We conducted a three-step animal study and found that when an animal BBS model was created using endobiliary RFA, a previously reported method, there were serious adverse events and a high probability of failure. The process of BBS formation was first examined histopathologically, and several methods to prevent adverse events were added to create an optimal BBS model using endobiliary RFA. Eventually, through this study, we demonstrated its reproducibility.
Ideal animal models used for research have challenging problems of high price and low reproducibility, so small animals such as mice and rats are mainly used. However, there is a significant problem in that endoscopic procedures and endoscopic devices must be directly verified, necessitating the use of large animal models in those studies. In terms of cost, this inevitably becomes a burden for researchers and device manufacturers. As mentioned in the Introduction, to determine the effect of a biliary stent or device, a model for improving BBS must be primarily used. The merit is greater in terms of cost if the procedure is performed directly on the animal, but the cost becomes a bigger problem because the test stent or device must be evaluated after the model is artificially created. Since previous animal BBS models using endobiliary RFA have already been demonstrated to some extent, our main purpose was to understand the process of making a BBS and add several new processes to improve the reproducibility.
According to the results in the first step, unexpected sudden death occurred in 40% of the pigs before the BBS model was created. This is a critical and important issue indicating that this animal experimental model could not be used. The cause of death was a problem, with CBD drainage prior to the onset of the BBS, which resulted in cholangitis-induced sepsis. The ERC test performed on the surviving animals showed that the degree of BBS was so severe that the guidewire was not inserted in one case. As suggested in a previous study, since a BBS caused by RFA is induced slowly, it seemed that sudden CBD obstruction would not be a problem. However, in practice, critical problems occurred before the BBS was completely formed7.
Based on the results of the first step, we placed a plastic stent to prevent BBS induced biliary obstruction and secondary cholangitis in the second-step trial. As shown in all clinical studies on intraductal RFA in malignant biliary obstruction, a stricture is an unconditional adverse event after RFA. Therefore, biliary stenting with a plastic or metal stent was performed in all patients to prevent this occurrence19,20,21,22. Based on this, we tried inserting a plastic stent for the same purpose. However, one animal with early stent migration died suddenly on the fourth day, and stent migration was also observed in two animals during follow-up. Perhaps, unlike a malignant biliary stricture, since a plastic stent was inserted before a benign stricture was formed, stent migration may have easily occurred. On microscopic examination (Fig. 6), it was found that an abscess had formed on the entire wall of the APC application area as the result of severe inflammation at the beginning of the procedure. BBS-induced cholangitis may cause sepsis, but sepsis could also be caused by abscess formation in the CBD wall. As a result, it was critical to administer broad-spectrum antibiotics during the follow-up period, and the animals’ unexpected sudden death could be avoided. In addition, it was found that fibrosis at the APC application site began on the tenth day and was completed by about the 20th day. This finding suggests that this BBS animal model can be used three weeks after RFA application to the CBD.
Based on the results of the previous two steps, both pigtail-type stents were replaced to prevent stent migration. The biggest concern with adding this step was that as the stent is inserted, stricture and upstream dilation may not be created as expected. Because stent migration occurred at four weeks in one out of 11 pigs (9.1%), which was lower than we expected, it is thought that unexpected sudden death could be prevented by good stent maintenance in the early follow-up period. According to the laboratory tests (Table 2), obstructive cholangitis did occur even if the stent had not migrated from the CBD. As we presumed, stent function was well-maintained at an early stage to prevent sudden death, but the stent function became problematic after two weeks due to the thin diameter of the stent. As a result, while this induced less upstream CBD dilation, there was no problem in maintaining survival for four weeks due to the administration of antibiotics. Upstream CBD dilation is a phenomenon that accompanies the formation of a BBS and is not a necessary result. However, it would be easier to insert the test device or stent when dilation is formed wider than the original diameter.
Based on the new findings of this three-step study, we would like to present guidelines for creating an optimal benign biliary stricture animal model (Table 4). Because RFA settings differ depending upon the RFA generator and probe used, the guidelines for RFA settings are limited. However, it is thought that if the electrode length is too long, there is a high risk of adverse events occurring before the BBS is formed. In contrast, if the electrode length is too short, it may be difficult to use it to assess the treatment effect of stents or devices used for BBS. For the above reasons, we recommend a length of about 20 mm. To ensure maximum contact between the probe and the CBD mucosa during RFA application, a biliary sphincterotomy in advance, and endoscopic suction are recommended. If the initial CBD diameter is not wide, the RFA probe could easily contact the CBD wall. Therefore, biliary sphincterotomy might not be necessary, and rather, promote the migration of plastic stents. Additional tips for preventing adverse events include administering intravenous broad-spectrum antibiotics during the follow-up period, inserting biliary plastic stents after RFA, and considering a double-pigtail-type stent to prevent stent migration.
The limitations of this experimental study were as follows. 1) This study used an in vivo animal model, and the sample size was small. 2) RFA settings differed depending upon the RFA generator and probe used. 3) In the unexpected death of animals, it was necessary to guess the cause of death as an autopsy finding due to that accurate laboratory analyses could not be performed during the autopsy. Nevertheless, although additional large-scaled studies are required to observe the reproducibility of this BBS animal model, our current findings provide basic and standardizing techniques for creating an optimal BBS animal model. Thus, the present study results yielded important insight into the potential benefits of the BBS animal model, which be helpful in the development of a new device, technology, or treatment strategies. Since making a large animal model is difficult and requires very sophisticated techniques, researchers may encounter numerous difficult processes, and unexpected adverse events may occur. It is most important for researchers wanting to make this model learn the methods for minimizing adverse events from well-trained researchers before beginning their work. Nevertheless, we think that the learning curve of researchers can be greatly shortened by referring to the optimal methods we introduced as a result of this study. Also, if the optimal method of making the BBS animal model is well established, it will be helpful in various ways in the future. For example, according to a recently reported animal study, pancreatic duct ligation causes pancreas atrophy and eventually, reduced premalignant pancreatic lesions23. If the method of inducing pancreatic atrophy by applying endoluminal RFA to the main pancreatic duct is well established24, it will have great potential for use clinically as a pancreatic cancer treatment method.
In conclusion, endobiliary RFA combined with a biliary plastic stent resulted in a safe and reproducible BBS animal model. Also, we first examined the process of BBS formation through time-sequential histopathological examinations. Our detailed experimental results will provide basic information and a basis for future BBS animal model production.