Clinical characteristics and outcome of patients with combined hepatocellular-cholangiocarcinoma—a European multicenter cohort

Background There is no clear consensus on the optimal systemic treatment regimen in combined hepatocellular-cholangiocarcinoma (cHCC-CCA) patients. We describe clinical characteristics and outcome of cHCC-CCA patients, with a special focus on patients receiving palliative systemic therapy, including immune checkpoint inhibitors (ICIs). Methods In this European retrospective, multicenter study, patients with histologically proven cHCC-CCA treated at four institutions between April 2003 and June 2022 were included. In patients receiving palliative systemic therapy, outcome was compared between cytotoxic chemotherapy (CHT)- and non-cytotoxic CHT (nCHT)-treated patients. Results Of 101 patients, the majority were male (n = 70, 69%) with a mean age of 64.6 ± 10.6 years. Only type of first-line treatment was independently associated with overall survival (OS). Palliative systemic therapy was administered to 44 (44%) patients. Of those, 25 (57%) patients received CHT and 19 (43%) had nCHT (n = 16 of them sorafenib) in systemic first line. Although there was no significant difference in overall response rate (ORR; CHT versus nCHT: 8% versus 5%), disease control rate (24% versus 21%), and median progression-free survival {3.0 months [95% confidence interval (CI) 1.4-4.6 months] versus 3.2 months (95% CI 2.8-3.6 months), P = 0.725}, there was a trend towards longer median OS in the CHT group [15.5 months (95% CI 8.0-23.0 months) versus 5.3 months (95% CI 0-12.5 months), P = 0.052]. However, in multivariable analysis, type of first-line regimen (CHT versus sorafenib) was not associated with OS. ORR in patients receiving ICIs (n = 7) was 29%. Conclusions In patients with cHCC-CCA, OS, progression-free survival, ORR, and disease control rate were not significantly different between individuals receiving CHT and patients receiving nCHT. Immunotherapy may be effective in a subset of patients. Prospective studies are needed to identify optimal systemic treatment regimens in cHCC-CCA.


INTRODUCTION
Primary liver cancer is the sixth most common malignant tumor and the third most common cause of cancer-related death globally. 1 Whereas hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (iCCA) represent the majority of cases, combined hepatocellularcholangiocarcinoma (cHCC-CCA) accounts for only up to 5% of primary liver tumors. 1,2 Per latest definition, classical cHCC-CCA requires the unequivocal exhibition of both cholangiocytic and hepatocytic differentiation within the same tumor on routine histopathology with hematoxyline eosin (H&E) staining, regardless of the percentage of each component. 2,3 Risk factors linked to the development of cHCC-CCA are largely unknown but may be similar to those reported for HCC. 4 Resection is the only curative treatment option, but recurrence rate is high, and many patients eventually require systemic therapy during the course of the disease. 4,5 Given the low prevalence of cHCC-CCA, difficulties with diagnosis, and the exclusion of patients with cHCC-CCA from clinical trials, there is no clear consensus on the optimal systemic treatment regimen. 4 Platinum-based chemotherapy (CHT) and the tyrosine kinase inhibitor (TKI) sorafenib are frequently used, 4 but the evidence level is low, as data were mainly derived from case series and small retrospective studies. [6][7][8][9][10] Data on immunotherapy with immune checkpoint inhibitors (ICIs), which have proven efficacy in HCC 11 and CCA, 12 are lacking in patients with cHCC-CCA. 4 In this retrospective, European multicenter study, we describe clinical characteristics and outcome of patients with histologically diagnosed cHCC-CCA, with a special focus on patients receiving palliative systemic therapy, including ICIs.

Study design
In this European multicenter study, patients with histologically proven cHCC-CCA, who were treated at four institutions in Austria and Germany between April 2003 and June 2022, were retrospectively included. The retrospective analysis of the data was approved by the local ethics committee of the Medical University of Vienna. Consent forms were waived due to the retrospective character of the study.

Patients and definitions
Eligible patients were adults (>18 years) with histologically proven cHCC-CCA. Patients with other primary liver tumors (e.g. HCC, CCA, fibrolamellar carcinoma), hepatic metastasis due to other primary malignancies, missing or incomplete histological reports, and patients with insufficient records were excluded from this study. Patient characteristics, laboratory parameters (within 90 days before baseline), tumor characteristics, information on previous/current treatments, and Eastern Cooperative Oncology Group performance status (ECOG PS) were collected from the clinical documentation system. Diagnosis of cHCC-CCA was based on pathological assessments of biopsy samples (n ¼ 52, 52%) and samples from surgical procedures (n ¼ 49, 49%). Different types of first-line treatments were categorized as follows: surgical treatment (resection, liver transplantation), local ablation (radiofrequency/microwave ablation, percutaneous ethanol instillation), locoregional treatment [transarterial chemoembolization (TACE), SIRT, radiation], palliative systemic therapy, and best supportive care (BSC). Baseline was defined as date of histological diagnosis (overall cohort) and start of systemic therapy (palliative systemic therapy cohort), respectively.

Palliative systemic therapy cohort
Patients with cHCC-CCA receiving palliative systemic therapy were assigned into a cytotoxic CHT versus non-cytotoxic CHT (nCHT) group. The CHT group included all conventional cytotoxic chemotherapies (i.e. platinum and non-platinum based chemotherapies) whereas TKIs and ICIs were included in the nCHT group. Decisions guiding the choice for type of first-line treatment were made in multidisciplinary tumor boards and were based on histological results, guidelines for HCC and CCA, and local therapy standards. All patients receiving palliative systemic therapy before 2007 (the year of sorafenib approval) received cytotoxic CHT (n ¼ 4). Patients who received classical adjuvant CHT after surgical resection or systemic therapy in combination with TACE for liver-limited disease were excluded from this cohort. One patient who received CHT after R1 resection was included. Three patients who received systemic therapy in combination with locoregional treatment were not included in either group. One of these patients, however, received conventional CHT after having progressed under the combination treatment and was therefore included at the time of CHT start.

Efficacy and definition of main outcomes
Duration of systemic therapy was defined as time from treatment initiation until date of treatment stop or death. Overall survival (OS) was defined as the time from diagnosis (overall cohort) or systemic therapy start (systemic therapy cohort) until date of death or last contact; patients who were alive/lost to follow-up were censored at last contact. Progression-free survival (PFS) was defined as time from systemic therapy start until date of radiological progression or date of death/last follow-up, whatever came first; patients who were alive/lost to follow-up without progression were censored at last contact. Overall response rate (ORR) was defined as the proportion of patients with complete response (CR) or partial response (PR), whereas disease control rate (DCR) was defined as the proportion of patients with CR, PR, or stable disease (SD) as best objective response. Best objective response was assessed according to the modified Response Evaluation Criteria in Solid Tumors (mRECIST). 13

Statistics
Statistical analyses were carried out using IBM SPSS Statistics 27 (SPSS Inc., Armonk, NY) and GraphPad Prism 9 (GraphPad Software, La Jolla, CA). Sankey plots were created using SankeyMATIC (https://sankeymatic.com/ about/). Continuous variables were reported as mean AE standard deviation or median (interquartile range), and categorical variables were shown as numbers (n) and proportions (%) of patients. Comparisons of proportions and of continuous variables were carried out by chi-square test and unpaired Student's t-test, respectively. Median estimated follow-up was calculated by the reverse Kaplane Meier method. The KaplaneMeier method was used to calculate survival curves and comparison was carried out by the log-rank test. In patients with incomplete dates, i.e. available month and year but missing day, the 15th of the respective month was used for calculations. Multivariable analysis was carried out by Cox regression analysis and variables with a P value <0.05 in univariable analysis were included. A two-sided P value 0.05 was considered statistically significant.
We next compared the outcome of cHCC-CCA patients treated with CHT (n ¼ 25) versus those treated with sorafenib (n ¼ 16) in systemic first line, and observed no differences regarding baseline characteristics (Supplementary  Table S1

Outcome of patients with cHCC-CCA treated with immunotherapy
Of seven patients receiving ICIs during their course of therapy, three (43%) patients received atezolizumab plus bevacizumab, two (29%) patients were treated with nivolumab alone, and one (14%) patient each received pembrolizumab alone and nivolumab in combination with TACE, respectively ( When comparing sorafenib with cytotoxic CHT, our results are partly in line with data from the United States, 7 Japan, 6 Korea, 8 and France. 10 In a retrospective, monocentric United States cohort of patients with cHCC-CCA receiving systemic therapy, DCR was lowest in those treated with sorafenib (n ¼ 7) compared with patients treated with gemcitabine plus platinum (n ¼ 41) or gemcitabine AE fluoropyrimidine (n ¼ 16) (DCR, 20% versus 78.4% versus 38.5%). Survival was also shorter in the sorafenib-treated subgroup (median OS, 9.6 versus 11.5 versus 11.7). 7 Similarly, in a retrospective, Japanese multicenter study including 36 patients with cHCC-CCA treated with systemic therapies, patients receiving sorafenib (n ¼ 5) had a shorter survival than those treated with fluorouracil plus cisplatin (n ¼ 11) and gemcitabine plus cisplatin (n ¼ 12) (median OS, 3.5 versus 11.9 versus 10.2 months). 6 Hence, both studies attested sorafenib only limited efficacy. The conclusions of both studies, however, are limited by the very small number of patients in the sorafenib group and were recently challenged by two larger studies. 8,10 A retrospective, single-center study from Korea found no significant differences in survival outcomes between sorafenib-(n ¼ 62) and cytotoxic CHT-treated (n ¼ 37) patients (median PFS, 4.2 versus 2.9 months; median OS, 10.7 versus 10.6 months). 8 Similar results were obtained in a French retrospective multicenter study, where OS and PFS of cHCC-CCA patients (n ¼ 25) treated with TKIs did not significantly differ from patients (n ¼ 54) receiving platinum-based regimens (median PFS, 2.8 versus 4.1 months; median OS, 8.3 versus 11.9 months). 10 Patients with cHCC-CCA may also be candidates for immunotherapy, as ICIs are effective in both HCC 11 and CCA. 12 Only recently, a subgroup of cHCC-CCA was identified showing features of a sustained intratumoral immune response as well as activated gene signatures predicting response to immunotherapy in HCC. This subclass was associated with improved survival and may benefit from ICIs. 14 Clinical data on immunotherapy in this rare liver tumor are limited, however, to a case report of a patient with metastatic cHCC-CCA who achieved a complete remission on third-line treatment with pembrolizumab. 15 In our cohort, a total of seven patients received ICI-based systemic therapies in different treatment lines, and achieved a promising ORR of 29% and a DCR of 43%. These data strongly suggest that ICI-based regimens should be further evaluated in patients with cHCC-CCA.
Limitations of this study include the retrospective nature with all its known potential shortcomings, the long recruitment period potentially influencing our results by

ESMO Open
changing clinical practices, as well as the large heterogeneity among included patients. Despite being one of the largest cohorts reporting on palliative systemic therapy (and the largest reporting on immunotherapy) in patients with cHCC-CCA published so far, the sample size is still small. Thus, confirmation of these results in large prospective trials would be desirable, but such studies may be difficult to conduct due to the low prevalence and difficulties in diagnosing cHCC-CCA.
In conclusion, type of first-line treatment was independently associated with worse OS in patients with cHCC-CCA. In the subgroup of patients who received palliative systemic therapy, there was a trend towards longer OS in cytotoxic CHT-treated patients (versus nCHT/sorafenib group), whereas other efficacy endpoints (PFS, ORR, and DCR) were not different. In multivariable analysis, type of first-line regimen (CHT versus sorafenib) was not associated with OS. Similarly to HCC and CCA, a proportion of patients with cHCC-CCA seem to benefit from ICIs.