: MicroRNA.

In a prospective cohort of 100 patients with BSs, Kurzawinski et al[23] evaluated the diagnostic performance of combining exfoliative bile cytology with brush cytology. The combined approach demonstrated greater sensitivity compared with bile cytology alone (69% vs 33%). Furthermore, a meta-analysis conducted by Burnett et al[24], including over 1500 patients, reported an overall sensitivity of 42% for brush cytology with a negative predictive value of 58% (Table 1).

Diagnostic accuracy of biliary brushings can be influenced by various lesion-related, operator-dependent, and/or patient-related factors. Proper handling and processing of brushing specimens are vital. The implementation of rapid on-site evaluation during ERCP brushing has demonstrated increased sensitivity of 74.6% and specificity of 97.5%[25]. Rapid on-site evaluation involves real-time assessment of specimen adequacy by a cytopathologist during the procedure. Evidence suggests that its use not only improves diagnostic yield but also reduces the number of sampling passes needed to reach a definitive diagnosis[26].

The addition of sheath-rinse specimens has been investigated as a strategy to increase cytological yield during ERCP. So et al[27] assessed its effect on sample cellularity and diagnostic performance by comparing brush-wash and sheath-rinse techniques. The use of sheath-rinse specimens was associated with an increase in sensitivity (from 59.2% to 69.9%) and overall diagnostic accuracy (from 62.5% to 72.3%), indicating a meaningful improvement in cytological evaluation.

The location of the BS can also influence diagnostic performance. However, studies have shown mixed results regarding its significance. The nature of the stricture, such as being compressive vs non-compressive, can affect cytology results but requires further investigation[28,29]. Furthermore, the proficiency of the endoscopist and cytopathologist notably affects diagnostic accuracy. Experienced practitioners are more likely to obtain and correctly interpret adequate samples, reducing false-negative and false-positive rates[30]. Age has also been identified as an independent factor influencing diagnostic outcomes. Younger patients may have a higher probability of obtaining a correct diagnosis with brush cytology compared with older individuals[31,32]. Some research indicates that older patients have a slightly higher likelihood of positive cytology results with an odds ratio of 1.02 per year increase in age. Conversely, other studies suggest that younger patients (under 69 years) exhibit higher sensitivity in brush cytology with a reported sensitivity of 84.2%[28]. Also, elevated total bilirubin levels are associated with increased sensitivity of brush cytology up to a threshold of 20 mg/dL. Beyond this level, severe jaundice may negatively impact diagnostic accuracy[32].

The method of sample collection, including the number of brushings and the pressure applied, plays a crucial role in obtaining adequate cellular material. Techniques like stricture dilatation and repeat brushing have been proposed to enhance sample quality without causing a significant increase in procedure-related adverse events[33]. In a study by Abbasi et al[34], application of negative pressure during biliary brush cytology was proposed as a novel sampling strategy, yielding sensitivity of 74.2% and perfect specificity in their cohort.

A randomized controlled trial demonstrated that increasing the number of brushing passes from 10 to 30 significantly improved diagnostic sensitivity, rising from 38% to 57%. In contrast, performing brush cytology after prior stricture dilation did not lead to a meaningful increase in cancer detection rates[5,35]. Therefore, according to the abovementioned study, when it comes to brush passes the more the merrier although there are no guidelines that dictate the optimal minimum number of passes.

The combined use of biliary brushing and bile aspiration during ERCP has been shown to enhance diagnostic performance. Cytological analysis of aspirated bile demonstrated superior cellular adequacy (92.8% vs 35.7%) and higher sensitivity (66.6% vs 17.6%) compared with brush cytology alone[5]. In a study of 76 patients with BSs, Sugimoto et al[36] reported a sensitivity of 32% for bile aspiration as a standalone method; however, when bile aspiration was performed after brushing, sensitivity increased to 70%. Similarly, Roth et al[37] found that analyzing both brush specimens and biliary fluid collected before and after brushing further improved sensitivity, reaching 84% (Table 1). Nevertheless, the diagnostic yield of brush cytology is limited by the often inadequate amount of sampled tissue due to submucosal tumor growth and the presence of extrinsic malignancy with biliary involvement or cellular atypia due to underlying inflammation and long-term biliary stenting[38].

Improving diagnostic accuracy requires refinement of ERCP sampling techniques alongside the incorporation of advanced molecular assays. Because brushing material and bile fluid originate directly from the affected ductal environment, they provide a biologically relevant substrate for identifying tumor-associated molecular alterations[39].

FISH serves as a cytogenetic adjunct to conventional brushing by identifying tumor-associated chromosomal alterations through fluorescent DNA probes. Aberrations such as polysomy are reported in the majority of malignant biliary tumors with detection rates approaching 80%. Nevertheless, the technique depends on adequate preservation of neoplastic cells and may yield limited results in paucicellular specimens[40] (Table 1).

Furthermore, the addition of advanced molecular diagnostics offers the potential to significantly enhance diagnostic accuracy. Molecular diagnostics, such as next-generation sequencing (NGS) and liquid biopsy, seem to be effective for identifying malignancy at an early stage by detecting genetic alterations and mutation profiling of DNA in brush or biopsy samples[18,41].

Molecular characterization of brush-derived samples has expanded beyond cell-based techniques. In addition to cytologic evaluation and FISH analysis, which require adequate epithelial cell capture, sequencing technologies can interrogate DNA fragments present in the supernatant of processed brush specimens. This approach allows detection of tumor-associated mutations even when cellular yield is suboptimal, offering a complementary strategy to traditional cytopathologic assessment[35]. According to a recent prospective trial by Singhi et al[41], NGS on brush cytology samples revealed sensitivity rates of 83%.

In the study by Kushnir et al[40], conventional cytology alone yielded a sensitivity of 26%. The addition of FISH increased sensitivity to 44%, while incorporation of molecular profiling further improved it to 56%. When all three modalities were applied together, the overall sensitivity reached 66% (Table 1).

Techniques such as liquid-based cytology can improve sample preservation and diagnostic yield. Bile-based molecular testing represents an expanding diagnostic frontier in the evaluation of BSs. Tumor-related molecules detectable in bile ranging from proteomic and metabolomic markers to miRNA signatures may provide clinically meaningful information that complements conventional cytologic and histologic assessment[42,43]. Yoshida et al[44], reported that bile contains a higher concentration of RNA species packaged within exosomes compared with circulating serum. This observation supports the concept that bile-derived exosomal miRNAs may represent a valuable source of disease-specific biomarkers for biliary tract cancers.

NOVEL TECHNIQUES AND MOLECULAR DIAGNOSTICS APPROACHES FOR INDETERMINATE BILIARY TRACT STRICTURES

FISH

FISH serves as an adjunct cytogenetic technique that enhances conventional cytology by identifying chromosomal copy number alterations within sampled biliary epithelial cells. Through probe-based detection of specific chromosomal loci, it enables recognition of aneuploidy patterns frequently associated with malignancy[45,46]. Typical aberrations include polysomy of chromosomes 3, 7, and 17 along with deletion of the 9p21 region. Although highly specific, its diagnostic performance depends on adequate cellular material and may be limited in paucicellular specimens. Such chromosomal alterations are not limited to biliary tumors but are observed across a broad spectrum of epithelial malignancies[47]. FISH analysis has been shown to identify pancreatobiliary malignancies with 93% sensitivity and 100% specificity[47].

Data assessing FISH in biliary brushings demonstrate variable sensitivity but consistently high specificity. In conventional CCA sensitivity typically ranges from approximately one-third to two-thirds of cases, whereas specificity approaches 100%. In primary sclerosing cholangitis-associated disease, diagnostic sensitivity appears more heterogeneous, likely reflecting inflammatory background changes, and specificity remains comparatively robust[35]. Khan et al[48] reported that adding FISH to conventional brush cytology improved sensitivity from 65% to 85% and increased the negative predictive value from 49% to 74% compared with cytology alone (Table 1). Evaluating 498 cases with parallel cytology and FISH analysis, Fritcher et al[49], reported superior diagnostic yield when cytogenetic testing was integrated into the assessment strategy. The reported sensitivity of FISH (43%) exceeded that of conventional cytology alone (20%). Importantly, the presence of polysomy was strongly associated with malignant disease, conferring an approximately 78-fold increase in the likelihood of cancer compared with negative FISH findings. Moreover, polysomy carried a substantially greater malignant potential than isolated trisomy 7 (98% vs 48%).

In a recent prospective study, combining brushing cytology and FISH increased sensitivity by approximately 15% in the primary diagnosis of BSs compared with cytology alone[50]. In individuals with primary sclerosing cholangitis, the detection of polysomy by FISH raises concern for underlying biliary neoplasia, particularly when accompanied by elevated serum CA19-9 levels. In this population in which chronic inflammatory changes often limit the diagnostic accuracy of conventional brush cytology, the addition of FISH has been shown to improve both sensitivity and specificity[30,51-53].

By identifying defined chromosomal copy number alterations, FISH can enhance diagnostic sensitivity for CCA and reduce the need for repeat endoscopic procedures. Accumulating evidence supports its adjunctive value in evaluating malignant strictures. A recent systematic review and meta-analysis reported pooled sensitivity and specificity estimates of 57.6% and 87.8%, respectively[54].

However, the disadvantage of FISH is that the technique is expensive and requires significant technical expertise. Moreno Luna et al[55], who assessed the clinical utility of cytology, digital imaging analysis, and FISH for the identification of malignant pancreas to BSs, stated that the cost of FISH is about three to four times the cost of cytology. Nevertheless, FISH has the potential to contribute to a reduction in overall healthcare expenditures by minimizing the need for repeated diagnostic procedures. The diagnostic clarity provided by FISH can also inform more nuanced and high-stakes clinical decision making, particularly in determining the appropriateness of resource-intensive therapeutic interventions, such as surgical resection. Nevertheless, a comprehensive and rigorous assessment of the cost effectiveness of these advanced cytologic techniques remains essential.

NGS

NGS combines a high analytical sensitivity and allows rapid simultaneous analysis of genetic material. NGS panels target multiple oncogenic mutations commonly detected in CCA and pancreatic carcinoma, the most common causes of malignant BSs.

Approximately 30%-50% of patients with CCA and pancreatic carcinoma harbor identifiable genomic alterations that may be amenable to targeted therapeutic interventions[56]. Evidence suggests that individuals treated with therapies matched to molecular profile of their tumor demonstrate superior response rates and improved survival outcomes compared with those receiving conventional, non-targeted treatment approaches[57]. Current European Society for Medical Oncology (ESMO) recommendations emphasize the importance of comprehensive molecular characterization in biliary tract cancers. Tissue acquisition strategies should be individualized according to the tumor site and technical feasibility. Available approaches include percutaneous core needle sampling, endoscopic brush-based cytology during ERCP, and endoscopic ultrasound-guided fine needle aspiration or biopsy[58,59].

Regarding the tumor location, Jusakul et al[60] pointed out that different anatomical locations of biliary cancer do not determine the molecular subtype as tumors in different locations can show molecular similarities. Genomic analyses suggest that biliary tract cancers cannot be fully characterized by anatomical origin. Tumors from separate locations may cluster within similar molecular subgroups, and lesions from a single anatomical site may diverge significantly at the genomic level. Importantly, patient outcomes seem more closely linked to underlying mutational landscapes than to tumor topography, underscoring the therapeutic relevance of molecular profiling.

Targeted NGS performed on bile duct brushings has shown high diagnostic accuracy in distinguishing benign from malignant strictures while simultaneously revealing clinically actionable genomic alterations[61]. NGS can be performed on cell-free DNA present in brush specimen preservative fluid obtained during ERCP. Unlike cytology and FISH, which depend on intact epithelial cells captured on the brush, NGS can be applied to cell-free DNA present in the residual preservative fluid following cytocentrifugation of brushing specimens[62]. Briefly, the NGS workflow can be RNA-based or DNA-based and thus starts with DNA and/or RNA extraction from a tumor sample followed by library generation involving sample preparation and target enrichment and then sequencing and data analysis. Sequencing platforms range from comprehensive genome-wide approaches (whole genome sequencing, whole exome sequencing, whole transcriptome sequencing) to clinically applicable targeted gene panels, the latter being more practical due to reduced cost and data complexity. An advantage of targeted NGS includes greater throughput, reduced data analysis burden, and lower cost and is therefore well suited to clinical practice. Targeted NGS provided greater diagnostic yield than histopathologic evaluation alone in detecting advanced biliary epithelial dysplasia, nearly doubling sensitivity (73% vs 48%) while preserving high specificity. Importantly, combining molecular analysis with routine pathology further improved case detection (sensitivity 83%) and maintained a specificity of 99%, highlighting the complementary value of the two approaches[62]. The most frequent gene analysis for indeterminate BSs work up is KRAS, TP53, SMAD4, CTNNB1, PIK3CA, ERBB2, BRAF, GNAS, and STK11[60,62].

NGS seems more sensitive than FISH, especially when combined with cytology. In a study by Dudley et al[62], who analyzed biliary brushing samples from 73 cases and pancreatic brushing samples from 8 cases, the 67% sensitivity from cytology alone improved to 76% when cytology was combined with FISH and to 85% when cytology was combined with NGS (Table 1). Sequencing results revealed KRAS as the predominant alteration, present in the majority of mutation-positive cases with TP53 ranking second in frequency. While KRAS mutations indicate neoplastic transformation, they occur early in tumorigenesis and are observed in premalignant conditions, limiting their discriminatory value between precursor and invasive disease. Conversely, TP53 alterations showed a much closer correlation with confirmed carcinoma on follow-up. These observations highlight the necessity of clinicopathologic correlation, especially when isolated KRAS mutations are identified.

Prospective evidence indicates that evaluating cell-free DNA recovered from brushing fluid during ERCP can substantially enhance diagnostic performance in BSs, increasing detection rates from 22% to 100%. By comparison, the addition of FISH achieved a diagnostic yield of 90%[39]. In addition, in a published study by Singhi et al[41], the combination of NGS and histology achieved a sensitivity of 83% with a specificity of 99% in 252 patients.

Molecular profiling using NGS is vital for accurate diagnosis and clinical decision making and can improve patient outcomes by distinguishing malignant from benign lesions and guiding targeted treatment selection for optimal outcomes. In patients with BSs bile duct brushing with targeted NGS revealed sensitive and specific diagnosis and identified potentially actionable genomic alterations[57]. ESMO Guidelines recommend early molecular profiling in patients diagnosed with CCA before or during first-line therapy for unresectable or metastatic disease[63].

miRNAs

miRNAs are short, non-coding RNA molecules, approximately 18-24 nucleotides in length, that modulate gene expression at the post-transcriptional level. They function mainly by targeting messenger RNAs and can be detected by PCR-based methodologies. A single miRNA can regulate many genes and messenger RNAs. Ambros and Horvitz[64] first discovered miRNAs in 1984 while studying a nematode. Various studies have shown that miRNAs are gene regulators conserved across species, and they can help in early cancer diagnosis. Lawrie et al[65] first identified miRNAs as biomarkers for cancer in 2008 in the serum of patients with B-cell lymphoma.

miRNAs are implicated in all steps of biliary and pancreatic carcinogenesis by functioning as oncogenes or onco-suppressor genes. Over the last years several miRNAs have been studied in patients with CCA, pancreatic adenocarcinoma, and chronic pancreatitis. miRNAs can be easily extracted from biopsies, body fluids, bile, and brushing tissue. Bile contains miRNAs that are stable and have prospective clinical utility as disease marker panels. Due to its direct anatomical proximity to the tumor site, bile may contain higher concentrations of tumor-derived molecular components compared with peripheral circulation. Although bile aspiration during ERCP is technically feasible and minimally burdensome, it remains underutilized in clinical practice. Incorporating bile-based molecular analysis could broaden the diagnostic framework by offering tumor-related insights beyond conventional cytology[44].

Bloomston et al[66] found that seven miRNAs (miR-99, miR-100, miR-100-1/2, miR-125a, miR-125b-1, miR-199a-1, miR199a-2) were overexpressed in resected pancreatic tissue in patients with pancreatic adenocarcinoma compared with healthy patients and those with chronic pancreatitis. Later, Zhang et al[67] analyzed the expression of 95 miRNAs but eventually concluded that eight miRNAs were significantly upregulated in most pancreatic cancer tissues and cell lines including miR196a, miR190, miR-186, miR-221, miR-222, miR-200b, miR15b and miR-95. The upregulation ranged from 70% to 100%[67]. miRNAs have also been used as serum biomarkers in patients with pancreatic cancer enhancing the diagnostic performance in cancer detection, especially miR-16 in combination with CA19-9[68]. Elevated expression of miR-10b, miR-30c, miR-106b, miR-155, and miR-212 was identified in both plasma and bile samples in the cohort analyzed by Cote et al[69], and this combined miRNA profile demonstrated strong diagnostic performance for pancreatic cancer.

Subsequently, miRNAs were studied in patients with CCA to assess their predictive value in diagnosis, prognosis, and therapy[70]. In 2019, Le et al[71] first studied miRNA expression in brush cytology specimens collected during ERCP. A total of 35 tissue samples were studied. Cytology specimens were first evaluated following miRNA isolation using miRNeasy Mini Kit (Qiagen, Hilden, Germany). Quantitative real-time PCR was performed for has-miR-16-5p, has-miR-21-5p, has-miR-196a and hse-miR-221-3p. Expression of miR-16, miR196a, and miR-221 showed a clear statistical significance between malignant and benign BSs. Additionally, combining routine cytology and the miR196-a single marker expression levels detected malignancy with a sensitivity of 84.6% compared with routine cytology alone with no false positives. miR-196-a proved to be the best biomarker for pancreatobiliary brushings in general. Although this was a single-center retrospective study with a small sample size, it clearly paved the way for future research in this field (Table 1).

Analysis of stored bile specimens by Kuniyosh et al[72] revealed significant upregulation of four miRNAs, miR-1275, miR-6891-5p, miR-7107-5p, and miR-3197 in patients with pancreatic cancer or CCA compared with benign controls. Notably, integrating bile cytology with miR-1275 expression achieved a sensitivity of 77.5% and a specificity of 100%. In a separate prospective cohort, Liu et al[73] evaluated bile collected during ERCP and percutaneous transhepatic cholangiography in 78 individuals presenting with obstructive jaundice, including cases of pancreatic ductal adenocarcinoma, CCA, and benign disease. Differential expression profiling identified miR-340 and miR-182 as markers capable of distinguishing malignant from non-malignant pancreaticobiliary conditions. Diagnostic performance analysis demonstrated moderate accuracy (area under the curve = 0.79; 95% confidence interval: 0.70-0.88) with sensitivity of 65% and specificity of 82%. Malignant samples exhibited upregulation of miR-182-5p and downregulation of miR-340-5p. Importantly, the differential expression of miR-340-5p persisted across subgroup analyses of both pancreatic ductal adenocarcinoma and CCA vs benign disease.

Growing evidence supports the use of exosome-derived molecules as potential biomarkers within liquid biopsy platforms. Exosomes are nano-sized extracellular vesicles released by cells that encapsulate nucleic acids, proteins, and other bioactive molecules, facilitating intercellular communication and reflecting the molecular characteristics of their cells of origin[43,44,74].

Severino et al[75] reported that bile samples from patients with CCA contained significantly higher concentrations of extracellular vesicles compared with samples from individuals with benign BSs. This difference was attributed to the increased metabolic activity characteristic of malignant disease. Extracellular vesicles are membrane-bound particles released by cells that include exosomes, microvesicles, and apoptotic bodies. These vesicles transport biologically active molecules such as DNA, RNA, proteins, and lipids shielded within a lipid bilayer that preserves their stability in bodily fluids and enables their use in diagnostic applications.

Analysis of bile obtained during ERCP identified a distinct miRNA profile miR-191, miR-486-3p, miR-1274b, and miR-484 in patients with CCA and pancreatic cancer as well as in individuals with chronic pancreatitis or biliary stone disease. In that cohort this miRNA signature demonstrated perfect discriminatory performance for CCA, reaching 100% sensitivity and 100% specificity.

In a recent study, Hipler et al[76] reviewed miRNAs successfully detected from the specimens obtained from the bile duct stents. miRNA assessment indicated increased expression of miR-16 and miR-223 among patients with malignancy although intergroup variability limited statistical discrimination between malignant and non-malignant cases. In contrast, cholangitis-associated samples demonstrated reproducible elevations in miR-16, miR-21, and miR-223 when compared with non-inflammatory controls. These observations imply that inflammatory processes within the biliary tract may significantly modulate miRNA profiles. Consequently, future biomarker development should focus on identifying candidates with minimal susceptibility to inflammation-related confounding.

CONCLUSION

ERCP brush cytology remains a cornerstone in the evaluation of indeterminate BSs due to its wide availability, technical ease, and safety profile. However, its limited sensitivity necessitates continued refinement and the integration of complementary approaches. Technical improvements such as rapid on-site evaluation, sheath-rinse techniques, and increased brushing passes have demonstrated incremental gains in diagnostic yield. Combining cytology with bile aspiration, forceps biopsy, or other sampling modalities further enhances sensitivity without significantly increasing procedural risk. In parallel, molecular diagnostics are transforming the field. FISH improves detection of chromosomal abnormalities, and NGS provides mutation profiling and identifies potentially actionable therapeutic targets. Emerging biomarkers, including miRNA and bile-derived liquid biopsies, hold promise but require further standardization and validation. Overall, current evidence supports a combined diagnostic strategy that incorporates optimized cytologic sampling with adjunctive molecular analyses. Future research should prioritize prospective trials that evaluate these combined methods, establish standardized protocols, and assess cost-effectiveness in routine clinical practice. Ultimately, improving the diagnostic accuracy of ERCP-based sampling is critical to guiding timely and appropriate management, reducing unnecessary surgery, and improving outcomes for patients with malignant BSs (Figure 2).

Figure 2 Diagnostic algorithm for the evaluation of indeterminate biliary strictures. Schematic diagnostic algorithm for indeterminate biliary strictures showing the role of endoscopic retrograde cholangiopancreatography brush cytology, refinements, biopsy, and molecular diagnostics, culminating in a multimodal strategy to improve outcomes (in cases in which endoscopic retrograde cholangiopancreatography with brush cytology is not diagnostic, spyglass cholangioscopy is recommended). CA19-9: Carbohydrate antigen 19-9; CEA: Carcinoembryonic antigen; MRCP: Magnetic resonance cholangiopancreatography; ERCP: Endoscopic retrograde cholangiopancreatography; ROSE: Rapid on-site evaluation; FISH: Fluorescence in situ hybridization; NGS: Next-generation sequencing; miRNA: MicroRNA.