Clinical features and outcomes of combined hepatocellular carcinoma and cholangiocarcinoma versus hepatocellular carcinoma versus cholangiocarcinoma after surgical resection: a propensity score matching analysis

Background Combined hepatocellular carcinoma and cholangiocarcinoma (cHCC-CC) is an infrequent type of primary liver cancer that comprises hepatocellular carcinoma (HCC) and cholangiocarcinoma (CC). This study investigated the clinicopathological features and prognosis among cHCC-CC, HCC, and CC groups. Methods We prospectively collected the data of 608 patients who underwent surgical resection for liver cancer between 2011 and 2018 at E-Da Hospital, I-Shou University, Kaohsiung, Taiwan. Overall, 505 patients with cHCC-CC, HCC, and CC were included, and their clinicopathological features, overall survival (OS), and recurrence were recorded. OS and recurrence rates were analyzed using the Kaplan–Meier analysis. Results In the entire cohort, the median age was 61 years and 80% were men. Thirty-five (7.0%) had cHCC-CC, 419 (82.9%) had HCC, and 51 (10.1%) had CC. The clinicopathological features of the cHCC-CC group were more identical to those of the HCC group than the CC group. OS was significantly lower in the cHCC-CC group than in the HCC group but was not significantly higher in the cHCC-CC group than in the CC group. The median OS of cHCC-CC, HCC, and CC groups was 50.1 months [95% confidence interval (CI): 38.7–61.2], 62.3 months (CI: 42.1–72.9), and 36.2 months (CI: 15.4–56.5), respectively. Cumulative OS rates at 1, 3, and 5 years in cHCC-CC, HCC, and CC groups were 88.5%, 62.2%, and 44.0%; 91.2%, 76.1%, and 68.0%; and 72.0%, 48.1%, and 34.5%, respectively. After propensity score matching (PSM), OS in the cHCC-CC group was not significantly different from that in the HCC or CC group. However, OS was significantly higher in the HCC group than in the CC group before and after PSM. Furthermore, the disease-free survival was not significantly different among cHCC-CC, HCC, and CC groups before and after PSM. Conclusion The clinicopathological features of the cHCC-CC group were more identical to those of the HCC group than the CC group. The OS rate was significantly lower in the cHCC-CC group than the HCC group. However, after PSM, OS and disease-free survival in the cHCC-CC group were not significantly different from those in the HCC or CC group.


Patients and follow-up
This is a prospectively cohort study inclusive of 608 liver cancer patients underwent surgical resection from 2011 to 2018 at E-Da Hospital, I-Shou University, Kaohsiung, Taiwan. Of these, 103 patients were excluded because of metastatic liver tumors. Finally, our prospective study included 505 patients diagnosed with cHCC-CC, HCC, and CC confirmed by pathological findings (Fig. 1). This study was approved by the Institutional Review Board of E-Da Hospital, I-Shou University (EMRP32100N). Patients were diagnosed with cHCC-CC, HCC, and CC based on histological confirmation. Clinicopathological information, such as demographic data, etiology, liver cirrhosis, CP class, operative methods, tumor factors, alpha-fetoprotein (AFP) level, vascular invasion, metastasis, mortality, and recurrence were examined as our previous study. Liver cirrhosis was diagnosed based on pathologic findings. The liver preserved functional was evaluated using the CP scoring system [27].
Patients were followed up every 3-6 months through abdominal ultrasound, computed tomography, or magnetic resonance imaging. OS was defined as the time from the date of HCC diagnosis to the date of death, the last follow-up, or the end of the study in June 2019, whichever came first. disease-free survival (DFS) was defined as the time from the date of HCC diagnosis to the date of recurrence, the last follow-up, or the end of the study in June 2019, whichever came first.

Data analysis and statistics
Continuous data are expressed as medians and ranges. Categorical data are described as numbers and percentages. Normally distributed continuous variables were compared using Student's t test or one-way ANOVA test, and Wilcoxon rank-sum statistics were applied when two groups were compared and continuous variables were not normally distributed. The chi-squared test was used to compare categorical variables. Recurrence rates by various disease statuses were calculated and expressed per 100 person-years. Mortality rates by

Conclusion:
The clinicopathological features of the cHCC-CC group were more identical to those of the HCC group than the CC group. The OS rate was significantly lower in the cHCC-CC group than the HCC group. However, after PSM, OS and disease-free survival in the cHCC-CC group were not significantly different from those in the HCC or CC group.
Keywords: Combined hepatocellular carcinoma and cholangiocarcinoma, Overall survival, Recurrence, Prognosis  disease statuses were calculated and expressed per 100 person-years. OS and DFS was evaluated using the Kaplan-Meier method. Statistical differences in OS among subgroups were examined using the log-rank test. Median OS is presented as the median and 95% confidence interval (CI).
Logistic regression was used to perform PSM by using patients' sex, age, cirrhosis status, CP class, tumor size, tumor number, and AJCC stage to reduce bias in the analysis. Each group was matched with the control group (cHCC-CC group or CC group) according to the generated PSM by using a caliper width of 0.05. On the completion of matching, baseline covariates were compared using the paired t-test or Mann-Whitney U test for continuous variables and the chi-square test for categorical variables. A p value of < 0.05 was considered statistically significant. All statistical analyses were performed using SPSS version 23.0 (SPSS, Chicago, IL, USA).

Patients' baseline characteristics
Overall, 505 patients were included in this study (Fig. 1). The clinicopathological features of all cohorts are presented in Table 1. In the entire cohort, the median age was 61 years, the majority (80%) were men, approximately half had hepatitis B virus (HBV, 45.0%), onefourth had hepatitis C virus (HCV, 24.6%), and 20.2% had a history of alcohol use. Approximately 32.3% of patients had liver cirrhosis, with the majority (96.8%) having CP class A. Several patients (45.2%) had tumors ≥ 5 cm in size, and 10.2% had multiple tumors.
The cHCC-CC group had the highest proportion of patients with hypertension, smoking, alcohol use, CP class A, tumor size ≥ 5 cm, microvascular invasion, lymph node metastasis, AJCC stage I-II, recurrence per 100 person-years, and mortality per 100 person-years. By contrast, the HCC group had more patients with old age, male, cirrhosis, Edmondson-Steiner grades I-II, tumor number, AFP level ≥ 200 ng/mL, ICG%, operative margin > 1 cm, macrovascular invasion, and antiviral therapy. Moreover, the CC group was noted to more likely have diabetes mellitus, HBV, HCV, CP class A, major hepatectomy, and distal metastasis, as presented in Table 1.
Significant differences were observed between cHCC-CC and HCC groups in terms of age, diabetes mellitus status, tumor size ≥ 5 cm, lymph node metastasis, mortality, mortality per 100 person-years, and median follow-up time. Furthermore, significant differences were observed between cHCC-CC and CC groups in terms of diabetes mellitus, alcohol use, cirrhosis, AFP level ≥ 200 ng/ mL, operative margin > 1 cm, major hepatectomy, microvascular invasion, and macrovascular invasion. In addition, significant differences were noted between HCC and CC groups in terms of sex, hypertension, cirrhosis, Edmondson-Steiner Grades I-II, tumor size ≥ 5 cm, AFP level ≥ 200 ng/mL, operative margin > 1 cm, major hepatectomy, microvascular invasion, macrovascular invasion, lymph node metastasis, AJCC stage I-II, recurrence, recurrence per 100 person-years, mortality, mortality per 100 person-years, and median follow-up time.

Overall survival in different groups after PSM
PSM was performed using sex, age, cirrhosis status, CP class, tumor size, tumor number, and AJCC stage, and no significant differences were noted regarding crucial features (Tables 2, 3).
Comparing cHCC-CC and HCC groups after PSM (Table 2), there were 35 patients each in cHCC-CC and HCC groups. OS was not significantly different between cHCC-CC and HCC groups (p = 0.632, Fig. 3a).
When cHCC-CC and CC groups were compared after PSM (Table 2), no significant intergroup difference was noted regarding the recurrence rate (p = 0.213, Fig. 4b).

Discussion
cHCC-CC is a rare type of PLC [1,2], and its incidence accounts for 0.4-14.2% of PLC [1][2][3][4][5]. Our study results indicated that there were 35 (7.0%) patients with cHCC-CC out of the 505 patients who underwent surgery for PLC. The clinicopathological features of cHCC-CC are more identical to those of HCC than CC. The OS rate was significantly lower in the cHCC-CC group than in the HCC group. The OS rate was not significantly higher in the cHCC-CC group than in the CC group. After PSM, no significant differences were noted regarding the OS rate between the cHCC-CC group and the HCC or CC group. However, the OS rate was significantly higher in the HCC group than in the CC group before and after PSM. In addition, no significant differences were noted in terms of the DFS among cHCC-CC, HCC, and CC groups before and after PSM.
The clinicopathological features of the cHCC-CC group resembled those of the HCC group more than the CC group. Upon comparing cHCC-CC and HCC groups, no significant differences were observed regarding most demographic features, comorbidity, laboratory data, surgical methods, pathological characteristics, and tumor factors (including Edmondson-Steiner Grades I-II, tumor size, tumor number, microvascular invasion, macrovascular invasion, lymph node metastasis, AJCC stage I-II, and tumor recurrence) except for factors such as age, diabetes mellitus, tumor size ≥ 5 cm, lymph node metastasis, mortality, and mortality per 100 person-years.
When cHCC-CC and CC groups were compared, significant differences were noted regarding diabetes mellitus, alcohol use, cirrhosis, AFP level ≥ 200 ng/mL, operative margin > 1 cm, major hepatectomy, microvascular invasion, and macrovascular invasion. Our study's observation that the clinicopathologic features of cHCC-CC resembled those of the HCC group more than the CC group is inconsistent with the results of previous studies [7][8][9][10][11][12]14].
Our study determined that median OS was 50.1 months in the cHCC-CC group after surgical resection. This median OS was higher than that reported in previous studies, which concluded that the median OS of patients with cHCC-CC ranged from 20 to 47 months [3, 5, 7, 10, 12, 15-21, 24, 25]. Nevertheless, our study's finding of significantly lower OS in the cHCC-CC group before PSM is consistent with the results of previous studies [2,5]. Notably, OS in the cHCC-CC group was not significantly different compared with the HCC or CC group after PSM. This result differs from those of previous studies after the stagematched analysis [2,5,14]. We are the first to present the fact in the literature that no significant differences related to OS were observed between cHCC-CC and HCC groups after PSM. This finding is probably because the clinicopathologic features of cHCC-CC are similar to those of HCC, especially those related to tumor factors and tumor recurrence. Therefore, identical recurrence rates could have resulted in similar OS in cHCC-CC and HCC groups. Our study revealed that the recurrence rate was not significantly different among cHCC-CC, HCC, and CC groups before and after PSM. Our results are consistent with those of previous studies that revealed that the DFS was not significantly different among cHCC-CC, HCC, and CC groups [2,5,14]. This finding is probably because of the similarity in the clinicopathologic features of cHCC-CC and HCC groups, especially regarding tumor factors, such as Edmondson-Steiner grades I-II, tumor size, tumor number, microvascular invasion, macrovascular invasion, lymph node metastasis, distal metastasis, AJCC stage I-II, and tumor recurrence.
The limitations of our study were the small sample size of cHCC-CC and CC groups. This small sample size could have resulted in statistically nonsignificant differences related to OS and recurrence after PSM. Second, we did not analyze the molecular markers, tumor markers, and immunohistochemical characteristics of patients with cHCC-CC.

Conclusions
The clinicopathologic features of cHCC-CC resembled those of HCC more than CC. The OS rate was significantly lower in the cHCC-CC group than in the HCC group. The OS rate was not significantly different between cHCC-CC and CC groups. After PSM, the OS rate in the cHCC-CC group was not significantly different than that in the HCC or CC group. In addition, the DFS was not significantly different among cHCC-CC, HCC, and CC groups before and after PSM.