Survival and Recurrence in Liver Transplant Patients With Intrahepatic Cholangiocarcinoma and Hepatocellular Carcinoma

Introduction

Hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC) are the 2 major forms of primary liver cancer. While HCC has been studied extensively, the post-treatment prognosis of ICC is unclear, despite a steady increase in the global incidence. The prognosis of patients with ICC remains dismal, due primarily to high recurrence rates following surgical resection, systemic chemotherapy, and even liver transplantation (LT) [1–4]. Given these unfavorable outcomes, ICC has traditionally been regarded as a contraindication for LT [5]; however, a subset of patients initially listed for LT after a radiological diagnosis of HCC harbor ICC incidentally, which is actually misdiagnosed HCC or undetected ICC, as confirmed by histopathological evaluation of the explanted liver [6–8]. These incidental findings have raised important clinical questions regarding the prognosis and management of ICC in the context of LT [9,10].

Advances in molecular oncology have strengthened the cancer stem cell theory, which posits that primary liver cancers – including HCC, ICC, and combined hepatocellular–cholangiocarcinoma (cHCC-CCA) – originate from a common population of stem or progenitor cells [11,12]. In cirrhotic livers, ICC can either present as a distinct entity or coexist with HCC or cHCC-CCA, thereby complicating preoperative diagnosis and management. The incidence and clinicopathological features of ICC in LT recipients remain unclear.

In this study, we aimed to evaluate the long-term outcomes of 25 patients with pathological diagnosis of ICC with or without concurrent HCC or cHCC-CCA upon examination of the liver explant after LT. Furthermore, we propose tailored post-transplant surveillance strategies to facilitate cost-effective, long-term follow-up care in this unique patient population.

Material and Methods

STUDY DESIGN AND PATIENT SELECTION:

This observational study was conducted in a single-center retrospective cohort. A comprehensive review of the institutional LT database was conducted to identify cases with pathologically-confirmed ICC in explanted livers. Over the 20-year period from January 2003 to December 2022, 25 patients were identified as having ICC, either as a single primary tumor or in conjunction with HCC or cHCC-CCA. During the same timeframe, 6611 adult patients underwent primary LT, indicating that ICC accounted for approximately 0.4% of all adult LT cases. The medical records of all included patients were analyzed retrospectively, and follow-up data were collected through July 2025 or until the time of death using a combination of institutional medical records and information obtained from the National Health Insurance Service. The study protocol was approved by the Institutional Review Board (IRB No. 2025-0952). Given the retrospective nature of the study, the requirement for informed consent was waived. All procedures were conducted in accordance with the ethical standards outlined in the Declaration of Helsinki (2013 revision).

PREOPERATIVE EVALUATION, SURGICAL PROCEDURES, AND POST-TRANSPLANT FOLLOW-UP:

The standard pretransplant evaluation protocol, including cancer screening and malignancy assessment, has been reported previously [6,13–15]. The Asan Medical Center criteria for HCC were primarily applied to patient selection [15]. Tumor staging of ICC and HCC was conducted in accordance with the 8th edition of the American Joint Committee on Cancer (AJCC) classification system [16]. Very early-stage ICC was defined as a solitary lesion measuring ≤2 cm in diameter [7,8]. The immunosuppressive regimens for LT recipients diagnosed with a malignancy have been reported previously [17]. Post-transplant surveillance was performed monthly during the first year, every 3 months for the subsequent 4 years, and then at intervals of 3–4 months thereafter. Tumor recurrence (TR) was based on radiologic findings, histologic confirmation, or both. All-type TR includes any recurrence of ICC or concurrent HCC and cHCC-CCA. Management of TR followed the established treatment guidelines for post-transplant liver malignancies [14,17,18].

STATISTICAL ANALYSIS:

Quantitative data are reported as either medians with ranges or as means with standard deviations, depending on data distribution. TR and overall survival (OS) were estimated using the Kaplan-Meier method, with group differences assessed by the log-rank test. A P value <0.05 was considered statistically significant. All statistical analyses were performed using SPSS software (version 22; IBM Corp., Armonk, NY, USA) and MedCalc (version 23.2.1; MedCalc Software Ltd., Ostend, Belgium).

Results

CLINICOPATHOLOGICAL FEATURES:

The clinical features of the 25 LT recipients with ICC are presented in Table 1. The mean age was 56.9±6.3 years and 23 (92.0%) patients were male. Primary liver diseases were hepatitis B virus (HBV) infection in 17 patients, hepatitis C virus (HCV) infection in 4, alcoholic liver disease in 2, and other liver cirrhosis in 2. The mean model for end-stage liver disease score was 12.5±8.3.

The preoperative diagnosis for 24 patients was HCC; ICC or cHCC-CCA was preoperatively suspected in only 1 patient due to marked elevation of CA19-9 level. Treatments before LT included transcatheter arterial chemoembolization (TACE; n=15), radiofrequency ablation (RFA; n=1), hepatic resection for HCC (n=1), and hepatic resection for HCC combined with TACE (n=2). Only 6 patients (24.0%) did not undergo any treatment for HCC prior to LT.

The mean and median levels of pretransplant tumor markers were 90.9±178.3 ng/ml and 15.4 ng/ml (range: 2–678 ng/mL) for alpha-fetoprotein (AFP; institutional reference, 7.5 ng/mL); 50.5±80.3 mAU/mL and 22 mAU/ml (range: 9–376 mAU/mL) for proteins induced by vitamin K absence or antagonist-II (PIVKA-II; institutional reference, 40 mAU/mL); 42.3±46.6 U/mL and 22.5 U/mL (range: 2.6–186 U/mL) for carbohydrate antigen 19-9 (CA19-9; institutional reference, 37 ng/mL); and 3.1±2.4 ng/mL and 2.7 ng/mL (range: 0.7–9.1 ng/mL) for carcinoembryonic antigen (CEA; institutional reference, 5 ng/mL), respectively. Elevated AFP was observed in 17 of 25 patients, PIVKA-II in 5 of 22 patients, CA19-9 in 7 of 25 patients, and CEA in 3 of 21 patients.

The types of LT were as follows: deceased-donor whole LT in 3 (12.0%) patients and living-donor LT in 22 (88.0%) including right liver graft implantation in 19 (76.0%) and dual-graft implantation in 3 (12.0%). Six recipients (24.0%) received ABO blood group-incompatible LT.

The pathological findings are summarized in Table 1. For ICC, the mean and median tumor diameter was 2.7±1.6 cm and 2.2 cm (range: 0.5–7.0 cm), respectively; the number of ICC tumors was 1 in 22 (88.0%) patients, 2 in 2 (8.0%), and 3 in 1 (4.0%). According to the 8th AJCC tumor staging system, these ICC tumors were classified as stage IA (19 recipients; 76.0%), stage IB (1; 4.0%), stage II (3; 12.0%), and stage IIIB (2; 8.0%). Very early-stage ICC comprising a single mass less than 2 cm in diameter was found in 7 (28.0%) patients. The growth pattern of ICC in all 25 cases was the mass-forming type.

Concurrent viable second primary liver malignancy was detected in 10 patients: HCC in 9 and cHCC-CCA in 1. Six liver nodules showing TACE-induced complete pathological response were identified in 3 patients, but such non-viable nodules were not analyzed in this study. Since there was only 1 case of cHCC-CCA, it was included in the concurrent HCC group for convenience.

The mean maximal and median viable HCC diameters were 2.6±0.7 cm and 2.5 cm (range: 1.5–3.5 cm), respectively. The number of viable HCC tumors was as follows: 1 in 7 patients, 2 in 1 patient, 7 in 1 patient, and 8 in 1 patient. Microvascular invasion was identified in 2 patients. These concurrent HCCs were classified as AJCC stage IA (n=1), stage IB (n=5), and stage II (n=4). Concurrent HCC was present in 4 of 7 patients with very early-stage ICC.

PATTERNS AND TREATMENT OF TUMOR RECURRENCE:

No patient died of perioperative complications after LT. During the mean follow-up period of 64.6±62.5 months, all-type TR occurred in 18 recipients (72.0%).

The initial sites of TR were intra- and extra-abdominal lymph nodes (n=7), the liver graft (n=5), lungs (n=3), adrenal glands (n=1), duodenum (n=1), and bone (n=1).

In 15 patients with ICC alone, 11 (73.3%) experienced TR. ICC recurrence in 1 patient who underwent pulmonary metastasectomy was confirmed pathologically. Initial treatment of recurrent ICC lesions included systemic chemotherapy (n=7), RFA (n=2), pulmonary metastasectomy (n=1), and no specific treatment (n=1). Ten patients finally underwent systemic chemotherapy after progression of recurrent ICC.

In 10 patients with concurrent HCC or cHCC-CCA, 7 (70.0%) experienced all-type TR. Two patients were diagnosed with ICC recurrence, which was confirmed pathologically in 1 patient who underwent pulmonary metastasectomy and systemic chemotherapy. ICC recurrence in the other patient was diagnosed through high elevation of CA19-9 level; this patient underwent radiotherapy for adrenal metastasis. HCC recurrence was diagnosed in 5 patients; it was pathologically confirmed in 3 patients who underwent pulmonary metastasectomy (n=1), adrenalectomy (n=1), and paraesophageal lymphadenectomy (n=1); these 3 patients finally underwent systemic chemotherapy after progression of recurrent HCC. The other 2 patients, who showed elevated AFP and/or PIVKA-II levels, initially underwent RFA (n=1) and external beam radiotherapy (n=1). Finally, both underwent systemic chemotherapy after progression of recurrent HCC.

TUMOR RECURRENCE AND SURVIVAL OUTCOMES:

During a follow-up period up to 227 months, 18 (72.0%) patients experienced all-type TR. Among these, 17 (94.4%) died due to ICC recurrence (n=13) or HCC recurrence (n=4). The 6-month, 1-year, 3-year, 5-year, and 10-year all-type TR rates were 32.0%, 52.0%, 60.0%, 72.0%, and 72.0%, respectively (Figure 1A).

During follow-up, 19 of 25 patients (76.0%) died due to all-type tumor recurrence (n=17), pneumonia (n=1), or unknown cause (n=1). The 6-month, 1-year, 3-year, 5-year, 10-year and 15-year OS rates were 92.0%, 80.0%, 56.0%, 47.0%, 32.2%, and 10.7%, respectively (Figure 1B).

After stratifying patients according to concurrent viable HCC and cHCC-CCA, the 1-year, 3-year and 5-year all-type TR rates were 53.3%, 60.0%, and 73.3%, respectively, for 15 recipients with ICC alone; and 50.0%, 60.0%, and 70%, respectively, for 10 recipients with concurrent viable HCC and cHCC-CCA (P=0.832; Figure 2A). The 1-year, 3-year, 5-year, 10-year, and 15-year OS rates were 73.3%, 53.3%. 46.7%, 25.0%, and 12.5%, respectively, for 15 recipients with ICC alone; and 90.0%, 60.0%, 48.0%, 48.0%, and 0%, respectively, for 10 recipients with concurrent viable HCC and cHCC-CCA (P=0.533; Figure 2B).

In 20 patients, after exclusion of 5 patients showing HCC recurrence, 13 patients had ICC recurrence, and all of them died. In addition, 2 patients died from pneumonia (n=1) and from unknown causes (n=1). The 6-month, 1-year, 3-year, 5-year, and 10-year ICC recurrence rates were 30.0%, 50.0%, 55.0%, 65.0%, and 65.0%, respectively (Figure 3A). The 6-month, 1-year, 3-year, 5-year, 10-year, and 15-year OS rates were 90.0%, 80.0%, 55.0%, 49.5%, 32.1%, and 16.0%, respectively (Figure 3B). Of these 20 ICC patients without HCC recurrence, the AJCC tumor stage of ICC had no significant prognostic effect on ICC recurrence (P=0.394; Figure 4A) or OS (P=0.395; Figure 4B).

When we compared the timing of TR in 18 patients showing all-type TR, we found no significant difference between the ICC and HCC recurrence groups (P=0.368; Figure 5A). In addition, comparison of post-recurrence survival revealed no significant difference between the ICC and HCC recurrence groups (P=0.098; Figure 5B).

TUMOR RECURRENCE AND SURVIVAL OUTCOMES IN PATIENTS WITH VERY EARLY-STAGE ICC:

Of 7 recipients with very early-stage ICC, 3 experienced all-type TR: ICC recurrence occurred in 1 patient, and HCC recurrence in 2 patients. In 5 patients, after excluding 2 patients with HCC recurrence, 1 patient experienced ICC recurrence at 3 months after LT. The 6-month, 1-year, 3-year, 5-year, and 10-year ICC recurrence rates were all 20.0% (Figure 6A). The 6-month, 1-year, 3-year, 5-year, 10-year, and 15-year OS rates were all 80.0% (Figure 6B).

Discussion

ICC is rarely identified in explanted livers following LT, primarily because patients with confirmed or strongly suspected ICC are excluded from transplantation during pre-LT evaluation due to historically poor post-transplant outcomes [1–4,19,20]. Consequently, most cases of ICC in LT recipients are either misdiagnosed as HCC prior to LT, or are discovered incidentally upon pathological examination of the explanted liver [6–8]. In the present study, 24 out of 25 patients were diagnosed initially with HCC before LT, with 10 ultimately found to have ICC coexisting with HCC or cHCC-CCA. Notably, 19 patients had undergone HCC-directed treatment before LT, and 3 patients had 6 non-viable nodules, showing a complete pathological response. This raises the possibility that some of the 15 patients classified as having ICC alone may have had undetected concurrent HCC before LT. An international multicenter study [8] identified ICC, with or without coexisting HCC, in 0.3% of explanted livers (81 of 25 016 LT recipients across 17 centers), and in 1.1% of those who were preoperatively diagnosed with or suspected of having HCC (81 of 7503 patients). These findings align with our data, which show an ICC incidence rate among adult LT recipients of 0.4% (16 of 6611).

Occurrence of synchronous double primary HCC and ICC is rare, accounting for only 0.25% of all primary liver cancers [21]. In our previous study analyzing hepatic resection of primary liver malignancies over a 5-year span (2010–2014), we identified 3723 cases of HCC, 548 of ICC, and 145 of cHCC-CCA. Among these, double primary HCC and ICC cases comprised just 0.2% (10 of 4416) [22]. In a separate study involving LT recipients, we observed 16 cases of ICC, either as single or double primary tumors, within explanted livers over a 12-year period (2003–2014), representing an ICC incidence of 1.2% (16 of 1,293) and a double primary liver malignancy incidence of 0.6% (8 of 1,293) among adult LT recipients with primary liver malignancies [6]. This suggests that the detection rate of double primary liver cancers may be higher in LT cohorts than in resection cases, likely due to complete removal and thorough pathological examination of the entire liver following LT, which facilitates identification of previously unrecognized ICC or HCC lesions.

Co-occurrence of ICC with HCC or cHCC-CCA raises intriguing questions regarding their shared pathogenesis. One proposed mechanism involves differentiation from hepatic progenitor cells, which can give rise to both hepatocytic and cholangiocytic lineages. Several hypotheses have been suggested to explain the simultaneous emergence of HCC and ICC. The first is that these 2 tumors arise independently, with 1 tumor type undergoing trans-differentiation into the other. The second hypothesis is that a common progenitor cell undergoes malignant transformation and differentiates completely or partially into both HCC and ICC phenotypes [11]. An alternative explanation points to the role of chronic HBV infection, particularly in HBV-endemic regions. A Chinese study reported that all 35 patients with synchronous double primary HCC and ICC were HBV-positive [21], supporting the well-established link between HBV infection and hepatocarcinogenesis [23]. The HBV X protein, a key viral regulatory protein, is involved in HBV-related liver tumorigenesis [24]. In addition, epidemiological evidence suggests a causal relationship between chronic HBV infection and ICC development, especially in geographic areas where HBV is prevalent [25–27]. A cohort study from Taiwan reported an age-adjusted hazard ratio of 4.8 for ICC in patients with chronic HBV infection [28]. Similarly, a Korean single-center cohort study found that HBV-infected individuals had a 2.3-fold higher risk of developing ICC than non-infected controls [29]. This association was supported by 3 meta-analyses that reported hazard ratios ranging from 2.66 to 5.55 [30–32]. Collectively, these findings support the hypothesis that chronic HBV infection contributes to development of HCC, ICC, and cHCC-CCA, and even synchronous double primary HCC and ICC.

Given its rarity, post-transplant prognosis of ICC has been investigated in only a limited number of studies, with reported 3- to 5-year OS rates after LT between 20% and 40% [5,6]. As a result, ICC continues to be regarded as a contraindication for LT. An analysis of the United Network for Organ Sharing/Organ Procurement and Transplantation Network (UNOS/OPTN) database identified 280 cases of cholangiocarcinoma, including both perihilar and intrahepatic subtypes, among 71 224 LT recipients between 1987 and 2005, with an incidence rate of 0.4% [19]. The 5-year OS rate in this cohort was 38%, but dropped to less than 20% in patients with incidentally diagnosed cholangiocarcinoma at explant livers. A Spanish multicenter study of 23 patients with ICC reported 5-year disease-free survival (DFS) and OS rates of 27% and 42%, respectively, with perineural invasion and advanced tumor stage (III–IVA) identified as independent predictors of a poor prognosis [33]. Similarly, an American single-center study involving 13 ICC patients reported 1-year and 3-year DFS rates of 67% and 42%, respectively. Notably, none of the 4 patients with well-differentiated ICC had recurrence, whereas 78% of those with moderately-differentiated tumors did, suggesting that LT may be a feasible option for select patients with favorable tumor histology, particularly well-differentiated ICC [20]. By contrast, the present study found no significant differences in survival outcomes based on tumor stage, except for patients with very early-stage ICC, underscoring the inherently aggressive nature of ICC regardless of tumor staging.

A retrospective multicenter study involving 29 LT recipients with ICC reported that the risk of post-transplant TR was significantly associated with factors such as larger tumor size and volume, the presence of microscopic vascular invasion, and poor histological differentiation. Notably, none of the 8 patients classified as having very early-stage ICC experienced recurrence, in contrast to a 36.4% recurrence rate among those with solitary tumors measuring >2 cm, or those with multifocal disease (P=0.02) [7]. Similarly, an international multicenter analysis of 81 LT recipients, which comprised 48 patients with ICC alone and 33 with both ICC and HCC, reported markedly better outcomes for patients with very early ICC. In this group, the 1-year, 3-year, and 5-year recurrence rates were 7%, 18%, and 18%, respectively, compared with 30%, 47%, and 61% for patients with more advanced disease (p=0.01). Corresponding actuarial OS rates at 1, 3, and 5 years were 93%, 84%, and 65%, respectively, for the very early ICC group versus 79%, 50%, and 45%, respectively, for the advanced ICC group (p=0.02). These findings suggest that LT may be considered, albeit cautiously, for patients with cirrhosis and very early-stage ICC [8]. In the present study, 3 of 7 patients with very early ICC had concurrent HCC. After excluding 2 cases that recurred due to HCC, only one of the remaining 5 patients developed ICC recurrence. Consequently, the 10-year TR and OS rates for very early-stage ICC in our cohort were 20.0% and 80.0%, respectively. These findings suggest that patients with liver cirrhosis and suspected very early ICC may be considered for LT under stringent selection criteria; however, at our institution, biopsy-proven ICC of any size remains a contraindication for LT.

Among the 25 patients included in this study, 10 presented with double primary liver cancers at the time of LT, meaning that each malignancy carried its own risk of post-transplant recurrence. Overall, 18 of the 25 patients experienced TR, with ICC recurrence observed in 13 cases and HCC recurrence in 5. Notably, all recurrences occurred within the first 5 years after LT. Post-recurrence survival outcomes were uniformly poor, regardless of recurrent tumor types. These findings underscore the need for vigilant and comprehensive surveillance targeting both ICC and HCC within the first 5 years after transplantation.

This study has several limitations. As a retrospective analysis conducted at a single center in an HBV-endemic region with a limited sample size, it inherently carries constraints associated with such study designs. To better characterize the real-world clinicopathological features and post-transplant outcomes, large-scale, multi-regional, multicenter investigations are warranted. Notably, a key strength of this study lies in the complete follow-up of survival outcomes for all patients.

Conclusions

ICC can be diagnosed incidentally in the explanted liver. Most such patients have a history of HCC treatment, and more than one-third have HCC or cHCC-CCA as double primary liver cancers. The prognosis of incidentally-detected ICC following LT is poor due to high recurrence rates; the exception is patients with very early-stage ICC. Considering that all ICC and HCC recurrences occurred during the first 5 years after LT, strict post-transplant surveillance is mandatory during this high-risk period. After 5 years after transplant, less stringent surveillance for both ICC and HCC can be applicable as part of a cost-effective lifelong post-transplant follow-up strategy.