Preoperative diagnosis of EBNET is challenging because its lack of specific diagnostic indicators and extremely low incidence. Michalopoulos et al.  reported that preoperative diagnosis was only in 4 cases of the 150 EBNET cases between 1959 and 2012. The diagnosis of EBNET mainly relies on the postsurgical pathological and immunohistochemical examination. For this patient, the following questions were raised: (1) how should the surgeons make accurate preoperative diagnosis of EBNET? (2) Whether R1 resection of EBNET G1 requires further treatment and the therapeutic plan?
CgA can be elevated in both functional and non-functional NETs and can be a promising serum marker, but the sensitivity of CgA measurements in patient with NETs is only about 60–90% with a specificity of less than 50% . Neuron-specific enolase (NSE) also has been utilized as a serum marker for NETs, and NSE can be elevated in 30–50% of NETs, especially in patients with high-grade tumors . Functional NETs can produce some hormonal substances, such as somatostatin, polypeptides, serotonin, and calcitonin. It is useful to measure specific hormones in functional NETs. Both CA19-9 and CA-125 are commonly used serum tumor markers for the preoperative diagnosis of cholangiocarcinoma . However, Wang et al.  reported that the positivity of CA19-9 was only 15.0%, CA72-4 was 7.5%, CEA was 17.5%, and AFP was 15.0% in primary hepatic neuroendocrine tumors. Furthermore, serum tumor markers are normal in most EBNET G1/G2, but they tend to be higher than normal levels in EBNET G3 or extrahepatic biliary neuroendocrine carcinomas [10,11,12,13,14]. Our patient with the hilar bile duct NETs G1 had serum tumor markers within normal limits.
Study  showed that the shape of EBNET could be divide into nodular, intraductal-growing, and periductal-infiltrating type. The imaging manifestations of enhanced CT in arterial-phase were as follows: (1) intraductal-growing type of EBNET indicated a higher density than the hepatic parenchyma, and this was helpful for distinguishing from the intraductal-growing type of cholangiocarcinoma showing a lower density than the hepatic parenchyma; (2) nodular type mainly showed equal density compared to hepatic parenchyma; (3) periductal-infiltrating type showed thickening of the bile duct wall and sudden blockage, which was similar to the distal bile duct cholangiocarcinoma. The imaging manifestations of MRI were as follows: (1) T1WI indicated that all tumors were lower SI than the hepatic parenchyma; (2) T2WI indicated that 80% of tumors were higher SI than the hepatic parenchyma; (3) DWI indicated that all tumors were higher SI than the hepatic parenchyma. The MRI manifestations of our patient were consistent with the study (Fig. 1).
Endoscopic and biopsy techniques can be used for preoperative diagnosis of EBNET. Lesions can be detected and biopsied by choledochoscopy, endoscopic retrograde cholangiopancreatography (ERCP), endoscopic ultrasound (EUS) and SpyGlass. Sano et al.  reported a case of well-differentiated EBNET diagnosed successfully by EUS-guided fine-needle aspiration biopsy. Besides, biliary brush cytology has been widely used in the preoperative diagnosis of biliary diseases, and the cytology of bile and bile duct brush specimens were also helpful for preoperative diagnosis of biliary NET [10, 16]. We regretted that the patient did not have bile juice cytology and step biopsy before surgery. Maybe bile juice cytology and step biopsy can help us make a correct preoperative diagnosis. The patient is very young, just 24 years old. This also reminds us that we should not easily diagnose cholangiocarcinoma in young patients, and more comprehensive preoperative examination is needed.
The unexpected pathological examination challenging us to give proper treatment. A multidisciplinary discussion proposed four options for this patient: (1) a second operation should be performed, and the entire tumor can be removed. But the new procedures would bring more surgical trauma to this patient. (2) The well-differentiated neuroendocrine cells are known to overexpress somatostatin receptors (SSR). The somatostatin analogs (SSA) or peptide receptor radionuclide therapy (PRRT) may be a good option for patients with well-differentiated NET G1. Martyn et al.  reported that compared with the placebo group, the SSA group could significantly prolong progression-free survival (PFS) in patients with metastatic enteropancreatic neuroendocrine tumors (Ki-67 < 10%). Ersin et al.  found that PFS in the SSA group was 21 months, which was better than the chemotherapy group as their first-line treatment for NET (Ki-67 ≤ 20%). PRRT has become an effective treatment for the NETs which express sufficient SSR. Both 90Y and 177Lu were used as radioactive isotopes in PRRT for NET. 177Lu–tetraazacyclododecanetetraacetic acid–octreotide (177Lu-DOTATATE) therapy was recommended for patients with SSR-positive NET in the US in January 2018 and in Europe in September 2017 . To make sure SSR is positive, SSR imaging such as 68 Ga-DOTATATE PET/CT can be given firstly before PRRT is administered. Studies [20, 21] showed that 68 Ga-DOTATATE PET/CT was safer and efficient for diagnosis and treatment management of NET, and should be the preferred imaging method for preliminary diagnosis, selection of patients for PRRT, and localization of unknown primary tumors. (3) Systemic chemotherapy and targeted therapy may be a treatment choice for this patient, the chemotherapy can improve resectability and control tumor progression. However, chemotherapy is mainly used for patients with high-grade and metastatic NETs. And the role of chemotherapy is still undetermined in well-differentiated NETs, and still lacks standard indications . Many targeted drugs are under research, but few targeted drugs have entered phase III clinical trials. At present, only sunitinib and everolimus are FDA-approved targeted drugs for NETs . Targeted therapy gives hope for low-grade and intermediate-grade NETs. A phase III trial  showed that everolimus could significantly improve PFS for patients with advanced, well-differentiated, and non-functional NETs. However, there is no specific evidence to clarify that chemotherapy or targeted therapy can benefit the hilar bile duct NET G1 with R1 resection. (4) NET G1 may be considered indolent, no further treatment required, and requires regular follow-up.
Our patient chose no further therapy after three cycles of adjuvant chemotherapy and follow-up strictly at our outpatient service. At present, the patient has been following up for 24 months without recurrence or disease progression. Therefore, we believe that postoperative prophylactic intravenous chemotherapy is beneficial for NETs, especially for patients in G3. Many more clinical trials are ongoing, and these results will be clarified in the future. We know little of biliary NETs because of its rarity. There are currently no guidelines for the diagnosis and treatment of biliary NETs. Here we reported a case of perihilar bile duct NETs G1 with R1 resection, as far as we know this is the first report. More information about biliary NETs must be registered.