Diagnostic value of peroral choledochoscopy combined with modified biopsy forceps for malignant biliary strictures: a two-center retrospective study Authors: Geqi Zhu, Yu Jin, Cuihua Qi, Shuxin Tian, Hao Liu, Weigang Chen, Rong Lin, Guochen Shang Published: 2026 ABSTRACT Background and aimsPathological evaluation is crucial for diagnosing biliary strictures. This study aimed to evaluate the diagnostic efficacy and safety of peroral cholangioscopy combined with modified biopsy forceps for malignant biliary strictures.MethodsWe retrospectively collected data from inpatients with suspected extrahepatic malignant biliary strictures who underwent peroral cholangioscopy at Union Hospital of Tongji Medical College, Huazhong University of Science and Technology, and The First Affiliated Hospital of Shihezi University from January 2020 to March 2025. The experimental group used peroral cholangioscopy combined with modified biopsy forceps, and the control group used peroral cholangioscopy. We compared the diagnostic efficacy and safety of the two biopsy methods.ResultsA total of 158 patients were included in the study, with 79 patients each in the experimental group and the control group. The detection rate of carcinoma in situ (high-grade intraepithelial neoplasia/severe dysplasia) was significantly higher in the experimental group than in the control group (32.9% vs 6.3%, P < 0.001). Regarding diagnostic efficacy and safety, the accuracy (93.7% vs 72.2%, P < 0.001), sensitivity (93.2% vs 71.4%, P < 0.001), negative predictive value (50.0% vs 8.3%, P = 0.014) and negative likelihood ratio (0.068 vs 0.286) of the experimental group were significantly superior to those of the control group. There was no significant difference between the two groups in terms of specificity (100.0% vs 100.0%, P = 1.000), positive predictive value (100.0% vs 100.0%, P = 1.000), or the incidence of adverse events (7.6% vs 8.9%, P = 1.000).ConclusionsCompared with POCS-guided biopsy using dedicated forceps, peroral cholangioscopy combined with modified biopsy forceps demonstrates higher diagnostic efficacy for malignant biliary strictures and maintains a good safety profile.Graphical abstract Biliary strictures (BS) are abnormal narrowings of the intrahepatic or extrahepatic bile duct drainage system caused by various factors, usually leading to obstructed bile flow and cholestasis resulting in obstructive jaundice. Although symptoms vary based on severity, the condition may not cause symptoms or abnormalities in liver tests during the early stages [1]. The etiology of biliary strictures is complex and diverse. Based on their pathological nature, they can be divided into benign biliary strictures (BBS) and malignant biliary strictures (MBS). BBS are mostly caused by non-tumor factors such as postoperative iatrogenic injury, stones, inflammation, extrinsic compression, and autoimmune diseases like primary sclerosing cholangitis, IgG4-related cholangitis, and autoimmune pancreatitis [2]. Studies indicate that approximately 15% to 24% of patients with suspected MBS are confirmed to have benign strictures by pathology after surgical resection [3]. For patients who actually have BBS, unnecessary surgical intervention is the primary clinical consequence resulting from the misdiagnosis of the nature of the biliary stricture [4]. MBS are mainly caused by cholangiocarcinoma (CCA), gallbladder cancer, pancreatic cancer, liver cancer, and metastatic cancer invading or compressing the bile duct. These diseases are associated with a high degree of malignancy and are difficult to diagnose [5]. CCA is the most common disease causing MBS. It is a malignant tumor originating from the bile duct mucosal epithelium and can be divided into intrahepatic cholangiocarcinoma (ICC) and extrahepatic cholangiocarcinoma (ECC) based on the anatomical location. ECC is further divided into perihilar cholangiocarcinoma (pCCA) and distal cholangiocarcinoma (dCCA) [6]. Currently, the differential diagnosis between benign and malignant biliary strictures remains a clinical challenge. Cell brushing and biliary biopsy performed during endoscopic retrograde cholangiopancreatography (ERCP) are the main methods to determine the nature of BS. Since ERCP is performed under non-direct visualization and maneuvering biopsy forceps into the common bile duct is technically difficult, obtaining lesion tissue is challenging. This often leads to insufficient sampling and inaccurate positioning, which affects the diagnostic accuracy for malignant biliary strictures (MBS) and fails to meet clinical requirements. Peroral cholangioscopy allows for direct visualization of the internal structure of the bile duct and facilitates targeted biopsy of lesions, resulting in a higher diagnostic accuracy for MBS [7]. However, the tissue samples obtained by POCS biopsy forceps are small. Especially for non-mass-forming lesions, POCS biopsy is very difficult, which may lead to reduced diagnostic sensitivity. Combined with our preliminary study, which found that using modified biopsy forceps for biliary biopsy can improve the accuracy and sensitivity of diagnosing MBS and reduce the difficulty of biopsy [8], and considering that the visualization features of POCS allow for direct observation of the stricture and precise operation, this study hypothesizes that applying POCS combined with modified biopsy forceps can further enhance the diagnostic capability for MBS and effectively solve the difficulty of POCS biopsy. In summary, this study aims to evaluate the diagnostic value and safety of POCS combined with modified biopsy forceps for MBS. Patients and methods Study subjects This study was a dual-center, retrospective, historical control study. The study subjects were 158 inpatients who underwent POCS for suspected extrahepatic MBS due to obstructive jaundice at Union Hospital of Tongji Medical College, Huazhong University of Science and Technology, and The First Affiliated Hospital of Shihezi University from January 2020 to March 2025. Based on our previous research findings that ERCP combined with the modified biopsy forceps improved the diagnostic accuracy and sensitivity for malignant biliary strictures and reduced biopsy difficulty [8], our research group has utilized the modified biopsy forceps for POCS biliary biopsy since July 2023 to further evaluate its diagnostic value when combined with POCS. Therefore, in this study, patients who underwent POCS-guided biopsy using dedicated forceps from January 2020 to June 2023 were enrolled in the control group (hereafter referred to as the ‘POCS group’), and patients who underwent POCS combined with the modified biopsy forceps from July 2023 to March 2025 were enrolled in the experimental group (see Fig. 1 for the detailed patient screening process). Preoperative and postoperative data, including baseline characteristics, laboratory tests, imaging examinations, and pathological results, were collected through the hospital electronic medical record system.Fig. 1Flowchart of patient screening and enrollment Flowchart of patient screening and enrollment The inclusion criteria were as follows: (1) age 18–85 years; (2) first-time performance of POCS; (3) imaging examinations suggesting extrahepatic MBS, including abdominal B-mode ultrasound, computed tomography (CT), magnetic resonance cholangiopancreatography (MRCP), and magnetic resonance imaging (MRI); (4) agreement to undergo POCS and signing of informed consent; and (5) performance of biliary biopsy during POCS. The exclusion criteria were as follows: (1) contraindications to ERCP and POCS; (2) postoperative gastric changes; (3) previous history of malignant tumors; (4) biliary strictures caused by pancreatic cancer; and (5) missing clinical or relevant case data. All procedures were performed in accordance with the Declaration of Helsinki. This study was reviewed and approved by the Ethics Committees of Union Hospital of Tongji Medical College, Huazhong University of Science and Technology, and The First Affiliated Hospital of Shihezi University (Ethics Approval Numbers: [2024] 0156-03 and KJ2024-411-01, respectively). All study subjects signed informed consent forms. Operational methods Modification process of biopsy forceps Currently, pediatric gastroscopy biopsy forceps are commonly used for biliary biopsy in clinical practice. To facilitate easier access to the biliary tract, reduce operational difficulty, and improve the diagnostic accuracy for MBS, we modified a type of pediatric gastroscopy biopsy forceps (Changmei, Jiangsu, China). The original diameter of these biopsy forceps was 2 mm, the opening diameter of the jaw was 6 mm, and the length was 1600 mm. After repeated tests regarding the bending angle and position, we determined that bending the forceps manually by approximately 40°–45° at a distance of 1.2–1.5 cm from the tip of the biopsy forceps was the optimal modification plan [8]. During this study, all modified biopsy forceps were prepared according to this standardized protocol, as shown in Fig. 2. The modified biopsy forceps caused no damage during repeated passage through the biopsy channels of the duodenoscope and peroral cholangioscope, and the tissue sampling effect was good.Fig. 2Comparison of the biopsy forceps before and after modification. A Conventional biopsy forceps; B Modified biopsy forceps Comparison of the biopsy forceps before and after modification. A Conventional biopsy forceps; B Modified biopsy forceps Main equipment Electronic duodenoscope: TJF260V, Olympus Medical Systems, Japan; High-frequency electrosurgical unit: ESG-100, Olympus Medical Systems, Japan; Duodenal sphincterotome: Autotome RX44, Boston Scientific, USA; Guidewire: Dreamwire 0.035 inch × 450 cm, Boston Scientific, USA; Disposable biopsy forceps: FB-20 K-A3, Jiangsu Changmei Medtech Co., Ltd., China; Peroral cholangioscope: MTCDS11001, Micro-Tech (Nanjing) Co., Ltd., China. Operational procedure All ERCP procedures were performed by experienced endoscopists who had completed more than 500 ERCP cases. After confirming that all preoperative preparations were complete, the duodenoscope was advanced to the mid-descending duodenum to locate the duodenal papilla. A sphincterotome with a guidewire was used to cannulate the bile duct, aligning with the duct’s axis. After successful cannulation, bile was aspirated for biliary decompression. Subsequently, cholangiography was performed to confirm the location and length of the stricture (as shown in Fig. 3A). A small endoscopic sphincterotomy of approximately 3–5 mm was performed in the 11–12 o’clock direction. The sphincterotome was withdrawn while maintaining the guidewire in place. Then, the POCS was inserted into the bile duct along the guidewire through the biopsy channel of the duodenoscope. Bile was aspirated to perform a visual impression diagnosis of the stricture site (as shown in Fig. 3B), observing the characteristics of the bile duct epithelium and vascular morphology at the stricture. Subsequently, POCS biopsy forceps were inserted for biopsy. After the biopsy was completed, the POCS was withdrawn. In the experimental group, the modified biopsy forceps were used to perform an additional biopsy at the BS site. The insertion of the biopsy forceps was performed gently, keeping the direction consistent with that of the biliary guidewire. Upon reaching the BS, the biopsy was repeated 3–5 times (as shown in Fig. 3C).Fig. 3Appearances of malignant biliary strictures under fluoroscopy and peroral cholangioscopy, and the fluoroscopic presentation during biopsy with modified biopsy forceps. A Fluoroscopic view of the malignant biliary stricture; B Peroral cholangioscopy view of the malignant biliary stricture; C Biopsy of the target lesion using modified biopsy forceps under fluoroscopy Appearances of malignant biliary strictures under fluoroscopy and peroral cholangioscopy, and the fluoroscopic presentation during biopsy with modified biopsy forceps. A Fluoroscopic view of the malignant biliary stricture; B Peroral cholangioscopy view of the malignant biliary stricture; C Biopsy of the target lesion using modified biopsy forceps under fluoroscopy Data collection, follow-up, and evaluation indicators Data collection content Patient characteristics, including sex, age, location of the biliary stricture, total bilirubin (TBIL), alanine transaminase (ALT), aspartate aminotransferase (AST), and alkaline phosphatase (ALP), were collected. The occurrence of procedure related adverse events, including postoperative pancreatitis, cholangitis, postoperative infection, bleeding, and perforation, was recorded. All adverse events were defined according to international consensus standards [9–12]:Post-ERCP pancreatitis: defined according to Cotton criteria as new or worsening abdominal pain within 24 h after the procedure, together with a serum amylase level more than three times the upper limit of normal, requiring hospitalization or prolongation of planned admission.Bleeding: defined as hematemesis, melena, or a decrease in hemoglobin greater than 2 g/dL within 30 days after the procedure, or the need for endoscopic hemostasis, blood transfusion, or surgical intervention.Cholangitis: defined as a fever above 38.5 °C after the procedure, with right upper quadrant pain, jaundice, or elevated inflammatory markers, requiring antibiotic therapy.Perforation: defined as new retroperitoneal or intra-abdominal free air confirmed by imaging or surgery, or signs of peritonitis.Severity of adverse events was graded according to the ASGE classification: Post-ERCP pancreatitis: defined according to Cotton criteria as new or worsening abdominal pain within 24 h after the procedure, together with a serum amylase level more than three times the upper limit of normal, requiring hospitalization or prolongation of planned admission. Post-ERCP pancreatitis: defined according to Cotton criteria as new or worsening abdominal pain within 24 h after the procedure, together with a serum amylase level more than three times the upper limit of normal, requiring hospitalization or prolongation of planned admission. Bleeding: defined as hematemesis, melena, or a decrease in hemoglobin greater than 2 g/dL within 30 days after the procedure, or the need for endoscopic hemostasis, blood transfusion, or surgical intervention. Bleeding: defined as hematemesis, melena, or a decrease in hemoglobin greater than 2 g/dL within 30 days after the procedure, or the need for endoscopic hemostasis, blood transfusion, or surgical intervention. Cholangitis: defined as a fever above 38.5 °C after the procedure, with right upper quadrant pain, jaundice, or elevated inflammatory markers, requiring antibiotic therapy. Cholangitis: defined as a fever above 38.5 °C after the procedure, with right upper quadrant pain, jaundice, or elevated inflammatory markers, requiring antibiotic therapy. Perforation: defined as new retroperitoneal or intra-abdominal free air confirmed by imaging or surgery, or signs of peritonitis. Perforation: defined as new retroperitoneal or intra-abdominal free air confirmed by imaging or surgery, or signs of peritonitis. Severity of adverse events was graded according to the ASGE classification: Severity of adverse events was graded according to the ASGE classification: ① Mild: hospitalization ≤ 3 days (or prolongation of planned admission ≤ 3 days); ② Moderate: hospitalization 4–10 days; ③ Severe: hospitalization > 10 days, need for ICU care, or need for endoscopic, radiological, or surgical intervention; ④ Death: death directly or indirectly attributable to the procedure. Diagnostic criteria and follow-up The final diagnosis was used as the gold standard for the qualitative diagnosis of biliary strictures. The final diagnosis was established through a comprehensive assessment of pathological diagnosis, imaging examinations, laboratory tests, and follow-up status. All pathological diagnoses were determined by two experienced senior pathologists specializing in pancreaticobiliary diseases, who were blinded to clinical information, endoscopic findings, and group allocation during slide review and re-evaluation.Pathological diagnosis: All included patients underwent POCS-guided biopsy to obtain bile duct tissue specimens. All pathological sections were independently reviewed by two senior pathologists using a blinded method; disagreements were resolved by consensus. Among the pathological results, specimens described as “carcinoma,” “suspicious for carcinoma,” or “high-grade intraepithelial neoplasia” were classified as pathology-positive and diagnosed as malignant biliary stricture (MBS); all other results were classified as pathology-negative and diagnosed as benign biliary stricture (BBS).Follow-up: Patients with negative pathology were followed up for more than 6 months by regular telephone calls or return visits. Follow-up content included imaging examinations, laboratory tests, biopsy pathological results, surgical pathological results for some patients, and clinical outcomes. The follow-up deadline was September 30, 2025.Final diagnosis: For patients with a pathological diagnosis of MBS, the final diagnosis was confirmed as MBS. For patients with a pathological diagnosis of BBS, the final diagnosis was determined based on at least 6 months of clinical follow-up results (imaging progression, surgical pathology, or clinical outcomes). If disease progression occurred after more than 6 months of follow-up, the final diagnosis was MBS; if imaging re-examination during at least 6 months of follow-up showed no progression or improvement of the lesion, the final diagnosis was BBS [8]. Pathological diagnosis: All included patients underwent POCS-guided biopsy to obtain bile duct tissue specimens. All pathological sections were independently reviewed by two senior pathologists using a blinded method; disagreements were resolved by consensus. Among the pathological results, specimens described as “carcinoma,” “suspicious for carcinoma,” or “high-grade intraepithelial neoplasia” were classified as pathology-positive and diagnosed as malignant biliary stricture (MBS); all other results were classified as pathology-negative and diagnosed as benign biliary stricture (BBS). Pathological diagnosis: All included patients underwent POCS-guided biopsy to obtain bile duct tissue specimens. All pathological sections were independently reviewed by two senior pathologists using a blinded method; disagreements were resolved by consensus. Among the pathological results, specimens described as “carcinoma,” “suspicious for carcinoma,” or “high-grade intraepithelial neoplasia” were classified as pathology-positive and diagnosed as malignant biliary stricture (MBS); all other results were classified as pathology-negative and diagnosed as benign biliary stricture (BBS). Follow-up: Patients with negative pathology were followed up for more than 6 months by regular telephone calls or return visits. Follow-up content included imaging examinations, laboratory tests, biopsy pathological results, surgical pathological results for some patients, and clinical outcomes. The follow-up deadline was September 30, 2025. Follow-up: Patients with negative pathology were followed up for more than 6 months by regular telephone calls or return visits. Follow-up content included imaging examinations, laboratory tests, biopsy pathological results, surgical pathological results for some patients, and clinical outcomes. The follow-up deadline was September 30, 2025. Final diagnosis: For patients with a pathological diagnosis of MBS, the final diagnosis was confirmed as MBS. For patients with a pathological diagnosis of BBS, the final diagnosis was determined based on at least 6 months of clinical follow-up results (imaging progression, surgical pathology, or clinical outcomes). If disease progression occurred after more than 6 months of follow-up, the final diagnosis was MBS; if imaging re-examination during at least 6 months of follow-up showed no progression or improvement of the lesion, the final diagnosis was BBS [8]. Final diagnosis: For patients with a pathological diagnosis of MBS, the final diagnosis was confirmed as MBS. For patients with a pathological diagnosis of BBS, the final diagnosis was determined based on at least 6 months of clinical follow-up results (imaging progression, surgical pathology, or clinical outcomes). If disease progression occurred after more than 6 months of follow-up, the final diagnosis was MBS; if imaging re-examination during at least 6 months of follow-up showed no progression or improvement of the lesion, the final diagnosis was BBS [8]. Evaluation indicators Based on the pathological diagnosis and the final diagnosis, the sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (PLR), and negative likelihood ratio (NLR) were compared between the two groups. Additionally, disease prevalence and safety profiles were evaluated. Statistical methods Data analysis was performed using SPSS 27.0, R 4.4.0, and MedCalc software. During the analysis, data were preprocessed to ensure the accuracy, completeness, and consistency of the collected data. Normally distributed continuous variables were expressed as mean ± standard deviation, and intergroup comparisons were performed using the independent-samples t-test. Non-normally distributed continuous variables were expressed as median (IQR), and intergroup comparisons were performed using the Mann–Whitney U test. Categorical data were expressed as frequency (%), and intergroup comparisons were performed using the Chi-square test or Fisher’s exact test, as appropriate. To reduce bias caused by rare events in this study, univariate and multivariate analyses were both performed using Firth’s penalized likelihood logistic regression (bias-reduced logistic regression). Univariate Firth regression analysis was performed on variables that might affect biopsy accuracy, including sex, age, stricture location, TBIL, ALT, AST, ALP, and the type of biopsy forceps. Variables with P < 0.100 in the univariate Firth analysis were then included in the multivariate Firth regression analysis to identify independent influencing factors. The results of the multivariate regression analysis were visualized using a forest plot. For model diagnostic performance evaluation, overall model significance was assessed by the likelihood ratio test; model discrimination was assessed by plotting the receiver operating characteristic (ROC) curve and calculating the area under the curve (AUC) with its 95% confidence interval (CI); model calibration was assessed by the Hosmer–Lemeshow test. All statistical tests were two-sided, and P < 0.05 was considered statistically significant. Results Baseline characteristics A total of 158 patients were included in this study, consisting of 79 patients in the POCS combined with modified biopsy forceps group and 79 patients in the POCS group. There were no statistically significant differences between the two groups regarding sex, age, number of biopsy passes, stricture location, TBIL, ALT, AST, and ALP (P > 0.05), as shown in Table 1.Table 1Patients characteristicsCharacteristicsPOCS + Modified Forceps Group (n = 79)POCS Group (n = 79)Overall (n = 158)P valueSex1.000Male, n (%)41 (51.9)41 (51.9)82 (51.9)Female, n (%)38 (48.1)38 (48.1)76 (48.1)Age (years) (mean ± SD)/median (IQR)65.5 ± 10.665.0 (58.0–70.0)65.0 (58.0–71.0)0.115Number of biopsy passes (times)median (IQR)5.0 (4.0–6.0)6.0 (4.0–6.0)5.0 (4.0–6.0)0.054Stricture location0.439Distal bile duct, n (%)64 (81.0)60 (75.9)124 (78.5)Hilar bile duct, n (%)15 (19.0)19 (24.1)34 (21.5)TBIL (μmol/L)median (IQR)135.0 (77.0–233.5)158.0 (67.0–204.7)141.2 (74.5–213.9)0.907ALT (U/L)median (IQR)110.0 (79.0–174.0)120.0 (56.0–263.0)110.0 (64.5–212.0)0.608AST (U/L)median (IQR)117.0 (75.0–175.0)117.0 (67.0–175.0)117.0 (74.7–175.0)0.464ALP (U/L)median (IQR)342.0 (204.0–573.0)334.0 (200.0–552.0)337.5 (200.0–555.5)0.820Continuous data are expressed as mean ± SD or median and IQR. Categorical data are expressed as number (percentage)IQR Interquartile range, SD Standard deviation, TBIL Total bilirubin, ATL Alanine transaminase, AST Aspartate aminotransferase, ALP Alkaline phosphatase Table 1Patients characteristicsCharacteristicsPOCS + Modified Forceps Group (n = 79)POCS Group (n = 79)Overall (n = 158)P valueSex1.000Male, n (%)41 (51.9)41 (51.9)82 (51.9)Female, n (%)38 (48.1)38 (48.1)76 (48.1)Age (years) (mean ± SD)/median (IQR)65.5 ± 10.665.0 (58.0–70.0)65.0 (58.0–71.0)0.115Number of biopsy passes (times)median (IQR)5.0 (4.0–6.0)6.0 (4.0–6.0)5.0 (4.0–6.0)0.054Stricture location0.439Distal bile duct, n (%)64 (81.0)60 (75.9)124 (78.5)Hilar bile duct, n (%)15 (19.0)19 (24.1)34 (21.5)TBIL (μmol/L)median (IQR)135.0 (77.0–233.5)158.0 (67.0–204.7)141.2 (74.5–213.9)0.907ALT (U/L)median (IQR)110.0 (79.0–174.0)120.0 (56.0–263.0)110.0 (64.5–212.0)0.608AST (U/L)median (IQR)117.0 (75.0–175.0)117.0 (67.0–175.0)117.0 (74.7–175.0)0.464ALP (U/L)median (IQR)342.0 (204.0–573.0)334.0 (200.0–552.0)337.5 (200.0–555.5)0.820Continuous data are expressed as mean ± SD or median and IQR. Categorical data are expressed as number (percentage)IQR Interquartile range, SD Standard deviation, TBIL Total bilirubin, ATL Alanine transaminase, AST Aspartate aminotransferase, ALP Alkaline phosphatase Patients characteristics Number of biopsy passes (times) median (IQR) TBIL (μmol/L) median (IQR) ALT (U/L) median (IQR) AST (U/L) median (IQR) ALP (U/L) median (IQR) Continuous data are expressed as mean ± SD or median and IQR. Categorical data are expressed as number (percentage) IQR Interquartile range, SD Standard deviation, TBIL Total bilirubin, ATL Alanine transaminase, AST Aspartate aminotransferase, ALP Alkaline phosphatase Comparison of diagnostic results In the experimental group, the pathological diagnoses included 69 cases of MBS and 10 cases of BBS, resulting in a final diagnosis of 74 cases of MBS and 5 cases of BBS. In the control group, the pathological diagnoses included 55 cases of MBS and 24 cases of BBS, resulting in a final diagnosis of 77 cases of MBS and 2 cases of BBS. Comparison of the pathological results showed a statistically significant difference between the two groups in the diagnosis of MBS (P = 0.007). However, comparison of the final nature of the biliary strictures showed no statistically significant difference between the two groups in the diagnosis of MBS (P = 0.442), as shown in Tables 2 and 3, respectively. Furthermore, the detection rate of carcinoma in situ (high-grade intraepithelial neoplasia/severe dysplasia) was significantly higher in the experimental group than in the control group (P < 0.001), as shown in Table 4.Table 2Pathological diagnosis of biliary stricturePathological diagnosis, n (%)POCS + Modified Forceps Group (n = 79)POCS Group (n = 79)Overall (n = 158)P valueMBS69 (87.3)55 (69.6)124 (78.5)0.007aBBS10 (12.7)24 (30.4)34 (21.5)POCS Peroral cholangioscopy, MBS Malignant biliary strictures, BBS Benign biliary stricturesaStatistical significanceTable 3Final diagnosis of biliary stricture natureFinal diagnosis, n (%)POCS + Modified forceps group (n = 79)POCS group (n = 79)Overall (n = 158)P valueMBS74 (93.7)77 (97.5)151 (95.6)0.442Carcinoma in situ26 (32.9)5 (6.3)31 (19.6)Cholangiocarcinoma42 (53.2)49 (62.0)91 (57.6)Neuroendocrine carcinoma1 (1.3)0 (0)1 (0.6)IPMN with cancerous change0 (0)1 (1.3)1 (0.6)Malignant biliary stricture confirmed by follow-up5 (6.3)22 (27.8)27 (17.1)BBS5 (6.3)2 (2.5)7 (4.4)POCS Peroral cholangioscopy; MBS Malignant biliary strictures; BBS Benign biliary strictures; IPMN Intraductal papillary mucinous neoplasmTable 4Detection results of carcinoma in situ (high-grade intraepithelial neoplasia/severe dysplasia)Final (pathological) diagnosis, n (%)POCS + Modified forceps group (n = 79)POCS group (n = 79)Overall (n = 158)P valueCarcinoma in situ (high-grade intraepithelial neoplasia/severe dysplasia)26 (32.9)5 (6.3)31 (19.6) < 0.001aNon-carcinoma in situ (non-high-grade intraepithelial neoplasia/severe dysplasia)53 (67.1)74 (93.7)127 (80.4)POCS Peroral cholangioscopyaStatistical significance Table 2Pathological diagnosis of biliary stricturePathological diagnosis, n (%)POCS + Modified Forceps Group (n = 79)POCS Group (n = 79)Overall (n = 158)P valueMBS69 (87.3)55 (69.6)124 (78.5)0.007aBBS10 (12.7)24 (30.4)34 (21.5)POCS Peroral cholangioscopy, MBS Malignant biliary strictures, BBS Benign biliary stricturesaStatistical significance Pathological diagnosis of biliary stricture POCS Peroral cholangioscopy, MBS Malignant biliary strictures, BBS Benign biliary strictures aStatistical significance Table 3Final diagnosis of biliary stricture natureFinal diagnosis, n (%)POCS + Modified forceps group (n = 79)POCS group (n = 79)Overall (n = 158)P valueMBS74 (93.7)77 (97.5)151 (95.6)0.442Carcinoma in situ26 (32.9)5 (6.3)31 (19.6)Cholangiocarcinoma42 (53.2)49 (62.0)91 (57.6)Neuroendocrine carcinoma1 (1.3)0 (0)1 (0.6)IPMN with cancerous change0 (0)1 (1.3)1 (0.6)Malignant biliary stricture confirmed by follow-up5 (6.3)22 (27.8)27 (17.1)BBS5 (6.3)2 (2.5)7 (4.4)POCS Peroral cholangioscopy; MBS Malignant biliary strictures; BBS Benign biliary strictures; IPMN Intraductal papillary mucinous neoplasm Final diagnosis of biliary stricture nature POCS Peroral cholangioscopy; MBS Malignant biliary strictures; BBS Benign biliary strictures; IPMN Intraductal papillary mucinous neoplasm Table 4Detection results of carcinoma in situ (high-grade intraepithelial neoplasia/severe dysplasia)Final (pathological) diagnosis, n (%)POCS + Modified forceps group (n = 79)POCS group (n = 79)Overall (n = 158)P valueCarcinoma in situ (high-grade intraepithelial neoplasia/severe dysplasia)26 (32.9)5 (6.3)31 (19.6) < 0.001aNon-carcinoma in situ (non-high-grade intraepithelial neoplasia/severe dysplasia)53 (67.1)74 (93.7)127 (80.4)POCS Peroral cholangioscopyaStatistical significance Detection results of carcinoma in situ (high-grade intraepithelial neoplasia/severe dysplasia) POCS Peroral cholangioscopy aStatistical significance In this study, all 158 patients successfully underwent biopsy of the lesion tissue. Pathological diagnosis identified MBS in 124 patients (78.5%) and BBS in 34 patients (21.5%). Among the 34 patients with BBS on pathological diagnosis, 27 (79.4%) were ultimately diagnosed with MBS after follow-up (5 in the experimental group, 22 in the control group), and 7 (20.6%) were ultimately diagnosed as BBS (5 in the experimental group, 2 in the control group). As shown in Table 5, the overall agreement rate between the pathological diagnosis and the final diagnosis was 82.9% (131/158). Notably, the agreement rate was significantly higher in the experimental group than in the control group (93.7% vs. 72.2%, P < 0.001).Table 5Comparison of pathological and final diagnoses between the two groupsDiagnosis, n (%)POCS + Modified Forceps Group (n = 79)POCS Group (n = 79)Overall (n = 158)P valueConsistent74 (93.7)57 (72.2)131 (82.9) < 0.001aInconsistent5 (6.3)22 (27.8)27 (17.1)Consistent was defined as pathological diagnosis consistent with the final diagnosis; inconsistent was defined as pathological diagnosis of BBS but final diagnosis of MBS (false negative)aStatistical significance Table 5Comparison of pathological and final diagnoses between the two groupsDiagnosis, n (%)POCS + Modified Forceps Group (n = 79)POCS Group (n = 79)Overall (n = 158)P valueConsistent74 (93.7)57 (72.2)131 (82.9) < 0.001aInconsistent5 (6.3)22 (27.8)27 (17.1)Consistent was defined as pathological diagnosis consistent with the final diagnosis; inconsistent was defined as pathological diagnosis of BBS but final diagnosis of MBS (false negative)aStatistical significance Comparison of pathological and final diagnoses between the two groups Consistent was defined as pathological diagnosis consistent with the final diagnosis; inconsistent was defined as pathological diagnosis of BBS but final diagnosis of MBS (false negative) aStatistical significance Comparison of diagnostic efficacy and safety As shown in Table 6, the accuracy (93.7% vs 72.2%, P < 0.001), sensitivity (93.2% vs 71.4%, P < 0.001), and NPV (50.0% vs 8.3%, P = 0.014) of the experimental group were significantly superior to those of the control group. The prevalence of MBS was 93.7% (95% CI: 85.8%–97.9%) in the experimental group and 97.5% (95% CI: 91.2%–99.7%) in the control group, with no significant difference between the two groups (P = 0.442). The negative likelihood ratio was 0.068 (95% CI: 0.029–0.158) in the experimental group and 0.286 (95% CI: 0.201–0.407) in the control group. Given that specificity was 100% in both groups, the positive likelihood ratio was theoretically infinite.Table 6Comparison of diagnostic efficacy and incidence of adverse eventsPOCS + Modified forceps group (n = 79)POCS group (n = 79)Overall (n = 158)P valueAccuracy (%) (95%CI)93.7 (85.8–97.9)72.2 (60.9–81.7)82.9 (76.1–88.4) < 0.001aSensitivity (%) (95%CI)93.2 (84.9–97.7)71.4 (60.0–81.2)82.1 (75.0–87.9) < 0.001aSpecificity (%) (95%CI)100.0 (47.8–100)100.0 (15.8–100)100.0 (59.0–100)1.000PPV, percentage (%) (95% CI)100.0 (94.8–100)100.0 (93.5–100)100.0 (97.1–100)1.000NPV, percentage (%) (95% CI)50.0 (30.0–70.0)8.3 (6.0–11.5)20.6 (15.6–26.8)0.014aPLR, percentage (%) (95% CI) > 10 > 10 > 10NLR, percentage (%) (95% CI)0.068 (0.029–0.158)0.286 (0.201–0.407)0.179 (0.127–0.252)Prevalence, percentage (%) (95% CI)93.7 (85.8–97.9)97.5 (91.2–99.7)95.6 (91.1–98.2)0.442Adverse events (Total), n (%)6 (7.6)7 (8.9)13 (8.2)1.000Pancreatitis, n (%)5 (6.3)5 (6.3)10 (6.3)Bleeding, n (%)1 (1.3)2 (2.5)3 (1.9)Perforation, n (%)0 (0)0 (0)0 (0)Death, n (%)0 (0)0 (0)0 (0)POCS Peroral cholangioscopy, CI Confidence interval, PPV Positive predictive value, NPV Negative predictive value, PLR Positive likelihood ratio, NLR Negative likelihood ratioaStatistical significance Table 6Comparison of diagnostic efficacy and incidence of adverse eventsPOCS + Modified forceps group (n = 79)POCS group (n = 79)Overall (n = 158)P valueAccuracy (%) (95%CI)93.7 (85.8–97.9)72.2 (60.9–81.7)82.9 (76.1–88.4) < 0.001aSensitivity (%) (95%CI)93.2 (84.9–97.7)71.4 (60.0–81.2)82.1 (75.0–87.9) < 0.001aSpecificity (%) (95%CI)100.0 (47.8–100)100.0 (15.8–100)100.0 (59.0–100)1.000PPV, percentage (%) (95% CI)100.0 (94.8–100)100.0 (93.5–100)100.0 (97.1–100)1.000NPV, percentage (%) (95% CI)50.0 (30.0–70.0)8.3 (6.0–11.5)20.6 (15.6–26.8)0.014aPLR, percentage (%) (95% CI) > 10 > 10 > 10NLR, percentage (%) (95% CI)0.068 (0.029–0.158)0.286 (0.201–0.407)0.179 (0.127–0.252)Prevalence, percentage (%) (95% CI)93.7 (85.8–97.9)97.5 (91.2–99.7)95.6 (91.1–98.2)0.442Adverse events (Total), n (%)6 (7.6)7 (8.9)13 (8.2)1.000Pancreatitis, n (%)5 (6.3)5 (6.3)10 (6.3)Bleeding, n (%)1 (1.3)2 (2.5)3 (1.9)Perforation, n (%)0 (0)0 (0)0 (0)Death, n (%)0 (0)0 (0)0 (0)POCS Peroral cholangioscopy, CI Confidence interval, PPV Positive predictive value, NPV Negative predictive value, PLR Positive likelihood ratio, NLR Negative likelihood ratioaStatistical significance Comparison of diagnostic efficacy and incidence of adverse events POCS Peroral cholangioscopy, CI Confidence interval, PPV Positive predictive value, NPV Negative predictive value, PLR Positive likelihood ratio, NLR Negative likelihood ratio aStatistical significance In addition, a total of 13 adverse events occurred in the two groups (6 in the experimental group, 7 in the control group), and there was no significant difference in the incidence of adverse events between the two groups (7.6% vs. 8.9%, P = 1.000). Specifically, post-ERCP pancreatitis occurred in 10 patients (6.3%): 5 in the experimental group and 5 in the control group. All cases were mild and resolved within 3 days with conservative treatment including fasting, intravenous fluids, and analgesics; no patient developed pancreatic necrosis or required ICU care. Post-procedure bleeding occurred in 3 patients (1.9%): 1 in the experimental group and 2 in the control group. All cases were also mild. Two patients had no further bleeding after conservative treatment with fluids and hemostatic agents, and one patient had spontaneous resolution with observation; all resolved within 3 days, and none required blood transfusion or surgery. Univariate and multivariate binary logistic regression analysis of biliary biopsy diagnostic accuracy Univariate and multivariate Firth regression analyses were performed to identify factors influencing the diagnostic accuracy of biliary biopsy. In the univariate Firth regression analysis, biopsy forceps type (OR = 5.300, 95%CI: 2.089–15.781, P < 0.001) and ALP (OR = 0.998, 95%CI: 0.997–1.000, P = 0.024) were statistically significant. Variables with P < 0.100 in the univariate analysis were then included in the multivariate Firth regression analysis, which showed that biopsy forceps type (OR = 6.641, 95% CI: 2.491–20.870, P < 0.001), number of biopsy passes (OR = 1.377, 95% CI: 1.039–1.864, P = 0.026), and ALP (OR = 0.998, 95% CI: 0.997–1.000, P = 0.019) were independent influencing factors for the diagnostic accuracy of biliary biopsy. The results of the multivariate regression analysis were visualized using a forest plot (Fig. 4). Other variables, including sex, age, stricture location, TBIL, ALT, and AST, did not show statistical significance in either the univariate or multivariate analyses (P > 0.05), as shown in Table 7.Fig. 4Forest plot of multivariate Firth regression analysis for factors influencing the diagnostic accuracy of biliary biopsyTable 7Univariate and multivariate Firth regression analysisVariableUnivariate analysisMultivariate analysisOR95%CIP valueOR95%CIP valueBiopsy forceps5.300(2.089–15.781) < 0.001a6.641(2.491–20.870) < 0.001aNumber of biopsy passes1.265(0.976–1.673)0.0771.377(1.039–1.864)0.026sex0.491(0.201–1.129)0.0950.519(0.200–1.282)0.157age (years)1.016(0.978–1.055)0.399Stricture location0.849(0.280–2.224)0.751TBIL (μmol/L)0.998(0.994–1.002)0.359ALT (U/L)1.000(0.997–1.003)0.940AST (U/L)1.000(0.997–1.004)0.950ALP (U/L)0.998(0.997–1.000)0.0240.998(0.997–1.000)0.019OR Odds ratio, CI Confidence interval, TBIL Total bilirubin, ATL Alanine transaminase, AST Aspartate aminotransferase, ALP Alkaline phosphataseaStatistical significance Forest plot of multivariate Firth regression analysis for factors influencing the diagnostic accuracy of biliary biopsy Table 7Univariate and multivariate Firth regression analysisVariableUnivariate analysisMultivariate analysisOR95%CIP valueOR95%CIP valueBiopsy forceps5.300(2.089–15.781) < 0.001a6.641(2.491–20.870) < 0.001aNumber of biopsy passes1.265(0.976–1.673)0.0771.377(1.039–1.864)0.026sex0.491(0.201–1.129)0.0950.519(0.200–1.282)0.157age (years)1.016(0.978–1.055)0.399Stricture location0.849(0.280–2.224)0.751TBIL (μmol/L)0.998(0.994–1.002)0.359ALT (U/L)1.000(0.997–1.003)0.940AST (U/L)1.000(0.997–1.004)0.950ALP (U/L)0.998(0.997–1.000)0.0240.998(0.997–1.000)0.019OR Odds ratio, CI Confidence interval, TBIL Total bilirubin, ATL Alanine transaminase, AST Aspartate aminotransferase, ALP Alkaline phosphataseaStatistical significance Univariate and multivariate Firth regression analysis OR Odds ratio, CI Confidence interval, TBIL Total bilirubin, ATL Alanine transaminase, AST Aspartate aminotransferase, ALP Alkaline phosphatase aStatistical significance Regarding the diagnostic performance evaluation of the multivariate Firth regression model, the likelihood ratio test (χ2 = 25.79, df = 4, P < 0.001) indicated that the model was overall statistically significant. The discriminative ability of the model was evaluated by plotting the ROC curve, which showed an AUC of 0.793 (95% CI: 0.706–0.879), indicating good discriminative ability (Fig. 5). The calibration of the model was evaluated by the Hosmer–Lemeshow goodness-of-fit test, which showed that the model fit was good (χ2 = 2.53, df = 8, P = 0.960), indicating no significant difference between the predicted values and the actual observed values.Fig. 5ROC curve of the multivariate Firth regression model ROC curve of the multivariate Firth regression model Subgroup analysis by location of biliary stricture To further evaluate the diagnostic performance of POCS combined with the modified biopsy forceps in different locations of biliary stricture, we performed a subgroup analysis based on the location of the biliary stricture. The specific diagnostic performance indicators are shown in Table 8. In the hilar biliary stricture subgroup (n = 34), the experimental group achieved 100% accuracy and sensitivity, compared to 73.7% and 72.2% in the control group, respectively. Although this difference did not reach formal statistical significance (100% vs. 73.7%, P = 0.053), it showed a strong trend toward improved accuracy, suggesting a potential clinical advantage in high-difficulty sampling locations. In the distal biliary stricture subgroup (n = 124), the experimental group achieved 92.2% accuracy and 91.5% sensitivity, while the control group had 71.7% and 71.2%, respectively. The difference in accuracy was statistically significant (92.2% vs. 71.7%, P = 0.004). The results of each subgroup analysis showed that the sensitivity of the experimental group was better than that of the control group. Except for the hilar subgroup of the experimental group where specificity could not be calculated because there were no benign biliary stricture cases, the specificity and PPV in the other subgroups remained at high levels. Limited by the small number of benign biliary stricture cases, the NPV in the subgroups that could be calculated was generally low.Table 8Comparison of diagnostic performance between the two groups based on stricture locationsStricture locationGroupnAccuracy (%)(95% CI)P valueaSensitivity (%)(95% CI)Specificity (%)(95% CI)PPV (%)(95% CI)NPV (%)(95% CI)Hilar bile ductControl group1973.7(48.8–90.9)0.05372.2(46.5–90.3)100(2.5–100)100(75.3–100)16.7(0.4–64.1)Experimentl group15100(78.2–100)100(78.2–100)NA100(78.2–100)NADistal bile ductControl group6071.7(58.6–82.5)0.00471.2(57.9–82.2)100(2.5–100)100(91.6–100)5.6(0.1–27.3)Experimentl group6492.2(82.7–97.4)91.5(81.3–97.2)100(47.8–100)100(93.4–100)50(18.7–81.3)NA not applicable (due to the absence of negative cases), CI confidence interval, PPV positive predictive value, NPV negative predictive valueaP value indicates the comparison of diagnostic accuracy between the two groups in each subgroup Table 8Comparison of diagnostic performance between the two groups based on stricture locationsStricture locationGroupnAccuracy (%)(95% CI)P valueaSensitivity (%)(95% CI)Specificity (%)(95% CI)PPV (%)(95% CI)NPV (%)(95% CI)Hilar bile ductControl group1973.7(48.8–90.9)0.05372.2(46.5–90.3)100(2.5–100)100(75.3–100)16.7(0.4–64.1)Experimentl group15100(78.2–100)100(78.2–100)NA100(78.2–100)NADistal bile ductControl group6071.7(58.6–82.5)0.00471.2(57.9–82.2)100(2.5–100)100(91.6–100)5.6(0.1–27.3)Experimentl group6492.2(82.7–97.4)91.5(81.3–97.2)100(47.8–100)100(93.4–100)50(18.7–81.3)NA not applicable (due to the absence of negative cases), CI confidence interval, PPV positive predictive value, NPV negative predictive valueaP value indicates the comparison of diagnostic accuracy between the two groups in each subgroup Comparison of diagnostic performance between the two groups based on stricture locations 73.7 (48.8–90.9) 72.2 (46.5–90.3) 100 (2.5–100) 100 (75.3–100) 16.7 (0.4–64.1) 100 (78.2–100) 100 (78.2–100) 100 (78.2–100) 71.7 (58.6–82.5) 71.2 (57.9–82.2) 100 (2.5–100) 100 (91.6–100) 5.6 (0.1–27.3) 92.2 (82.7–97.4) 91.5 (81.3–97.2) 100 (47.8–100) 100 (93.4–100) 50 (18.7–81.3) NA not applicable (due to the absence of negative cases), CI confidence interval, PPV positive predictive value, NPV negative predictive value aP value indicates the comparison of diagnostic accuracy between the two groups in each subgroup Discussion Biliary stricture (BS) is one of the common diseases of the digestive system. The nature of BS can be initially assessed by combining clinical symptoms, serum tumor markers, and imaging examinations such as ultrasound, CT, and MRCP. When a patient with BS is highly suspected of having malignancy based on clinical symptoms, laboratory findings, and imaging, but lacks a pathological diagnosis, it is clinically referred to as indeterminate biliary stricture [13]. Although significant progress has been made in research on serum tumor markers and imaging diagnostic techniques, pathological results remain the gold standard for diagnosis. Therefore, biliary biopsy is crucial for differentiating between benign and malignant biliary strictures, for the preoperative evaluation of cholangiocarcinoma, and for determining the optimal intervention [14]. Common biopsy methods currently used for IBS include biliary cytology brushings via ERCP, ERCP tissue biopsy, POCS visual diagnosis and targeted biopsy, and endoscopic ultrasonography-guided fine needle aspiration/biopsy (EUS-FNA/B), among others. However, the most commonly used method is ERCP tissue biopsy. Due to the difficulty of biliary biopsy, the sensitivity of traditional ERCP biopsy has always been low; the sensitivities of biliary cytology brushings and tissue biopsy via ERCP are approximately 45% and 48%, respectively, and even when combining these two techniques, sensitivity only increases to 59.4% [15]. In the diagnosis of IBS, some studies indicate that the diagnostic sensitivity of POCS is similar to that of fluoroscopy-guided biopsy, at 54% and 64%, respectively. When the two methods are combined, sensitivity can increase to 80%. However, the insufficient diagnostic sensitivity of POCS-guided biopsy using dedicated forceps is mainly due to the small size and opening of the biopsy forceps, resulting in small sample volume. Furthermore, if the BS is severe, the POCS cannot be pushed past the stricture to the target lesion [16]. In contrast, EUS-FNA has a sensitivity of 75%–94% and an accuracy of 79%–84% [17]. In summary, the sensitivity of POCS biopsy in diagnosing IBS is higher than that of conventional ERCP biliary biopsy, but the diagnostic efficacy of both methods varies significantly among different studies, likely related to the etiology of BS, location of the stricture, variability in biopsy tissue sample size, and differences in operator technique. Although studies have attempted to improve and develop operational accessories for POCS or combine it with other techniques to increase diagnostic sensitivity, these methods are not yet fully mature. Furthermore, the clinical need to obtain sufficient tissue specimens from biliary lesions for pathological analysis has become increasingly prominent in recent years. Therefore, developing tissue acquisition methods that enable targeted biopsy is of significant importance for the precise diagnosis of the nature of biliary strictures [18]. Given the low diagnostic sensitivity of the conventional biopsy forceps previously used, the main reasons include the small sample volume, the difficulty of inserting the forceps into the bile duct, the challenge in accurate positioning on the lesion, and the inability to perform repeat biopsies easily. This is significant considering that biopsy specimen quality is related to diagnostic sensitivity [19], and at least three biopsies are required to achieve a correct diagnosis in 90% of BS patients [20]. Therefore, this study modified the pediatric gastroscopy biopsy forceps commonly used in clinical practice. By bending the tip of the forceps to a specific curvature, the difficulty of biopsy was reduced, allowing the forceps to enter the bile duct more easily. Furthermore, the modified forceps can enter the bile duct quickly without guidewire guidance and can be repeatedly advanced based on biopsy needs. Our preliminary study found that using modified biopsy forceps during ERCP significantly improved the accuracy (84.4% vs 53.1%, P = 0.003) and sensitivity (83.3% vs 50.0%, P = 0.006) for diagnosing MBS compared to conventional forceps, but diagnostic capability could still be further improved [8]. With the development and widespread use of POCS, biliary examination has entered the era of direct visualization, making POCS increasingly important in the diagnosis of biliary strictures. Hence, this study aimed to improve the diagnostic value for MBS in IBS patients by using POCS combined with modified biopsy forceps. Our results demonstrate that the diagnostic accuracy (93.7% vs 72.2%, P < 0.001) and sensitivity (93.2% vs 71.4%, P < 0.001) of POCS combined with modified biopsy forceps were significantly superior to POCS-guided biopsy using dedicated forceps. The significant improvement in diagnostic accuracy and sensitivity seen with the modified biopsy forceps, and the further enhancement when combined with POCS compared to ERCP, can be attributed mainly to two factors: first, the modified biopsy forceps reduce the difficulty of entering the bile duct and allows multiple biopsy passes; when combined with POCS, they can precisely target the stricture under direct visualization and improve the quality of each biopsy pass; second, the visualization feature of POCS allows direct observation of the stricture, enabling precise biopsy and fewer missed lesions, thus contributing to improved diagnostic accuracy and sensitivity. In addition, in this study there was no significant difference in the number of biopsy passes between the two groups (5.0 vs. 6.0, P = 0.054), and the control group had slightly more passes, yet its diagnostic sensitivity and accuracy were significantly lower than those of the experimental group. These findings indicate that the improvement in diagnostic performance in this study is not related to the number of biopsy passes and should be attributed to the design advantage of the modified biopsy forceps itself. The overall prevalence of MBS in the enrolled patients was 95.6%, and there was no significant difference in prevalence between the experimental group and the control group (93.7% vs. 97.5%, P = 0.442). Previous literature reports that MBS accounts for about 76%–85% of all biliary strictures [21]. However, the high prevalence in this study is mainly because both centers are regional referral centers for complex diseases, and the patients treated were mainly those with a high suspicion of MBS. This is consistent with the original intention of this study to evaluate the diagnostic performance of POCS combined with the modified biopsy forceps, but it also introduces the potential problem of spectrum bias. Considering that spectrum bias may affect the NPV in this study, we further calculated the likelihood ratio (LR), which is not affected by disease prevalence, to more objectively evaluate the diagnostic performance of POCS combined with the modified biopsy forceps. The results showed that the NLR in the experimental group (0.068, 95% CI: 0.029–0.158) was significantly better than that in the control group (0.286, 95% CI: 0.201–0.407), and the 95% CIs did not overlap. This indicates that POCS combined with the modified biopsy forceps has a stronger ability to rule out MBS, and this result is not affected by the high prevalence in our study population and better reflects the true diagnostic value of POCS combined with the modified biopsy forceps. Moreover, there was no significant difference in prevalence between the two groups, indicating that the disease spectrum was comparable between the two groups and that the difference in diagnostic performance was not caused by an uneven distribution of the disease spectrum. In this study, all 158 patients successfully underwent biopsy of the lesion tissue. After comprehensive assessment, we found a significant discrepancy between pathological diagnosis and final diagnosis in some patients: 34 patients (21.5%) had a benign pathological diagnosis, but 27 of them (79.4%) were confirmed as malignant after follow-up. This phenomenon reveals that false-negative results caused by sampling error are a key challenge in the current diagnosis of biliary strictures. Further comparison of the false-negative rates between the two groups showed that the false-negative rate was significantly higher in the control group than in the experimental group (27.8% vs. 6.3%). This indicates that POCS combined with the modified biopsy forceps reduced the risk of false-negative events by 77.3%, which is the most direct reflection of its improved diagnostic performance. From a clinical perspective, this improvement means that for every four malignant biliary strictures that might have been missed, three can be correctly diagnosed using POCS combined with the modified biopsy forceps. We believe the false-negative rate was reduced for the following reasons: the dedicated POCS biopsy forceps have a small jaw and limited penetration, making it difficult to obtain tissue from sufficient depth, leading to false-negative results. In contrast, the pre-curved tip of the modified biopsy forceps (40–45° at 1.2–1.5 cm from the tip) allows them to enter the bile duct at a more favorable angle, precisely target the stricture under direct POCS visualization, and achieve deeper submucosal tissue acquisition [8]. This physical advantage directly translates into higher diagnostic sensitivity (93.2% vs. 71.4%) and effectively reduces the false-negative rate. Studies have shown that the bending angle and tip design of biopsy forceps are key factors affecting tissue acquisition depth, and an appropriate bend can significantly improve the ability to penetrate the bile duct wall, thereby increasing biopsy specimen adequacy and quality [18]. Our results also indicate that the detection rate of carcinoma in situ (high-grade intraepithelial neoplasia/severe dysplasia) was significantly higher in the POCS combined with modified biopsy forceps group than in the POCS group (32.9% vs 6.3%, P < 0.001). Furthermore, the NPV of the POCS combined with modified biopsy forceps group (50.0% vs 8.3%, P = 0.014) was significantly superior to the POCS group. This demonstrates that POCS combined with modified biopsy forceps has an advantage over POCS alone in detecting precancerous biliary lesions and differentiating IBS. This is primarily because masses protruding into the lumen and dilated, tortuous mucosal vessels under POCS direct visualization usually suggest MBS [22], while BBS typically presents with smooth internal biliary wall mucosa, normal and clear vascular patterns, a fine granular appearance, and no papillary tissue or visible masses [23]. Biliary Intraepithelial Neoplasia is a non-invasive flat or (micro)papillary lesion where biliary epithelium grows into the duct lumen [24]. Because it does not form a mass or cause biliary obstruction, it is typically asymptomatic and undetectable by traditional imaging; the final diagnosis must be based on pathology. Adequate tissue sampling of BilIN, regardless of its grade, is crucial to avoid missing more severe lesions. Therefore, POCS combined with modified biopsy forceps allows for targeted and sufficient sampling of IBS lesions under direct vision, yielding more representative tissue specimens. This enhances the detection rate of precancerous lesions like BilIN and ultimately reduces false-negative events caused by insufficient sampling. Despite the ability of POCS combined with modified biopsy forceps to achieve direct visualization and targeted biopsy of biliary strictures, false-negative events persisted. We attribute this primarily to the fact that the majority of these malignant biliary strictures were infiltrative cholangiocarcinomas characterized by a desmoplastic reaction or submucosal growth pattern. Consequently, the overlying mucosa may exhibit only inflammatory changes even under direct visualization, thereby preventing the modified forceps from accessing the deeper malignant infiltrative tissue. Furthermore, anatomical constraints such as severe biliary stricture or acute ductal angulation restricted the ability of the modified forceps to obtain adequate pathological specimens despite visual guidance. Therefore, the establishment of a final diagnosis necessitates a comprehensive assessment integrating radiological imaging, laboratory findings, and longitudinal clinical follow-up. In conclusion, POCS combined with modified biopsy forceps compensates for the subjectivity and limitations inherent in relying solely on POCS visual diagnosis. By ensuring sufficient tissue acquisition, it improves the diagnostic ability for MBS and precancerous lesions, reducing clinical diagnostic uncertainty, with the goal of providing more accurate diagnoses and personalized treatment plans for MBS patients. Common adverse events following POCS include post-procedure pancreatitis, self-limiting abdominal pain, gastrointestinal bleeding, cholangitis, and perforation [25]. Since endoscopic sphincterotomy must be performed to insert the POCS into the common bile duct along the guidewire for biopsy, the incidence of adverse events such as post-procedure bleeding, cholangitis, and perforation is reported to increase to 7.0% [26]. Post-procedure pancreatitis is the most common adverse event of ERCP, with a reported incidence of 3.5% to 9.7%, but the incidence rate in patients undergoing biliary biopsy is less frequently reported [27]. Given the risk of biliary infection, existing evidence suggests that routine prophylactic antibiotic use before POCS may reduce the incidence of cholangitis [28]. In this study, the incidence of adverse events between the two groups was not statistically significant (7.6% vs 8.9%, P = 1.000). No severe adverse events occurred in either group, and all were cured after conservative medical treatment, suggesting that POCS combined with modified biopsy forceps can safely and effectively be used for biliary biopsy. Furthermore, the forceps modification method is simple and did not cause damage to the duodenoscope during use, suggesting that the modified biopsy forceps should be widely applied in clinical practice. We also used univariate and multivariate Firth regression analyses to explore the factors influencing the diagnostic accuracy of biliary biopsy. The use of the modified biopsy forceps (OR = 6.641, 95% CI: 2.491–20.870, P < 0.001) was a significant independent factor for diagnostic accuracy, indicating that the modified biopsy forceps can improve the diagnostic accuracy of biliary biopsy. The number of biopsy passes (OR = 1.377, 95% CI: 1.039–1.864, P = 0.026) was also an independent factor, suggesting that increasing the number of biopsy passes can improve diagnostic accuracy. ALP (OR = 0.998, 95% CI: 0.997–1.000, P = 0.019) was another independent factor, but its OR was very close to 1, indicating that although the correlation with diagnostic accuracy was statistically significant, its actual clinical significance may be limited. The biological mechanism underlying this association may be that serum ALP is a product of cholangiocyte response to cholestasis [29], and MBS typically causes more severe and persistent cholestasis, leading to increased ALP synthesis and impaired excretion, thus significantly elevating serum ALP levels [30]. Therefore, a higher ALP level may, to some extent, increase the suspicion of MBS. However, serum ALP levels overlap widely between benign and malignant biliary strictures and are affected by many factors such as the degree of obstruction, disease duration, and hepatic functional reserve; thus, ALP is not a specific marker for malignant biliary strictures. Moreover, baseline ALP levels did not differ significantly between the two groups, indicating that ALP is not a confounding factor for the difference in diagnostic performance between the groups. Therefore, the diagnostic advantage of the modified biopsy forceps cannot be attributed to differences in ALP levels between groups. In addition, the diagnostic performance evaluation of the multivariate Firth regression model showed good discrimination and calibration, and regardless of whether other independent factors were included in the model, the modified biopsy forceps remained the strongest independent factor for diagnostic accuracy. This study found no significant association between sex, age, stricture location, TBIL, ALT, or AST and the diagnostic accuracy of biliary biopsy (all P > 0.05). These factors did not show an influence in this study, but this may need further verification with a larger sample size. To further clarify the diagnostic value of POCS combined with the modified biopsy forceps in different locations of biliary stricture, we performed a subgroup analysis based on the location of the stricture. The results showed that in both the hilar and distal biliary stricture subgroups, the experimental group had better diagnostic accuracy and sensitivity than the control group. In the distal stricture subgroup, the difference in accuracy was statistically significant (P = 0.004); in the hilar stricture subgroup, the difference in accuracy approached statistical significance (P = 0.053). These findings suggest that POCS combined with the modified biopsy forceps has good diagnostic value in different locations of biliary stricture, with a more obvious advantage in distal strictures. However, because the number of patients with hilar strictures was small in this study (15 in the experimental group, 19 in the control group), this may affect the precise evaluation of the diagnostic performance of POCS combined with the modified biopsy forceps for hilar strictures. Future studies with larger sample sizes are needed to further validate its diagnostic value. Second, in each subgroup, the sensitivity of the experimental group was higher than that of the control group, while specificity and PPV remained at high levels in both groups. This suggests that POCS combined with the modified biopsy forceps improves the detection rate of malignant biliary strictures without significantly increasing the risk of false-positive events, indicating good clinical usefulness. The low NPV in some subgroups was mainly due to the small number of benign biliary stricture cases included. In addition, because biopsy of hilar biliary strictures is generally more technically challenging, the pre-curved tip design of the modified biopsy forceps may be more favorable for targeted biopsy at this location. The subgroup analysis results also showed that the experimental group achieved 100% accuracy and sensitivity, while the control group had only 73.7% and 72.2%, respectively. This finding further supports the clinical value of POCS combined with the modified biopsy forceps and suggests that its pre-curved tip design may effectively overcome the technical challenges of hilar stricture biopsy. Considering that there was no significant difference in the proportion of hilar strictures between the two groups, this indicates that the distribution of stricture location did not bias the study conclusions. In summary, POCS combined with the modified biopsy forceps shows good diagnostic performance for both hilar and distal biliary strictures, can provide more reliable tissue specimens for biopsy at different locations, and is worthy of further clinical application. However, this study has several limitations. First, this was a retrospective study, and the retrospective design is an inherent limitation. Although we used rigorous statistical methods and tried to control for potential confounding factors as much as possible, the biases inherent to retrospective studies cannot be completely eliminated. Therefore, the level of evidence of our conclusions is limited by the retrospective design. In addition, the sample size of this study is relatively small. Future multicenter prospective randomized controlled trials with larger sample sizes are needed to further validate the diagnostic value of the modified biopsy forceps. Second, because the modified biopsy forceps lack a guidewire, biopsy of intrahepatic biliary lesions is very difficult, and further research and design improvements are needed. Meanwhile, all modified biopsy forceps in this study were prepared according to the standardized protocol. However, due to the retrospective study design, we did not systematically record the exact bending angle and bending point distance for each modified biopsy forceps. Future prospective studies using tools such as a goniometer should precisely measure the bending parameters of the modified biopsy forceps to further explore the dose–response relationship between bending angle/distance and diagnostic performance. Third, this study used a retrospective analysis and a historical control study design, which may introduce non-contemporaneous control bias. However, the baseline characteristics of the two groups were comparable, and multivariate regression analysis still showed that the modified biopsy forceps were an independent factor for diagnostic accuracy after adjusting for potential confounders. This suggests that the improvement in diagnostic performance is mainly due to the improvement of the biopsy forceps themselves, rather than non-contemporaneous control bias or other confounding factors. Nevertheless, future prospective multicenter randomized controlled trials are still needed to further validate our findings. Finally, because of the retrospective study design, we could not systematically collect or record some parameters, including exact procedure time, cost-effectiveness analysis, quantitative assessment of biopsy sample size and specimen adequacy, and physicochemical and cytological examination of bile samples. Although these parameters could provide additional information on the clinical usefulness and diagnostic value of POCS combined with the modified biopsy forceps, our results still show that POCS combined with the modified biopsy forceps significantly improves diagnostic performance. Future prospective studies should standardize the collection of these relevant parameters to provide stronger evidence for the clinical application of the modified biopsy forceps. In summary, POCS combined with modified biopsy forceps improved the diagnostic accuracy and sensitivity for MBS, reduced the difficulty of biliary biopsy, and demonstrated good safety.