Prognosis value of microscopic bile duct invasion in hepatocellular carcinoma: A multicenter study

Abstract Objective To evaluate the prognostic significance of microscopic bile duct invasion (MiBDI) in hepatocellular carcinoma (HCC) following R0 resection. Patients and Methods Patients who underwent R0 resection for HCC at nine medical centers were stratified into five groups: neither bile duct nor vascular invasion (MiBDI−MVI−), microscopic bile duct invasion alone (MiBDI+MVI−), both microscopic bile duct and vascular invasion (MiBDI+MVI+), microscopic vascular invasion alone (MiBDI−MVI+), and macroscopic bile duct invasion (MaBDI). Overall survival (OS) was assessed using Kaplan–Meier analysis, and independent risk factors of OS were determined using Cox proportional hazards models. Results A total of 377 HCC cases were analyzed. The OS for MiBDI+MVI‐ was similar to that of MiBDI−MVI− (p > 0.05) but better than MiBDI+MVI+, MiBDI−MVI+, and MaBDI (all p < 0.05). Multivariate analysis indicated that MiBDI was not an independent risk factor for OS, while MVI and MaBDI were. Conclusions Overall survival (OS) in patients with MiBDI was superior to those with MVI and MaBDI. Isolated MiBDI did not influence OS in patients with HCC after R0 resection.


| INTRODUCTION
Hepatocellular carcinoma (HCC) is the predominant form of primary liver cancer, ranking fourth in morbidity and second in mortality among all cancers. 1 Hepatocellular carcinoma (HCC) is characterized by significant malignancy and strong invasiveness, making it prone to invasion and metastasis during tumor initiation and progression.Tumor cells infiltrate blood vessels, leading to vascular invasion or vascular tumor thrombus.While tumor cells infiltrate bile duct, leading to bile duct invasion (BDI) or bile duct tumor thrombus (BDTT).
Among these distinct anatomical invasion types, clinical research on vascular invasion is more comprehensive.Histopathological classifications of vascular invasion typically differentiate between microscopic (MVI) and macroscopic tumor thrombi (MaVI).MVI and MaVI have reported incidence rates of 15%-75% and 44%-62.2%,respectively. 2,3Both two types of vascular invasion are considered independent prognostic factors for poor outcomes in HCC and are incorporated into the 8th edition of the TNM staging system as essential elements for predicting prognosis and informing treatment decisions. 4In the 8th edition TNM staging, patients with either MVI or MaVI are classified as stage II and stage IIIB, respectively.Conversely, the incidence of BDI is lower, with notably fewer studies compared with vascular invasion.To date, researches on BDI primarily originated from small sample-size or case report-based retrospective studies.
As research has advanced, clinicians and pathologists have enhanced their understanding of BDI.In recent years, the detection rate of BDI has been increasing. 51][12][13][14] For patients with BDI who receive only palliative care or the best supportive treatment, the median survival interval is a mere 1.6-4.3months. 15,16Consequently, BDI is considered a risk factor affecting HCC prognosis, and several studies have confirmed that its prognosis is better than that of MaVI. 17,18However, unlike MVI, only a few single-center, small sample-size retrospective studies exist on microscopic bile duct invasion (MiBDI).A retrospective study from Japan suggested that, compared with MiBDI, macroscopic bile duct invasion (MaBDI) was a protective factor for prognosis. 19A matched study from South Korea indicated that, compared with HCC without BDI, the prognosis of MiBDI was worse. 20A retrospective propensity score matching study from China revealed that the prognosis of MaBDI was worse than that of MiBDI, and the overall survival did not significantly differ between HCC without BDI and HCC with MiBDI. 21The prognostic differences MiBDI and MaBDI remain a subject of debate.As for distinctions between MiBDI and MVI, no relevant research has been conducted to date.
The aim of the present study is to distinguish the survival differences of patients with MiBDI, MaBDI, and MVI, and to evaluate the prognosis significance of MiBDI for HCC.

| Ethics statement
This study was conducted in accordance with the Declaration of Helsinki and the Ethical Guidelines for Clinical Research.Informed consent was obtained from all patients before surgery.Approval was obtained from the Institutional Research Ethics Committee of Mengchao Hepatobiliary Hospital of Fujian Medical University, with the approval number 2020_077_01.The data are anonymous, and the requirement for informed consent was therefore waived.The inclusion criteria for this study were as follows: (1) confirmation of HCC and BDI by histopathology, (2) receipt of R0 resection, and (3) availability of complete clinical data and postoperative follow-up records.The exclusion criteria were as follows: (1) combined HCC intrahepatic cholangiocarcinoma, (2) recurrent or metastatic HCC, (3) presence of portal vein tumor thrombus (PVTT), or (4) co-occurrence with other malignancies.R0 resection was defined as the microscopically negative margin.Patients were further categorized based on their bile duct and vascular invasion status: no bile duct or vascular invasion (MiBDI−MVI−), microscopic bile duct invasion only (MiBDI+MVI−), both microscopic bile duct and vascular invasion (MiBDI+MVI+), microscopic vascular invasion only (MiBDI−MVI+), and macroscopic bile duct invasion (MaBDI).

| Clinicopathological characteristics
The clinicopathological characteristics that were included in the study were age, gender, presence of liver cirrhosis, history of hepatitis B virus infection, number of tumors, tumor size, presence of satellite nodules, tumor differentiation grade, and microvascular invasion (MVI).Cirrhosis was confirmed either by histopathology or by clinical diagnosis.Tumor differentiation grade was classified according to the Edmonson-Steiner grade.MVI was defined as vascular invasion of small vessels discovered under the microscope.Referring to the definition of MVI in the clinical guidelines, 22 the MiBDI was defined as the mass of cancer cells observed in the bile duct cavity under the microscope.

| Treatments
The therapeutic schedule for patients in this study was determined by a multidisciplinary team at each hepatobiliary medical center.The decision to hepatectomy was based on the clinical tumor stage, liver function, and residual liver volume of the patients.

| Postoperative Follow-up
All patients underwent REGULAR follow-up appointments after discharge, scheduled at 3-month intervals for the first 2 years, at 6-month intervals for 2-5 years, and annually thereafter.Follow-up examinations included laboratory tests for tumor biomarkers, coagulation function, liver function, and whole blood count, as well as abdominal ultrasound and/or contrast-enhanced computed tomography or magnetic resonance imaging.Overall survival (OS) was defined as the time from surgical resection to death or last follow-up, and follow-up for this study was censored on March 31, 2022.

| Statistical analysis
Continuous variables were presented as the median with quartile ranges and were compared using either Student's t-test or Mann-Whitney test.Categorical data were expressed as frequency (%) and were compared using either the chi-squared test or Fisher's exact test.Overall survival (OS) was evaluated using Kaplan-Meier survival curves.Univariate and multivariable analyses were performed using the Cox proportional hazard regression model, with the backward stepwise selection method used to determine independent prognostic factors.Statistical significance was defined as p-values less than 0.05.The statistical analysis was conducted using R version 4.2.2 (http:// www.r-proje ct.org/ ) and IBM SPSS Statistics software (version 24.0;IBM Corp., Armonk, NY).The R packages utilized were "readxl," "readr," "table 1," "ggpubr," "plyr," "survminer," "survival," and "ggplot2.

| Patients' characteristics
The study enrolled 377 cases of HCC who received R0 resection, including 139 cases in the group with neither bile duct nor vascular invasion (MiBDI−MVI−), 11 cases in the group with microscopic bile duct invasion alone (MiBDI+MVI−), 37 cases in the group with both microscopic bile duct and microscopic vascular invasion (MiBDI+MVI+), 56 cases in the group with microscopic vascular invasion alone (MiBDI−MVI+), and 156 cases in the group with macroscopic bile duct invasion (MaBDI).The patients' characteristics are detailed in Table 1.

| Microscopic histopathological characteristics of the BDI and MVI
Significant differences were observed between the microscopic features of bile duct and microvascular.The interlobular bile duct wall comprises epithelium and fibrous membrane, with the epithelium gradually transitioning from a single cuboidal to a single columnar shape, and an oval nucleus located at the base of the cell (Figure 1A,B).In contrast, the vascular intima consists of endothelium and subendothelium, with the endothelium lining the vascular lumen as a single layer of flat epithelium.The  nucleus is located in the middle, lightly stained, and the edge is thin (Figure 1C,D).2).In addition, further COX multivariate analysis which included MiBDI (n = 48), MVI (n = 186), and MaBDI (n = 134) as variables was conducted and the result still showed that MiBDI was not but MVI and MaBDI were the risk factors of OS (Table S1).6][7][8][9] The occurrence of microscopic bile duct invasion is even more infrequent.As a result, there has been limited focus on its diagnosis and treatment, leading to a scarcity of relevant studies.Clinicians possess inadequate understanding of this phenomenon and may mistakenly classify microscopic bile duct invasion as MVI. 23The misclassification of these two types of invasions is likely to result in an inaccurate prognostic evaluation and subsequently improper medical intervention.To our knowledge, there are no studies on the difference in prognosis between MVI and MiBDI.

| Overall survival
Given that the prognosis of patients with HCC relies on precise prognostic assessment and suitable treatment recommendations, we undertook this study to evaluate the prognostic significance of microscopic bile duct invasion.
A retrospective study showed that the prognosis of HCC with MaBDI was better than HCC with MiBDI, 19 while a propensity score matching study came to the opposite conclusion. 21Our results show that microscopic bile duct invasion is distinct from MVI, both morphologically and prognostically.From a morphological standpoint, there are clear differences between MiBDI and MVI.The bile duct epithelium primarily consists of cuboidal and columnar epithelial cells, while the vascular endothelium is mainly composed of single-layered, flat epithelial cells.These differences make it relatively easy to distinguish the two morphologically.Additionally, immunohistochemical staining can be used to differentiate between them.For instance, CD34 is associated with brownish-yellow staining of the cell membrane, indicating the presence of vascular endothelial cells.Conversely, CK19 does not stain liver cells but can specifically stain bile duct epithelium. 24In the survival analysis, we aimed to exclude the influence of MVI by grouping the subjects based on different types of tumor invasion.Our findings revealed that the OS of MiBDI+MVI− was not significantly different from that of MiBDI−MVI− (p = 0.32), but it was better than the outcomes for MiBDI+MVI+ (p = 0.0068), MiBDI−MVI+ (p = 0.048), and MaBDI (p < 0.001).Further multivariate analysis confirmed these results.Consequently, we believe that pure microscopic bile duct invasion does not impact the overall survival of HCC patients after R0 resection.Numerous studies have been conducted on vascular invasion.6][27] In this study, the prognosis of MiBDI is also superior to that of MVI and better than that of MaBDI.Therefore, we summarize this phenomenon as "microscopic invasion is better than macroscopic invasion, and bile duct invasion is better than vascular invasion."As for the possible reasons why BDI has a more favorable prognosis than vascular invasion, we speculated that this is because the corresponding distal organs are different.The invasion of tumor cells into the bile ducts and subsequent entry into the intestine via bile, coupled with the unfavorable conditions for tumor growth provided by the digestive juices in the intestine and the tumor cells are excreted along with the feces, bile duct invasion is less prone to distant metastases.Conversely, tumor cells that infiltrate blood vessels could disseminate throughout the body via the bloodstream, leading to the formation of metastatic cancer in various organs such as the lungs, bones, and brain, consequently impairing relevant organs.
This study has the following limitations: First, in the real world, there are far more HCC patients without BDI than those with BDI.To increase the sample size, we collected data through a multicenter study.However, due to the substantial workload, it was not feasible to collect HCC cases without BDI from all centers within the same period as a control, leading to potential selection bias in this retrospective study.Second, the number of patients with pure microscopic bile duct invasion is relatively small (only 11 cases), necessitating a larger sample and a well-designed clinical study for further validation.Lastly, the patients in this study are all from China, with most of them having hepatitis B infections.The prognostic role of microscopic bile duct tumor thrombus in non-hepatitis B-related liver cancer requires additional investigation.

F I G U R E 2
Kaplan-Meier survival curves of HCC with different tumor invasion status.