The identification of benign and malignant biliary strictures is critical to patient treatment and prognosis. Early diagnosis of benign biliary strictures can reduce unnecessary surgery and costs, and early detection of bile duct malignancies allows for early surgical intervention to improve surgical cure rates and improve patient prognosis. However, due to the particularity of the anatomical structure of the bile duct and the way of tumor growth, the current diagnostic methods still have insufficient sensitivity and accuracy. Therefore, accurate diagnosis of indeterminate biliary strictures is still a major clinical problem.
The diagnosis of biliary strictures is a complex and multiple approach. EUS-FNA is a method for obtaining biopsy tissue from biliary lesions and has 87% sensitivity and 87% accuracy for the diagnosis of indeterminate biliary strictures in a prospective single-center trial31. However, the shortcoming of this method is that it is limited to intrahepatic biliary strictures or proximal extrahepatic biliary strictures. Single-person operation of POCS allows insertion of the extrahepatic bile duct for direct visualization of the lesion and obtaining a targeted biopsy. In a meta-analysis of single-operator POCS including 335 patients, the sensitivity and specificity of visual representation for diagnosing indeterminate biliary strictures were 90% and 87%, respectively, whereas the sensitivity and specificity of single-operator POCS biopsy for diagnosing indeterminate biliary strictures were 69% and 98%, respectively32. However, differences in the interobserver agreement were found in various studies, which may be due to the lack of a standard consensus for the diagnosis of indeterminate biliary strictures in POCS33,34. In addition, POCS may have difficulty accessing the intrahepatic bile duct and obtaining sufficient tissue samples to confirm a malignancy, but pCLE allows adequate visualization of intrahepatic bile duct strictures because of its small diameter14. pCLE has a high accuracy rate in both intrahepatic and extrahepatic biliary strictures due to its fine outer diameter and access to the intrahepatic bile ducts, and there is no statistical difference between intrahepatic (100%) and extrahepatic biliary strictures (86%)17. In addition, for indeterminate biliary strictures where the mass is not visible on the image, it is difficult to obtain the specimen by biopsy, but pCLE can obtain a high diagnostic accuracy of 79% and an negative predictive value (NPV) of 100%30. pCLE is a new imaging technique capable of providing real-time microscopic tissue information in vivo to obtain images of the GI epithelium and sub epithelium at 1,000 magnification during endoscopy. Real-time high-resolution histological diagnosis of mucosa and submucosa tissue structure can achieve the purpose of real-time optical biopsy35. As an emerging endoscopic technique, pCLE is gradually spreading worldwide, and the recent American Society for Gastrointestinal Endoscopy guidelines on the management of biliary tract tumors mention pCLE as a useful alternative to existing diagnostic workup36. In many studies, pCLE has high sensitivity and accuracy for the diagnosis of indeterminate biliary strictures.
The meta-analysis showed that pCLE as a diagnostic tool for indeterminate biliary strictures had a pooled sensitivity of 0.88 and a pooled specificity of 0.79. We reported that the index Q value of sROC was 0.84 and AUC was 0.90, indicating that the overall diagnostic accuracy is very high. The range of DOR is from 0 to infinity, which can be used as an overall assessment of the accuracy of diagnosis. When the DOR value is 1.0, the test is an invalid diagnostic method37. Our study reported a pooled DOR of 24.63, which also illustrates the high overall diagnostic accuracy. Because PLR and NLR are more easily applied in the clinical setting, they were used as an indicator to assess diagnostic accuracy in the current meta-analysis. A PLR value of 3.67 indicated that patients with malignant biliary strictures were approximately three times more likely to be positive compared to patients without malignant biliary strictures. In contrast, the NLR is found to be 0.18, which means that if the pCLE test is negative, the patient has about a 18% chance of having malignant biliary strictures. Therefore, based on current data, PLR and NLR may be used as valuable tools for diagnosing uncertain biliary strictures in pCLE in the future.
We noted differences in diagnostic accuracy between subgroups. The diagnosis of pCLE for indeterminate biliary strictures has a Miami classification and a Paris classification, and we compared the diagnostic accuracy of the Miami and Paris classifications in a subgroup analysis. The Miami classification of pCLE diagnostic indeterminate biliary strictures was developed in 2012 and tested by a multicenter study38. The image features of Miami classification are shown in Table 3. In order to improve the specificity of pCLE diagnosis of indeterminate biliary strictures and reduce the false positive rate related to chronic inflammation, Caillol proposed the Paris classification related to benign inflammation changes28. The Paris classification is based on the Miami classification with the addition of 4 image features of benign inflammatory stenosis as shown in Table 3. The use of the Paris classification to evaluate the diagnostic accuracy of pCLE for indeterminate biliary strictures shows that the Paris classification can improve the specificity of the diagnosis of indeterminate biliary strictures24,29. Similar results were obtained by using the Paris classification in the subgroup of the present study, improving the specificity of the diagnosis from 77.6% to 85.0%. In future studies, each criterion of the Paris classification scheme for diagnosis of indeterminate biliary strictures needs to be evaluated and appropriate thresholds need to be set. In addition, the thickness of the reticular structure in the Paris classification also needs to be quantified.
Because pCLE can reach the biliary strictures by a catheter or a cholangioscopy, we compared the diagnostic accuracy of the two approaches in a subgroup analysis. The pCLE may reach the biliary strictures more accurately by direct visualization with a cholangioscopy, while the pCLE can only reach the biliary strictures by catheter under fluoroscopic guidance. The cholangioscopy can adjust the angle in the bile duct to make the mini probe to the site of interest diagnose indeterminate biliary strictures in pCLE and obtain a high accuracy of 93.3%17. There are similar results in the present study, where the sensitivity and specificity of pCLE for the diagnosis of indeterminate biliary strictures in the cholangioscopy subgroup were higher at 92.3% and 93.5%, respectively. Additionally, pCLE operators' skills should be improved through training, as highly skilled operators are able to achieve high accuracy16,18,39.
In most pCLE diagnostic studies on indeterminate biliary strictures, the CholangioFlex mini probe was used, which has an outer diameter of less than 1 mm and a resolution of 3.5 mm. However, some studies have used the GastroFlex mini probe, which increases the number of optical fibers and has better resolution and image quality19,22,23. Good image quality achieved 100% accuracy while poor image accuracy was only 79%, indicating that good image quality can significantly improve accuracy28. In addition, the Gastroflex probe had a larger outer diameter but no reduction in cannulation rate19. More accurate diagnoses were obtained in some studies using the Gastroflex probe, but the small number of Gastroflex probes used in these studies will require further confirmation in studies with large sample sizes. In the subgroup analysis of this study, the sensitivity and specificity of pCLE in the subgroup of using the GastroFlex mini probe for the diagnosis of indeterminate biliary strictures reached 93.3% and 80.6%. However, the diagnostic performance of the GastroFlex mini-probe will need to be further clarified by multi-center prospective studies in the future. In addition, if some objective diagnostic methods such as artificial intelligence and signal-to-noise ratio are used, it may improve the sensitivity and specificity of pCLE to diagnose indeterminate biliary strictures20.
Brush cytology and intraductal biopsy by ERCP are routinely performed to evaluate indeterminate biliary strictures40. We performed a meta-analysis of brush cytology and intraductal biopsies by ERCP. The meta-analysis showed that ERCP with brush cytology and intraductal biopsy as a diagnostic tool for indeterminate biliary strictures has a pooled sensitivity of 0.54 and a good specificity of 0.96. We reported that the index Q value of sROC is 0.61 and AUC is 0.65, indicating that the overall diagnostic accuracy is poor. There is no overlap between the 95% AUC confidence intervals of pCLE and brush cytology and intraductal biopsy in the diagnosis of indeterminate biliary strictures, indicating that there are significant differences between the two groups. The sensitivity of 0.54 for ERCP with brush cytology and intraductal biopsy is indeed disappointing. The reasons for the low sensitivity could be tumor-associated fibrosis, submucosal spread, or bile duct compression by external lesions. Desmoplastic tumors have a relatively small number of cells and are very firm, making sampling very difficult. Another reason for the poor diagnosis of ERCP with brush cytology and intraductal biopsy is the random sampling of tissues. The pCLE also allows for a more accurate diagnosis by locating different biliary strictures in real time. ERCP with brush cytology and intraductal biopsy combined with pCLE can obtain a more accurate diagnosis18,30.
This study has several limitations. First, as most of the included studies were retrospective studies, and there was potential selective bias needing further prospective studies to confirm. Secondly, the poor quality of some of the included studies may affect the results of the meta-analysis. Third, some patients in some of the included studies were followed up rather than pathology as the gold standard.
In conclusion, pCLE is a reliable and accurate method for diagnosing indeterminate biliary strictures, especially when reaching the biliary strictures by cholangioscopy. However, pCLE is expensive compared to common ERCP, and future studies are needed to confirm the cost-effectiveness. With the development of diagnostic classifications and advances in technology, pCLE will improve the accuracy of diagnosing indeterminate biliary strictures. The pCLE has the potential to overcome the limitations inherent in tissue sampling by ERCP by providing real-time microscopic images of the bile ducts to make an accurate diagnosis of indeterminate biliary strictures.