Cholangioscopy in the diagnosis and management of cholangiocarcinoma

Abstract

Cholangiocarcinoma is a late-presenting and aggressive malignancy that remains difficult to diagnose and treat. Early detection remains challenging due to its insidious onset and nonspecific symptoms, which necessitate advancements in diagnostic and therapeutic strategies. Conventional imaging modalities often lack sufficient sensitivity, while endoscopic retrograde cholangioscopy-based sampling is limited by a low diagnostic yield. However, digital single-operator cholangioscopy has emerged as a transformative tool, significantly enhancing diagnostic accuracy through direct visualization, targeted biopsies, and standardized classification systems. Beyond its diagnostic applications, cholangioscopy plays a critical role in therapeutic interventions, facilitating precise biliary drainage, stent placement, and intraductal therapies such as photodynamic therapy and radiofrequency ablation. The integration of artificial intelligence into cholangioscopy holds the potential to further refine real-time diagnosis and therapeutic decision-making. As cholangioscopy evolves, its expanding role in the multidisciplinary management of cholangiocarcinoma is reshaping clinical practice. This review highlights its growing impact and explores future directions for research and innovation in biliary disease management.

Key Words: Cholangiocarcinoma; Single-operator cholangioscopy; Monaco classification; Carlos-Medranda criteria; Photodynamic therapy; Radiofrequency ablation; Intraductal lithotripsy

Core Tip: Cholangiocarcinoma (CCA) remains a major challenge due to its late presentation and limited treatment options. Digital single-operator cholangioscopy has revolutionized the diagnosis and management of CCA by improving diagnostic accuracy through direct visualization, targeted biopsies, and standardized classification systems. It also facilitates therapeutic interventions, including biliary drainage, stenting, and intraductal therapies such as photodynamic therapy and radiofrequency ablation. The integration of artificial intelligence holds promise for further advancements, solidifying cholangioscopy’s role in the multidisciplinary management of CCA.

INTRODUCTION

Cholangiocarcinoma (CCA) is an uncommon but highly aggressive malignancy of the bile ducts, characterized by its insidious onset and generally poor prognosis[1,2]. Its incidence continues to rise globally, particularly among patients with chronic liver disease or underlying predisposing conditions such as primary sclerosing cholangitis[1,2]. Regional trends exhibit substantial geographic variation, with Southeast Asia exhibiting disproportionately higher rates[1-3]. Therefore, early detection and proper staging are crucial to optimize patient outcomes. While imaging techniques such as computed tomography, magnetic resonance imaging, and endoscopic ultrasound are widely used in diagnosis[4,5], cholangioscopy has emerged as a powerful tool for the direct visualization, staging, and management of CCA. This review outlines the evolving role of cholangioscopy in enhancing diagnostic yield, facilitating targeted interventions, and guiding therapeutic decisions in patients with CCA, particularly when conventional modalities fail to provide a definitive diagnosis.

PATHOGENESIS AND RISK FACTORS

CCA arises through a multifactorial process involving chronic inflammation, bile duct injury, and genetic predispositions[4]. Risk factors include: (1) Hepatobiliary diseases (e.g., primary sclerosing cholangitis); (2) Chronic liver diseases (e.g., viral hepatitis); (3) Genetic disorders such as Lynch syndrome; and (4) Parasitic liver fluke infections in endemic regions[5-7]. These factors induce chronic inflammation, fibrosis, and cellular dysregulation, culminating in malignant transformation[5-7].

CLASSIFICATION AND HISTOPATHOLOGY

CCA is classified based on its anatomical location along the biliary tree: (1) Intrahepatic CCA, arising proximally to the secondary bile ducts; (2) Perihilar CCA, arising between the second-order bile ducts and the insertion of the cystic duct; and (3) Distal CCA, arising distal to the insertion of the cystic duct to the ampulla of Vater[7]. Perihilar tumors are further categorized using the Bismuth-Corlette classification, which delineates tumor involvement relative to the biliary confluence and hepatic ducts[8]. Histologically, > 90% of CCAs are adenocarcinomas with glandular structures and desmoplastic stroma[9]. Molecular defects, including oncogenes (e.g., rat sarcoma viral oncogene homolog, erb-b2 receptor tyrosine kinase 2), mutations such as isocitrate dehydrogenase 1 and fibroblast growth factor receptor 2 mutations, and tumor suppressors (e.g., p53, SMAD family member 4), are associated with aggressive phenotypes[10].

DIAGNOSTIC CHALLENGES

The anatomical nature of CCA often delays symptom onset and diagnosis. Local progression may present with symptoms such as jaundice, pruritus, and abdominal pain, but these are often nonspecific. When CCA is suspected, transpapillary brushing (brush cytology) is a standard endoscopic diagnostic technique. Performed during endoscopic retrograde cholangiopancreatography (ERCP), brush cytology collects cellular samples from the bile ducts by brushing their lining[11-13]. The procedure begins with the placement of a guidewire under fluoroscopic guidance to ensure proper access to the target area[13-15]. A cytology brush, typically sheathed for protection, is then inserted into the strictured bile ducts. The sample is collected by repeatedly moving the brush across the stricture to exfoliate cells for analysis. The brush is subsequently removed, and the collected material is prepared for cytological examination. This method is cost-effective, widely available, and relatively simple to perform[12-14]. However, transpapillary brushing has a low diagnostic yield, with sensitivity ranging from 35% to 52% but a high specificity of 95%-98%[12]. Combining brush cytology with fluorescence in situ hybridization (FISH), a molecular cytogenetic technique used to detect specific DNA sequences, improves sensitivity to 43%-60% while maintaining a specificity of 91%-95%[13].

Despite its utility, brush cytology has several limitations. It presents challenges in the precise targeting of lesions, and sampling issues often arise due to inadequate cellularity, particularly in patchy or fibrotic strictures, reducing diagnostic accuracy[16]. Furthermore, a negative result frequently necessitates further diagnostic testing[16,17]. Stricture characteristics, such as length and size (> 1 cm for both), influence the likelihood of obtaining positive results, limiting their effectiveness for smaller or less distinct strictures[16,17]. Moreover, brush cytology is highly operator-dependent, with diagnostic yield varying based on sampling techniques, number of brushing passes, and sample preparation methods[16-18].

Another diagnostic method for indeterminate biliary strictures is fluoroscopic-guided biopsy sampling[18-21]. Performed during ERCP, this technique involves advancing biopsy forceps over a guidewire across the stricture under fluoroscopic guidance. After positioning at the target site, multiple samples are collected to ensure adequate tissue retrieval. Fluoroscopic-guided biopsy has demonstrated a sensitivity of 43%-81% and specificity of 90%-100% for detecting malignancies, particularly in proximal lesions that are more accessible[18-20]. Another technique involving modified biopsy forceps (fluoroscopy-shaped endobiliary biopsy forceps) has demonstrated a higher rate of sensitivity at 81.1% and 88.2% accuracy in per-patient analysis[18]. Overall, both methods have their drawbacks, including relatively low sensitivity, a high false-negative rate, and the need for multiple samples or complementary techniques to enhance diagnostic accuracy[18-22]. Sampling difficulties are common, especially in tight or fibrotic strictures, potentially leading to incomplete tissue collection and false negatives[18,21]. Additionally, fluoroscopic-guided biopsy requires a higher level of technical skill, as success is heavily dependent on operator expertise and the quality of imaging guidance, given the absence of direct endoscopic visualization[21,22]. This limitation may also increase the risk of major adverse events such as perforation and bleeding[21]. While a systematic review found no statistical difference in adverse events between brush cytology and intraductal biopsy sampling, more complications were reported in the fluoroscopic biopsy group, including two severe cases requiring surgical intervention[21].

According to guidelines from the American Society for Gastrointestinal Endoscopy, ERCP for biliary strictures of undetermined aetiology should include fluoroscopic-guided biopsy sampling alongside brush cytology rather than brush cytology alone[21]. This recommendation is based on a systematic review and meta-analysis of 21 observational studies involving 2726 patients, which compared ERCP with combined fluoroscopicbiopsy and brush cytology vs brush cytology alone. The study found that adding biopsy sampling increased the diagnostic yield by 20%[21]. The miss rate for malignancy was 58% with brush cytology alone, but improved to 41% with biopsy sampling alone[21]. Sensitivity was 40% for brush cytology, 52% for biopsy sampling, and 66% when the two techniques were combined[21]. While both methods demonstrated similar technical success rates, fluoroscopic-guided biopsy required greater expertise due to the lack of direct endoscopic visualization[21].

Successful biliary cannulation is extremely important in the evaluation of CCA. In up to 18% of ERCPs, difficult cannulation limits access to the biliary system, particularly in patients with altered anatomy or malignant strictures, which may delay diagnosis and treatment[23]. Prolonged cannulation attempts also increase the risk of post-ERCP pancreatitis[23]. Several techniques, including trans-pancreatic sphincterotomy and double-guidewire methods, have been shown to improve cannulation success and reduce complications. A 2022 meta-analysis of 17 randomized controlled trials (n = 2015) found that trans-pancreatic sphincterotomy significantly increased successful biliary access while lowering the incidence of post-ERCP pancreatitis[23]. Recognizing and managing difficult biliary access is therefore a critical preparatory step, as failed or traumatic cannulation can limit passage and increase procedural risk. Incorporating advanced access techniques and guidewire-assisted cannulation strategies enhances both feasibility and safety.

Diagnosing indeterminate biliary strictures and, by extension, CCA, remains challenging due to late symptom onset and the limitations of conventional diagnostic methods. While techniques such as brush cytology and fluoroscopic-guided biopsy contribute to diagnosis, they are hindered by low sensitivity, sampling challenges, and operator dependence. Combining these methods improves detection rates but does not fully overcome their limitations. Cholangioscopy represents a significant advancement by allowing direct visualization of the bile ducts and targeted tissue sampling, offering the potential to enhance diagnostic accuracy beyond traditional approaches.

PERORAL CHOLANGIOSCOPY

Peroral cholangioscopy was first introduced in the 1970s to enable direct visualization of the biliopancreatic tree for diagnosing and treating intraductal lesions[24-27]. The original method, known as the “baby-mother” system, required two endoscopists - managing the duodenoscope (the “other”) and the other handling the cholangioscope (the “baby”)[24]. However, this early approach was significantly limited by technical challenges, including poor steerability, high repair costs, and inadequate irrigation capabilities[24]. Over time, peroral cholangioscopy has undergone substantial advancements. The first single-operator cholangioscopy system was introduced in 2005, leading to the launch of the SpyGlass™ system in 2007 and the SpyGlass Direct Visualization System in 2015, both developed by Boston Scientific[24-26]. These innovations allowed a single endoscopist to perform the procedure independently, improving efficiency and usability. The latest digital single-operator cholangioscopy (DSOC) systems feature numerous technical improvements, including: Higher image resolution, enhanced visualization of suspicious strictures for more targeted biopsies, a wider field of view, a brighter light source, an increased therapeutic channel lumen diameter, a tapered tip, and four-way tip deflection[24,25]. Device innovation has continued, including new DSOC platforms such as EyeMAX™, which utilizes a 2 mm working channel capable of accommodating pediatric forceps (1.6 mm)[27]. A single-center study highlighted increased ease of use, higher user satisfaction, and superior biopsy facilitation[27]. With these technological improvements, both the diagnostic and therapeutic applications of cholangioscopy have expanded. Current diagnostic uses include evaluating indeterminate bile duct strictures, detecting choledocholithiasis, mapping intraductal tumors, and assessing dominant strictures in primary sclerosing cholangitis. Therapeutic applications will be discussed in the following section.

DIAGNOSTIC ROLE OF CHOLANGIOSCOPY

DSOC plays a crucial role in diagnosing CCA through two primary mechanisms: (1) Tissue biopsy acquisition; and (2) Visual interpretation of indeterminate strictures.

Biopsy diagnosis

Multiple studies have evaluated the diagnostic utility of DSOC-guided biopsy in biliary strictures of unknown aetiology, including CCA. A recent randomized multicenter trial at three tertiary care centers (61 patients: 32 in the study arm, 29 in the control arm) compared standard ERCP with tissue brushing to DSOC with DSOC-guided biopsy[28]. The results demonstrated significantly higher sensitivity for DSOC-guided biopsy (68.2% vs 21.4%, P < 0.1), improved visualization sensitivity (95.5% vs 66.7%, P = 0.05), and greater overall accuracy (87.1% vs 65.5%, P = 0.05)[28]. Positive predictive value and negative predictive value showed no significant differences. Adverse events, including cholangitis, cholecystitis, pancreatitis, and bleeding, occurred at comparable rates in both groups[28].

A systematic review by Rey Rubiano et al[29] analyzed the role of DSOC in diagnosing indeterminate biliary strictures. The pooled sensitivity from two primary studies (143 patients) was 64.9%, with a specificity of 100%, compared to ERCP with brushing and/or biopsy, which had a sensitivity of 51% and a specificity of 100%[29]. While adverse event rates were statistically similar, there was a noted risk of postprocedural cholangitis following DSOC[26]. A retrospective single-center analysis by Büringer et al[30] assessed DSOC in 196 patients, with 108 undergoing DSOC-guided biopsies. Of these, 93 cases were true positives and 15 were false negatives, resulting in a sensitivity of 86%, specificity of 99%, positive predictive value of 90.5%, negative predictive value of 98.5%, and a likelihood ratio of 86 - highlighting DSOC’s high diagnostic reliability for indeterminate strictures[30]. A systematic review and meta-analysis by Fujii-Lau et al[21] evaluated 13 studies (12 observational, 1 randomized controlled trial) involving 1529 patients undergoing ERCP with either cholangioscopy or other tissue acquisition methods (fluoroscopic-guided biopsy sampling, brush cytology, or both). The study found an incremental yield of 27% [95% confidence interval (CI): 9%-46%] with ERCP plus cholangioscopy vs ERCP alone in four observational studies and 41% (95%CI: 11%-72%) in the randomised controlled trial[21]. Sensitivity was significantly higher with DSOC [0.72 (95%CI: 0.66-0.77) vs 0.61 (95%CI: 0.57-0.66), P = 0.001][21]. No differences were noted in technical success, specimen adequacy, or adverse event rates between groups[21]. Recent prospective real-world series and multicenter evaluations further confirm DSOC’s improved diagnostic sensitivity over conventional ERCP-based sampling and support integration of adjunct diagnostic modalities such as intraductal ultrasound for challenging strictures. For example, a 2024 multicenter/diagnostic study reported higher sensitivity for DSOC-targeted biopsies vs brush cytology and emphasized DSOC’s complementary role alongside intraductal ultrasound[31].

Visual diagnosis

DSOC also facilitates direct visual interpretation of indeterminate biliary strictures. A 2020 systematic review and meta-analysis by de Oliveira et al[32] assessed the efficacy of DSOC for this purpose. Reviewing six studies comprising 283 procedures, the authors reported an overall pooled sensitivity of 94% (95%CI: 89%-97%) and specificity of 95% (95%CI: 90%-98%) for visually diagnosing biliary malignancies. Two additional studies evaluating DSOC’s visual interpretation capabilities reported sensitivity rates of 85%-90% and specificity rates of 80%-90%[33,34]. The high diagnostic value of DSOC in visually identifying malignancies has led to the development of standardized criteria for distinguishing between benign and malignant biliary strictures. These are described below.

Monaco criteria: This is a standardized classification system designed to evaluate and differentiate malignancy and benign biliary strictures during cholangioscopy. It utilizes 8 key visual features observed during DSOC to improve diagnostic accuracy for indeterminate biliary strictures[19]: (1) Presence of stricture (asymmetric or symmetric); (2) Presence of lesion (mass > duct, or nodule < duct, or polypoid appearance); (3) Mucosal features (smooth or granular); (4) Papillary projections [fingerlike (long or short)]; (5) Ulceration; (6) Abnormal vessels; (7) Scarring (local or diffuse); and (8) Pronounced pit pattern[19]. Research has shown that when utilizing the Monaco criteria, sensitivity ranges from 80% to 85%, and specificity is around 90%[33,35]. In the clinical setting, the Monaco criteria have greatly enhanced diagnostic confidence when diagnosing CCA. It allows for standardization for the interpretation of visual findings, especially when cytology or biopsy results might be inconclusive, and allows for high reproducibility, especially amongst highly experienced endoscopists[33,35].

Carlos-Medranda criteria: Robles-Medranda et al[36] developed a novel macroscopic classification for bile duct lesions. They performed a study where they analyzed images captured during cholangioscopy and correlated them to histology findings, allowing them to come up with a classification system that differentiated imaging findings into non-neoplastic and neoplastic lesions based on both morphological and vascular patterns (Figure 1). Utilizing this classification system, they were able to find that the sensitivity and specificity for neoplastic diagnoses was 96.3% and 92.3%, respectively[36]. Additionally, the classification system allowed for high reproducibility amongst observers for both neoplastic and subtype groups, but it was more pronounced in experts (> 80%) than non-experts (64.7%)[36].

Figure 1 Carlos-Medranda classification: Novel macroscopic classification for bile duct lesions. Adapted from Robles-Medranda et al[36].

Recent interobserver studies and an atlas of cholangioscopy imaging highlight both the potential of standardized image atlases and the persistent variability in visual interpretation. A 2024 single-center real-world series further supports this: Among 14 patients undergoing DSOC for indeterminate strictures, video-clip review of 29 cases using the Monaco criteria yielded almost perfect agreement for final diagnosis (κ = 0.871; 93.1% agreement) and accuracy of 73.6% for experts’ vs 64.4% for trainees[37].

From a financial standpoint, although DSOC has a higher upfront cost, studies have demonstrated its cost-effectiveness due to its ability to reduce the overall number of procedures required for malignancy diagnosis[21,38,39]. A study comparing DSOC with ERCP-guided brushing for FISH polysomy found that DSOC was more cost-effective, as its higher diagnostic yield minimized the need for multiple biopsies[39]. Despite significant advancements in DSOC’s role in diagnosing indeterminate biliary strictures and CCA, several barriers have limited its widespread adoption. As a specialized procedure, DSOC requires additional training to ensure proper execution. Moreover, its availability remains limited outside high-volume centers. Some studies have also reported biopsy-related sampling errors, which can impact diagnostic accuracy[39-43]. Another concern is the potential for tumor seeding in the proximal biliary tree if adequate drainage is not ensured, as DSOC involves the instillation of water or saline solution[21]. There is also increased concern for infection, as a study evaluating bacteraemia after cholangioscopy found a bacteraemia rate of 8.8% among 57 patients, with higher rates observed in cases involving tissue biopsy collection[44]. To mitigate this risk, experts recommend ensuring proper drainage before obtaining biopsies[21,44].

Even with these limitations, ASGE recommended the use of DSOC in the diagnosis of biliary strictures of undetermined aetiology in the following scenarios: (1) Non-distal biliary strictures where there is a high probability of adequate drainage of the critical liver segment; (2) Previous non-diagnostic ERCP without cholangioscopy; and (3) Centers with clinical expertise and easy access to the equipment[21]. Cholangioscopy has emerged as a powerful diagnostic tool for CCA. While certain limitations exist, the significant improvements in diagnostic accuracy, coupled with its economic advantages, make DSOC a valuable and evolving technique in biliary disease evaluation. Further research is necessary to enhance its utilization, particularly in refining visual diagnostic tools to guide tissue sampling. Table 1 for a comparison between the Carlos-Medranda criteria and the Monaco criteria (Table 1).

Table 1 Comparison of the Monaco and Carlos-Robles-Medranda classification systems for digital single-operator cholangioscopy visual diagnosis.

	Monaco classification	Carlos-Robles-Medranda classification
Ref.	Sethi et al[35], 2022	Robles-Medranda et al[36], 2018
Primary purpose	Standardize DSOC visual features distinguishing benign vs malignant biliary strictures	Provide a macroscopic classification correlating morphologic and vascular patterns with neoplastic vs non-neoplastic lesions
Key diagnostic parameters	Eight features: (1) Stricture symmetry; (2) Presence of lesion; (3) Mucosal surface; (4) Papillary projections; (5) Ulceration; (6) Abnormal vessels; (7) Scarring; and (8) Pit-pattern	Four morphologic patterns subdivided by vascularity: Non-neoplastic: Villous, polypoid, or inflammatory with regular vascularity; neoplastic: Flat, polypoid, ulcerated, or honeycomb with irregular/spider vascularity
Representative visual cues	Irregular vessels, papillary or nodular mucosa, asymmetric stricture, ulceration, or diffuse scarring	Disrupted vascular network, irregular or spider vessels, loss of normal pit pattern, polypoid or honeycomb architecture
Training/ease of use	Relatively simple checklist (8 binary variables); designed to facilitate teaching and reproducibility among non-experts	Requires detailed morphologic assessment; higher interpretive demand, but integrates vascular evaluation, improving histologic correlation
Diagnostic accuracy	Sensitivity approximately 80%-85%, specificity approximately 90% in expert hands	Sensitivity approximately 96%, specificity approximately 92% for neoplastic lesions
Inter-observer agreement	Moderate agreement overall (κ approximately = 0.31-0.52); improved with experience and structured training	Higher reproducibility - excellent among experts (κ approximately = 0.80-0.83) and substantial among non-experts (κ approximately = 0.65)
Strengths	Simple, reproducible, and suitable for multicenter teaching and video-library scoring; good standardization for visual training modules	Strong histopathologic correlation, incorporates vascular and morphologic features, higher diagnostic performance, and IOA
Limitations	Limited vascular assessment; relies heavily on mucosal pattern recognition; moderate IOA in trainees	More complex and time-consuming; may require advanced image quality and operator expertise
Overall summary	Practical and training-friendly system emphasizing accessibility and reproducibility	Pathology-driven classification offering superior accuracy and inter-observer reliability, but requiring more expertise

DSOC: Digital single-operator cholangioscopy; IOA: Interobserver agreement.

THERAPEUTIC ROLE OF CHOLANGIOSCOPY

DSOC can also be utilized as a therapeutic tool in managing CCA. The following section examines the role and advancements of DSOC in the treatment of CCA[43,45].

Drainage and stenting of biliary strictures

Digital DSOC provides high-resolution direct visualization of the biliary tree, enabling orifice identification and precise guidewire advancement through biliary strictures[46]. These advancements have enhanced the success rate of biliary drainage and stenting, particularly in angulated or tight strictures. A retrospective analysis of 167 procedures identified 30 cases of DSOC-assisted guidewire placements across biliary strictures, achieving a success rate of 46.2% in malignant strictures, and initial procedures were more successful than repeat attempts. Although adverse events such as pancreatitis, cholangitis, and bleeding also occurred (16.7%), DSOC proved to be a valuable alternative to invasive procedures like percutaneous transhepatic or endoscopic ultrasound-guided biliary drainage[47]. A case series by Kastelijn et al[48] studying clinical outcomes of DSOC selective cannulation for complex biliary stents after failed ERCP showed a success rate of 50%, all of which had improved bilirubin levels after. Only 1 out of the 10 patients had post-ERCP pancreatitis. These results highlight the value of cholangioscopy in the selective cannulation and stent insertion of complex biliary strictures.

Intraductal ablative therapies

Intraductal ablative therapies target malignant strictures via cholangioscopy to reduce tumor burden and improve stent patency.

Photodynamic therapy: Studies have suggested that photodynamic therapy (PDT) might be beneficial for the palliation of hilar CCA, which is an uncommon biliary cancer that is usually unresectable. PDT’s main role is to treat unresectable CC, improving the quality of life and prolonging survival[49]. PDT involves administering an IV photosensitizing agent that is preferentially absorbed by neoplastic tissue, followed by activation with a locally applied specific wavelength of light, leading to selective tumor cell destruction. Evidence shows that PDT can potentially improve survival and quality of life in patients with unresectable CCA[50-52]. The systematic review and meta-analysis by Mohan et al[51] showed that patients receiving PDT in conjunction with biliary stenting showed a statistically significant improvement in overall survival compared to stenting alone, resulting in a 51% reduction in the risk of death. The study also showed increased stent patency with PDT, suggesting a 39% reduction in stent occlusion. A randomized prospective study of 39 patients with nonresectable CCA by Ortner et al[50] highlighted that PDT, when combined with biliary stenting, was associated with improved stent patency and survival compared to stenting alone[52]. PDT offers extensive benefits for treating biliary strictures, particularly when targeting tumor cells and sparing healthy tissue. The laser light via PDT refracts through bile, allowing it to treat malignant tissue not in direct proximity to the laser fiber. Despite the advantages, PDT has limitations, including photosensitivity, which requires patients to avoid light post-treatment. Also, it is relatively more expensive than alternative therapies. Due to its limited penetration depth, deeper tumor cells may be unaffected by treatment. PDT is associated with procedural adverse events, such as those reported in Mohan et al[51] systematic review, showing about three times the odds of complications with PDT compared to stenting alone (pooled odds ratio = 2.79, 95%CI: 1.14-6.82). Further research needs to establish definitive conclusions on the role of PDT in the treatment of CCA.

Radiofrequency ablation: Radiofrequency ablation (RFA) uses thermal energy to induce coagulative tissue necrosis via a cholangioscope. DSOC-guided RFA has been explored for the palliative treatment of unresectable CCA[50]. The DSOC-guided administration of RFA is showing excellent results. Data from Yang et al[53] highlight the effectiveness of endoscopic RFA combined with stenting compared to stenting alone in the management of unresectable extrahepatic CCA. Specifically, the study found that patients in the RFA + stent group had a significantly longer stent patency period (6.8 months) compared to the stent-only group (3.4 months). These results underscore the potential benefit of using RFA with stenting to improve biliary drainage and reduce the recurrence of duct obstruction by CCA[53]. Another study by Owen et al[54] also indicated that locoregional therapy (including RFA) could improve biliary drainage and stent patency in patients with intrahepatic CCA. Another retrospective study with 12 patients with CCA who underwent RFA under DSOC guidance showed technical success of 100%, with only one patient developing post-operative cholangitis[55]. Both PDT and RFA are effective endoscopic therapies for malignant biliary strictures. PDT provides broader treatment coverage by targeting tumor cells through refracted laser light. However, it is more expensive and may cause photosensitivity and adverse events during the procedure. RFA, on the other hand, is less invasive and quicker. Conversely, it relies on direct tumor contact, making it less suitable for certain complex bile duct anatomy. Current evidence does not show a definitive superiority of one modality over the other, and there is limited data on long-term survival benefits post-intraductal therapy[56].

Although both PDT and RFA have demonstrated improved stent patency and survival compared to stenting alone, comparative data remain limited. A 2022 meta-analysis by Mohan et al[51] included 10 studies and found no statistically significant difference in overall survival between PDT and RFA (hazard ratio = 0.90, 95%CI: 0.67-1.20), though PDT was associated with marginally longer stent patency. RFA demonstrated a lower incidence of cholangitis and procedure-related adverse events, while PDT was limited by photosensitivity reactions and higher cost. Similarly, Mohammad and Kahaleh[50] reported that RFA may offer greater procedural efficiency and cost-effectiveness, whereas PDT provides broader luminal coverage for multifocal lesions. Collectively, these findings suggest that modality selection should be individualized - RFA favored for safety and accessibility, and PDT considered for extensive intraductal disease - pending high-quality randomized head-to-head trials[50,51].

Intraductal lithotripsy for associated bile duct stones

Conventional ERCP successfully treats most bile duct stones via sphincterotomy and stone extraction. In approximately 10% of patients, standard ERCP is unsuccessful[57]. These are usually complex biliary stones that are large, multiple, located in smaller ducts (intrahepatic/cystic ducts), and/or associated with strictures. Usually, additional invasive methods are needed to remove these stones, like mechanical lithotripsy or endoscopic balloon dilation[43]. DSOC has emerged to be useful in the management of these complex stones, which have failed standard ERCP treatment. Real-time visualization by DSOC enables the use of intracorporeal lithotripsy to dissolve these stones, using techniques like electrohydraulic or laser lithotripsy[57].

USE OF ARTIFICIAL INTELIGENCE IN CHOLANGIOSCOPY

Recent research studies have explored the integration of artificial intelligence (AI) with cholangioscopy to improve the diagnosis and treatment of CCA. A multicenter prospective trial involving 26 cholangioscopy examinations evaluated an AI-assisted system for detecting biliary strictures and managing choledocholithiasis. The study found that AI analysis had a diagnostic accuracy of 91%, which is more accurate than the standard endoscopists’ visual assessment. AI also improved the sensitivity and specificity for detecting malignancy, reaching 89% and 92%, respectively. This suggests that AI-enhanced cholangioscopy can provide real-time diagnostic support and potentially reduce the need for repeat procedures[58].

Additionally, a systematic review published in March 2023 emphasized that AI-assisted imaging techniques, which include machine learning algorithms, demonstrated a sensitivity of up to 94% and a specificity of 88% when differentiating malignant from benign biliary strictures[59]. AI can be integrated to aid in the automatic detection of biliary cancers in real-time. One study reported that AI-assisted cholangioscopy improved biliary malignancy detection rates by 30% compared to conventional methods and reduced the need for tissue biopsies by 25%[60].

The field is rapidly moving from retrospective model development to prospective, multicenter validation. A 2024 prospective feasibility trial demonstrated that real-time AI-assisted cholangioscopy can feasibly augment endoscopist interpretation and diagnostic yield, although larger effectiveness and implementation studies are still required[61]. In this trial, a real-time cholangioscopy AI-computer-aided diagnosis system achieved 94.4% overall accuracy (95%CI: 72.7%-99.9%) in distinguishing malignant from non-malignant biliary lesions, outperforming both brush cytology (62.5%) and forceps biopsy (75.0%) while delivering predictions at 8 frames per second[60]. These findings highlight the rapid maturation of AI-assisted cholangioscopy toward real-world application, though larger multicenter validation studies remain essential to confirm generalizability, assess workflow integration, and define its ultimate role in routine DSOC interpretation. Overall, the integration of AI in cholangioscopy shows promising results in the detection of CCA. AI-assisted cholangioscopy is showing promise of becoming a useful tool in clinical practice, reducing the use of invasive diagnostic methods and improving patient outcomes.

CONCLUSION

CCA is an aggressive malignancy with a rising global incidence. Early detection remains critical, though challenging due to its insidious onset and non-specific symptoms. While traditional diagnostic methods such as brush cytology, FISH, and fluoroscopic-guided biopsy have limited sensitivity, DSOC has emerged as a transformative tool. DSOC enhances diagnostic accuracy through direct visualization and targeted biopsies, supported by standardized criteria like the Monaco and Carlos-Medranda classifications. Beyond diagnostics, DSOC plays a pivotal role in the therapeutic management of CCA. It improves biliary drainage, enables precise stent placement, and facilitates intraductal ablative therapies like PDT and RFA, which have shown promising results in prolonging stent patency and improving survival in patients with unresectable CCA. Furthermore, DSOC aids in the management of complex bile duct stones, expanding its utility. Despite its advantages, cholangioscopy is resource-intensive and operator-dependent, necessitating further advancements to optimize accessibility and outcomes. Its integration into multidisciplinary care highlights its importance as a cornerstone in the evolving landscape of CCA diagnosis and management.

Future investigations should prioritize prospective, multicenter validation of AI models for real-time lesion characterization during DSOC, integrating standardized video datasets and assessing clinical impact on diagnostic yield and decision-making. Parallel efforts are needed to shorten the DSOC learning curve through structured training programs, validated classification systems, and AI-assisted decision support to enhance interobserver agreement and operator performance. Finally, development of minimally invasive, cholangioscopy-guided combinations of intraductal therapies, such as PDT or RFA, coupled with localized chemotherapy, represents a promising avenue to improve local disease control while minimizing systemic toxicity[61-63].