BilIN has been defined by Zen et al. in 2005 as microscopic lesions of flat or low-papillary dysplastic epithelium [13, 14]. BilIN is regarded as the biliary counterpart of pancreatic intraepithelial neoplasia (PanIN) [11]. It is characterized by the pathological findings of irregular nuclei, cell dysplasia and irregular polarity in cytology [12]. The microscopic appearances of BilIN are classified as flat, pseudopapillary or micropapillary [12, 15]. BilIN is divided into three grades (BilIN-1, BilIN-2, and BilIN-3) based on the following histological features: degree of cellular atypia, nuclear pseudostratification and apical surface protrusions, and loss of cellular polarity [1]. Low-grade and high-grade dysplasia are abbreviated as BilIN-1 and -2, and in situ carcinoma as BilIN-3 [4]. BillN is considered to progress with increasing neoplastic potential, with BilIN-3 shifting to CC in the multistep carcinogenesis sequence [14]. Metaplastic changes of the epithelium of the bile duct can be induced by chronic inflammation, and patients often have a preceding chronic biliary disease like primary sclerosing cholangitis (PSC), choledocholithiasis, and pancreaticobiliary malfunction. with BilIN [12, 16].

In our case, there were two atypical findings regarding BilIN. One is that it occurred in a patient with no history of a prior biliary disease or chronic inflammation. Although the mucosa around the tumor showed redness, the range was limited to the junction of the cystic and common hepatic ducts. The other atypical finding is the large tumor size of 25 × 10 mm. Although BilIN is usually unrecognizable by macroscopy, the tumor was clearly detected in CT and could be observed as an elevated mass in the cholangioscope. To the best of our knowledge, it was larger than any other case of BilIN without CC reported in the literature so far and was therefore very difficult to differentiate from CC. These two findings in our case suggest that the occurrence of BilIN can be triggered by factors other than inflammation and that the tumor can grow to a size large enough to be detected by image analyses without shifting to CC.

The tumor was ulcerated and hemorrhaging in the CBD when the patient was referred to our hospital. In our case, obstructive jaundice caused by the hematoma in the CBD led to the detection of BilIN before progressing to CC. CC is not a major cause of hemobilia, which is present in only 3% of cases [17]. It was unclear why and when the tumor developed the ulcer and bleeding. It is also unknown that BIlIN itself causes stenosis of the bile duct and obstructive jaundice. In PanIN, the stenosis of the main pancreatic duct combined with distal ductal dilatation and obstructive pancreatitis is often observed [18]. However, PanIN itself is not considered to cause duct stenosis, because PanIN is almost flat with no intraductal proliferation, but inflammation or fibrosis [18]. The inflammation detached dysplastic epithelium and leaked pancreatic juice may cause further inflammation, fibrosis, and finally the stenosis [18]. According to this discussion, we speculate that BilIN itself also does not cause bile duct stenosis as well as PanIN, but if it combined with some inflammation, it could cause the destruction of bile duct epithelium, the stenosis of the bile duct, and obstructive jaundice.

It is important to detect precursor lesions early in cancer therapy, although the prognosis of BilIN without CC has not been clarified and a uniform treatment guideline is lacking. And unlike endoscopic resection for adenoma in the gastrointestinal tract, a minimally invasive treatment in the bile duct has yet to be developed. The only curative treatment for BilIN prior to malignant transformation is surgical resection [10]. In surgical resection for CC, an inspection of the resection margin often reveals the presence of BilIN [13]. It is well-established that resection with a cancer-free margin is crucial for curative therapy, but the clinical significance of BilIN in the surgical margin is unclear and additional resection is still controversial. Matthaei et al. reported that BilIN found in the surgical margin of biliary tract cancer resection in 53%; however, it does not require additional resection because the patients with BilIN-positive at the surgical margin and BilIN-negative did not differ significantly regarding overall survival [13]. They also reported that survival even of patients with BilIN-3 was not shorter than that of patients without any BilIN at margin [13]. Several studies also reported as well but the number of patients reported was small, so further study with a larger population and longer term is required [19, 20]..

Many molecular and genetic alternations have been found to occur and accumulate in BilIN during the multistep process of cholangiocarcinogenesis [12, 21]. The expression pattern of molecules in the tumor tissue biopsy related to cell cycle and carcinogenesis might be helpful in the preoperative diagnosis of BilIN and differentiation from CC. Representative molecules considered that can be upregulated in BilIN are p21, p53, cyclin D1, S100P, MUC1, and MUC5AC [2, 11, 12]. Some of them can be upregulated in CC too [11, 12]. On the other hand, most BilIN shows negativity for MUC2 [22]. The expression of MUC1 is increased along with the progression of histological grade in BilIN and associated with a poorer outcome in CC [2, 23]. MUC5AC expresses in the early stage of cholangiocarcinogenesis [24]. Recent studies showed the expression of S100P was increased in BilIN-2 and -3 as well as cholangiocarcinoma, and Sato et al. proposed the diagnostic algorithm of histological grading of BilIN including the expression of S100P in the cytoplasm that helps the grading of BilIN1–3 [12, 18, 25]. The expression level of p53 is very low not enough to be detected by immunohistochemistry in non-neoplastic cells because its half-life is normally very short [11]. P53 is mutated in a large number of malignant neoplasms, then mutated p53 can be detected immunohistochemically based on its overexpression [11]. Nakanishi et al. reported that p53 was significantly upregulated in BilIN-3 and CC compared with BilIN-1 and -2, and suggested that the expression of p53 might be involved in the acquisition of invasive growth in CC [11]. Such research on BilIN is crucial to elucidate the multistep carcinogenesis sequence from BilIN to CC. In our case, the tumor showed weak expression of S100P, positive MUC5AC, negative MUC1, and not overexpressed p53, which is consistent with the pathological features of BilIN-1 and BilIN-2.